Samenvatting: Business Research Methods

Deze samenvatting is gebaseerd op collegejaar 2012-2013. Bekijk hier ons huidige aanbod.


Druk: 3e 2011

Auteur: Bryman & Bell

 

 

 

Chapter 1: Business Research Strategies

Introduction

Business research cannot exist by itself. It means that business research always intersects with such sciences as sociology, psychology and economics (including such fields as marketing, accounting and finance).
There are two points which are especially vital in studying methodology. Firstly, methodology is not the same for everyone - different organizations and people have their own vision of how these methods must work. Secondly, it is important to take into consideration in what environment business research occurs. For example, you can research business environment in a small company where the situation is stable, or a relatively big one which plays a role in an acquisition or merge, or a company which went bankrupt and now sales all of its assets. In all three cases, different methods of research should be used in order to get a valuable result.
Quite often business researchers conduct studies based on their own past experience, which probably are of particular interest for them.
There are different opinions of how business research method should be conducted. These opinions have lead to two different “modes” of knowledge production:
Mode 1: Knowledge production is driven by academic agenda. The base of discoveries is existing knowledge. Knowledge production is a linear process. Audience and customers of this mode is academic community.
Mode 2: Knowledge production is driven by a process that makes boundaries of disciplines to be exceeded. In this mode discoveries are closely related to the case of study. This process is less linear than in “mode 1”. The audience consists of academic, policy makers and practitioners.
Some researchers suggest that “mode 2” of knowledge production is in many ways better than the first one.
Any new researcher must deal with different questions while conducting the research. The most important are what the aim of function of business research is and who the audiences of this research are.
Moreover, it is also difficult to choose a topic of business research. There are four points that must be evaluated:
The influence of researcher. It is vital to understand how the researcher collects data, analyses it and interprets the results.
Difficulty of understanding past researches. Methodological researches, as a rule, are described less detailed than, for example, sociological ones. For new researchers it may cause some difficulties in interpreting the results.
Kinds of methods. In some cases, it is hard to find past examples of researches of the topic, in others there are quite a lot of them. This must also be taken into consideration.
Research progress. While making a business research, one should carefully investigate how former researchers carried out their studies. This will help to improve and develop the ways in which methodological research is performed.
Theory and Research
Theories are divided into two groups: grand theories and theories of middle range. As a rule, grand theories are more abstract and general than of middle range, which operate in a limited domain. Management and business researches are driven by theories of middle range.
Empiricism is a theory of knowledge which asserts that knowledge arises from experience. In other words, every theory must first be tested before it can be considered as knowledge.
Deductive and Inductive Theory
There are two types of relationship between theory and research. It depends on whether we talk about deductive or inductive theory.
Deductive approach is a method in which a researcher has some data and based on this data tries to come up with a theory which later can be applied to similar cases. Deductive approach of relationship between theory and research can be described by following diagram:
 

 

Principle of deductivism: The purpose of theory is to generate hypothesis that can be tested and then explain the given theory.
As a rule, deductive method looks like the diagram above. Nevertheless, there are some cases when not all steps of the sequence are relevant to research. There are different reasons for this, some of which are:

  • someone else has published theory about his/her findings based on the same data before the researcher develops his own theory
  • the data may become relevant to the theory only after data has been collected
  • the data may not fit with original hypothesis

Some researchers prefer the inductive approach. This approach is reverse to deductive one.
 

 

Principle of inductivism: knowledge is arrived through gathering data and information and then provides a basis for theory.
Both deductive and inductive strategies are associated with qualitative approach.
Epistemological Considerations
Positivism is a position that promotes methods of the natural sciences to the study of social reality and beyond. Positivism contains elements of both deductive and inductive strategies. As a matter of fact, deductive method is expressed a little more in positivism than method of induction.
Phenomenology is a philosophy that investigates how individualists make sense of the word around them. It also takes into account that philosophers perceive the world in their own way, therefore these preconceptions must bracket out.
Interpretivism is an alternative to the positivism orthodoxy that respects differences between people and the objects in the system. Therefore it allows scientists to be more subjective if we compare it with positivism. Interpretivism includes Weber’s notion of Verstehen, the hermeneutic-phenomenological tradition and symbolic interactionism.
Ontological Considerations
Objectivism is an ontological position that states that social phenomena and their meanings exist independent and separate from human actions.
Constructionism is another ontological definition. It suggests that such categories as organization and culture are pre-given. Human actions have no impact on them.
Relationship of Epistemology and Ontology to Business Research
Paradigms is “a cluster of beliefs and dictates for scientists in a particular discipline influence what should be studied, how research should be done, [and] how results should be interpreted” (Bryman 1988a: 4). One of the most important features of paradigms is its incommensurability. It means that two paradigms cannot be consistent with each other.
Each paradigm consists of assumptions which are represented in one of two ways:

  • objectivism - there is an external viewpoint from which it is possible to analyze the organization or culture
  • subjectivism - organization and culture are studied by only those individualists who are directly involved in its activities

Epistemology and ontology related to business research/Paradigms
The main influence is that ontological commitments influence the ways how research questions are formulated and how they are executed. A paradigm is a collection of beliefs and dictates which for scientists (in a particular discipline) influence what should be studied, how the research should be conducted and how the results should be interpreted. Paradigms can be divided in two ways
Objectivist. The organization will be viewed of an external viewpoint and will be encompassed of consistent process and structures
Subjectivist. The organization is socially constructed, used by individuals to make sense of social experience so it can be studied from the point of view of the individuals involved.
Each paradigm also constructs assumptions about the function and purpose of the research in business:

  • Regulatory. The purpose is to describe what goes on in the organizations, but no judgements and possibly making small changes.
  • Radical. Judgements are made how organizations ought to be and produce ideas to achieve this

There are four possible paradigmatic positions:

  • functionalist - studies organizations and bases his study on solving of the problems which leads to radical explanation
  • interpretative - believes that understanding of organization (or culture) must be based on the experience of those who works (or lives) there
  • radical humanist - thinks of organization as a social arrangement from which employees must be emancipated
  • radical structuralist - understands organization as a product of power relationships which might result in conflicts within organization

Research Strategy: Quantitative and Qualitative
There are a lot of scientists and researchers who think that the difference between quantitative and qualitative research is not useful anymore. Nevertheless, there are certain important issues which differ quantitative and qualitative approaches. They are: principal, epistemological and ontological orientations. Quantitative research use deductive strategy, positivistic epistemology and objective ontology. Qualitative approach deals with inductive, Interpretism and constructionism orientations.

 

Nevertheless, it is false to think that it is impossible for research to deviate from these orientations. For example, qualitative researched is usually concerned with generation of the theory (inductive method), but not with its testing (deductive method). However, there were cases when qualitative research was used to test the theory rather than create it.
Influences on the Conduct of Business Research
Business research is influenced by five factors. Three of them were already discussed earlier (theory, epistemology and ontology). The other two are values and practical considerations.
 

Values reflect either personal beliefs of society or the feelings that create value for researcher. In second case, researcher must use objective orientation, otherwise experiment would be biased. A value can be the choice of research area and methods, implementation of data collection, its analysis and interpretation, etc.
The last factor influencing business research, practical considerations, should never be underestimated. This term implies choices of research strategy, design, method, the nature of people and topic being studied, and some other issues than change from case to case depending on nature of study.

Chapter 2: Research Design

Introduction
A research design provides a guideline for collection, analysis and interpretation of data. A research method is a way for collecting data.
Criteria in Business Research
There are three most important criteria for the evaluation of business and management research. They are: reliability, replication and validity.
Reliability reflects a question of whether the results of a certain study are repeatable. After all, if different researchers would finish the same experiment with different outcomes, it would be a question who of them was right and who made a mistake in his research. A reliable research is that one which outcome is the same every time whoever the researcher is.
Sometimes researchers replicate the findings of others for different reasons. Therefore, studies must be replicable. The reason is straightforward - if a researcher does not describe his outcomes in details, it will the impossible for other experimenters to interpret the results.
Validity is concerned with the integrity of conclusions of a research. In other words, it examines whether the research is conducted in a right way. Scientists distinguish four types of validity.
Measurement validity (or construct validity). This type of validity deals mainly with quantitative research. It is concerned with a question whether the measurement really reflects the measuring variable. If it does not, then both validity and reliability are questionable.
Internal validity. This type of validity deals with the question whether a conclusion that incorporates relationship between two variables really works. If we suggest that certain independent variable causes variation in dependent variable, we must be sure that this is true and not other factor is the reason for this variation.
External validity. It concerns with a question of whether the results of an experiment also work beyond specific research context.
Ecological validity. This last type of validity checks that scientific findings also work in everyday life. For example, if an experiment is made in a laboratory or in a special room, there is a great probability that findings will be ecologically invalid.
A variable is an attribute on which cases vary. If an attribute does not vary, we say that it is constant. Constant attributes are of less interest to the researchers than variables. Two most common types of variables are independent and dependent variables.
There are four aspects of trustworthiness that are parallel to some quantitative research criteria described before:

  • Credibility (parallel to internal validity) - the extent to which findings are believable.
  • Transferability (parallel to external validity) - question whether findings in particular research apply to other contexts.
  • Dependability (parallel to reliability) - concerns with a question whether findings also apply in other studies.
  • Confirmability (parallel to objectivity) - deals with the problem whether researcher’s values and views interfered experiment to a high degree.

Naturalism has many different meanings; the most common ones are the following. Naturalism (1) means a commitment to adopting the ideology of natural science methods; (2) means being fair to the nature of the phenomenon being studied; (3) is a type of research that tries to minimize the interruption of methods of data collection.
Qualitative research often concerns with naturalism. It means that it tries to collect data that is valid in environment, not only in laboratory. By and large, qualitative research in a greater degree deals with ecological validity than quantitative one.
Research Designs
In the rest of the chapter, we will discuss five types of research design: experimental, cross-sectional, longitudinal, case study and comparative designs.
Experimental Design
One of the characteristics of experimental design is that it involves significant confidence in the strength and trustworthiness of casual findings - it means that this design is strong in terms of internal validity.
If we conduct an experiment, we have to manipulate an independent variable in order to find out its impact on dependent one. There is a big disadvantage though. This disadvantage is the reason that there is much less experiments than it could be. There are a lot of variables that we cannot manipulate, for example, gender, age, share prices or interest rates.
It is necessary to distinguish between two types of experiments - laboratory experiment and field experiment.  The former one is conducted in a laboratory, while the latter one takes place in real-life settings.
The classical experimental design looks like following: there are two groups, one is called experimental group, the other is called control group.  In case of studying an effect of treatment, experimental group gets the treatment, control group does not.
Before getting a treatment, both groups are being observed for a certain measurement (which the researchers are interested in). After treatment, both groups are evaluated again. After this evaluation, researchers are able to find out whether there is a difference between the two groups.
This can be represented as a following schema:

 

T1 - before getting a treatment.
T2 - after getting a treatment.
Obsexp,1 and Obsexp,2 - measurements of experimental group before and after treatment, respectively.
Obscon,1 and Obscon,2 - measurements of control group before and after treatment, respectively.
If we talk about research design, a question arises - what is the purpose of a control group. In fact, it eliminates the influence of factors on the experiment other than independent variable (in our case, treatment). If experimental group is affected by any external factor, the control group is affected as well (sometimes to less or more degree, but still affected). Therefore, in the end, when we compare two groups, these side effects cancel each other out. Here are some examples of such factors:
Testing. It is possible that experimental group, knowing that it is being studied, behaves other than in everyday life. The control group also experiences this effect and therefore this effect is omitted when comparing the groups.
History. Some events in the past may cause the changes in the observations. If there would not be control group, we could not be sure that our experiment is not the case.
Maturation. Like everything else, people change, and these changes may have effects on the dependent variables. Since these changes apply to control group as well, we can discount this effect.
Selection. If both experimental and control groups have been selected randomly, there is less probability that differences between Obsexp,2 and Obscon,2 are caused by pre-existing differences.
Ambiguity about the direction of casual influence. Sometimes it is difficult to determine which variable causes the other. In some cases, existence of control group may help to solve this problem.
There are five main threats to external validity of an experiment: interaction of selection and treatment, interaction of setting and treatment, interaction of history and treatment, interaction effects of pre-testing and reactive effects of experimental arrangements.
In laboratory experiment, researcher has greater influence on the study than in field experiment. In laboratory, there might be interaction of selection and treatment and interaction of setting and treatment. On the other hand, there is no possibility of pre-testing effects since there are no pre-tests.
Quasi-experiments have some characteristic of experiments, but do not fulfill all internal validity requirements. As a rule, experiments without control group cannot be considered as quasi-experiments; however, there are some exceptions from the rule.
Evaluation research is a study that mainly answers the question whether the intervention has achieved its anticipated goals. Sometimes in evaluation research it is not feasible to divide participants into two groups; therefore evaluation research is semi-experimental study.
A true experiment has greater confidence in internal validity than other kind of research (laboratory experiment outperforms field one at this characteristic). As a rule, experimental design is used more often in quantitative rather than in qualitative research.


Cross-sectional Design
A cross-sectional design involves the collection of data from many cases at a single point in time in order to collect data (quantitative or quantifiable) to determine the impact of two or more variables, which are then examined to detect patterns of allocation.
More than one case. Variation can be established only when one or more cases are examined - researchers are more likely to encounter variation in different variables when they have several sources.
At a single point in time. In cross-sectional design studies, all the data is collected more or less simultaneously. In experiments, in contrast, the intervals between data collection can be very long.
Quantitative or quantifiable data. In order to evaluate variation, it is necessary to have a systematic approach of measuring variation. Quantitative method is the best one from this point of view.
Patterns of association. In cross-sectional design it is possible to examine relationships between variables, but not to manipulate one of them.
Survey research is a type of experiment in which structured interviews and questionnaires are conducted in cross-sectional design.
If we think about cross-sectional design in terms of validity, reliability and replicability, we can state that internal validity is weak, external validity is quite strong, ecological validity varies from case to case. Degree of reliability, replicability and measurement validity depends on the researcher himself and the way he conducts the research.
In cross-sectional design, collected data comprises n observations and n cases.
          
Although most of cross-sectional studies are quantitative, there are some rare cases when qualitative research belongs to this design. As a rule, it is cases where unstructured or semi-structured interview is involved.
Longitudinal Design
Longitudinal design is a type of research that is used to record changes in one case over several points of time. As a table below shows, there are n observations in each of n points of time.

There are two types of longitudinal design: panel study and cohort study. In former type, data is collected from randomly selected individuals or organizations on at least two occasions. In latter one, individuals and organizations are selected in a way that they all have certain characteristic, for example, having certain experience in life.
Reliability, replicability and validity of longitudinal design are almost the same as of cross-sectional.
Case Study Design
Case study research entails detailed and intensive analysis of a single case. This case might be represented by an organization, event or even person.
The way in which case study design is different from previous ones is that it uses idiographic approach - researcher is usually concerned with specific unique features of the case. First three designs, as a rule, use nomothetic approach - researchers try to generate statements that can be applied regardless of time and place.
The main disadvantage of case study is that it has extremely weak external and ecological validity, since special findings that refer to a particular case have low probability of working in the world generally.
There are five types of cases:
Critical case. A single case is studied in order to accept or reject a certain hypothesis.
Unique case, also called extreme case, is common focus in clinical studies.
Revelatory case is studied when there is a possibility to study a phenomenon which previously was unavailable for research.
Representative case. It is also called typical case. This type of case illustrates an everyday situation or form of organization.
Longitudinal case. This approach examines how things, people or organizations change over time.
On the whole, case studies can be represented by both inductive (theory generation) and deductive (theory testing) approaches.
Sometimes there is more than one case that is studied. Such research can be considered as comparative design research, since it is about comparing two cases. The way to distinguish between multi-case design and comparative design is to determine the focus of the study. If it is the cases themselves and their features, then the research is a multi-case study. Meanwhile, the focus of comparative design is not cases and their unique contexts, but constructing general findings that have little to do with these contexts.
Sometimes the researcher spends several months or even years on a particular case-study research. In that situation, case-study research has some common features with longitudinal research, which also studies a case for a long interval of time.

Comparative Design
The last design left for discussion is comparative design. As its name itself suggests, the aim of this type of research is to compare two or more cases. Its schema looks like this:

The most famous comparative design deals with differences in culture and nations. It is cross-cultural, or international research. The aim of this research is to collect data about different cultures and nations.
There are two approaches of comparative business research - cross-cultural and intercultural approaches. The former one implies comparing different business variables in various countries; aim of the latter one is to study the interaction between people with different cultural backgrounds.
There are a lot of difficulties involved in feasibility of cross-cultural comparative research, some of which are data translation problems and insensitivity to cultural and national differences between cases.
Reliability, replicability and validity are the same as for cross-sectional design.
Comparative design is sometimes called “hybrid” because of its similarities to other designs: cross-sectional in quantitative research, case-study when it is qualitative, some features are common with experiments and quasi-experiments.
To sum up, there are 5 research designs, each of them (except cross-sectional, have several approaches.

 

 

Chapter 4: Getting Started: Reviewing the Literature

Introduction
Reviewing literature is an important part of under- and postgraduate dissertations. Sometimes this task involves problems, such as determining which literature to choose and how to combine it. Therefore, an important aspect that this research has is including and excluding certain literature from reviewing.
Reviewing the Existing Literature and Engaging with What Others Have Written
A literature review is something that shows that somebody is able to engage in scholarly review after having studied some literature in the same field.
While reading a literature, it is vital to do following things:

  • Taking good notes, including small details about material read.
  • Develop critical thinking skills, since review is not just summary of literature, and you must be critical in this review and have your own unbiased opinion.
  • Showing why your research is important. If you don’t the reader may either get bored while reading your review or just consider it not worth reading at all.
  • Always keeping in mind that having written a review, you may want to refer again the literature you have already studied in your conclusions or summaries.
  • Not trying to use absolutely everything you read in literature review. Sometimes a material may be important for research, but it is enough to just keep this information in head and probably use it in further research.
  • Not stopping reading the literature once you have already started designing your own research.

Systematic review is a method to improve quality of a research if it reflects the bias of researcher. Systematic review process has several advantages, the most important of which are more strongly evidence-based review and providing more reliable and understandable foundation for further studies.
However, there are some limitations, for example the fact that the process of identifying relevant qualitative literature is very time-consuming compared to quantitative one.
Meta-analysis is a process of reviewing a large number of quantitative studies about certain variable, summarizing the results and comparing them in order to determine whether this variable really has an effect on other. Unfortunately, meta-analysis is not always feasible since not all publishers include important details about their quantitative research (e.g. sample size).
Meta-ethnography is a method very similar to meta-analysis. The biggest difference is that meta-analysis deals with quantitative research, whereas meta-ethnography - with qualitative. Meta-ethnography involves seven steps: getting started, deciding what is relevant to the initial interest, reading the studies, determining how the studies are related, translating the studies into one another, synthesizing translations and expressing the synthesis.
The key stages of reviewing are following:
Planning the review. It is important to determine and clarify purpose of research and its boundaries and then to monitor its progress.
Conducting a review. This stage includes comprehensive unbiased search for literature and then applying techniques such as meta-analysis or meta-ethnography.
Reporting and dissemination. The review should be accessible, readable and easy understandable for practitioners.
Narrative review process is contrary to systematic. Former are less focused and more wide-ranging comparing to latter ones.
Searching the Existing Literature and Looking for Business Information
Nowadays, the most famous way of gathering data is electronic databases search. It is easy to get lost while trying to find the needed literature. You have to be critical when checking your sources. It is also significant to determine right keywords for topic you write about.
The scheme below suggests one of numerous ways of searching the literature.
 

 

 

Referencing Your Work
Referencing shows that you understand the history of development of your works.
Today the most common referencing style is Harvard style. After paraphrasing or quoting an author, you put brackets with author’s name and year of publication of the appropriate source.
The second popular approach is footnote, or numeric. The footnote at the bottom of the page provides more details about certain piece of text. Furthermore, it is easier to understand than Harvard style, since in latter one the references may be as long as a sentence or even a paragraph.
Avoiding Plagiarism
Plagiarism is when you take other peoples’ opinions, words or inventions and use it as one’s own. Different universities have different rules and punishment systems in regard to plagiarism, nevertheless all students should take plagiarism and its consequences seriously. Even advanced researchers can make small mistakes when quoting someone which then may result in serious penalties.

Chapter 5: Ethics in Business Research

Introduction
When conducting a research, we should not forget about values. It means that while studying environment or people we have to be ethical.
There are four reasons why works about ethics are so frustrating:

  • Researchers differ from each other in their ethics researches, thus their works differ dramatically.
  • Main elements of this topic do not change - main issues described in 1960s were still the same in 1990s.
  • Ethics issues often touch some well-known cases. For example, studies such as unofficial company actions (Dalton, 1959) may involve a lot of questions whether it was ethical to conduct that research.
  • The last problem is that ethical concerns exist only in some methods, while in others they do not. It implies that the latter ones are immune from such kind of researches, which is not fair and correct.

Covert observations are those experiments in which a researcher plays a role of a participant but not of a researcher, and then conducts a study base on his own experience. Quite often researchers refer to the fact that they had no choice but to conduct a covert observation because of unavailability of information and data. As a rule, this is not true - there are just anticipated difficulties in getting this data, but not a real impossibility. If researcher wants his covert observation to be ethical, he must firstly try to get the information and then, after failing to get it, conduct this type of research.
There are four stances of ethics:
Universalism - a stance which assumes that ethical precepts should never be broken. On the other hand, universalist researchers sometimes make unethical decisions and actions.
Situation ethics (also known as principled relativism) - can be described in two ways. First states that if ethics rules would never be broken, we would still have no idea about some fields of society. The second says that sometimes we have no other choice but to break these rules.
Ethical transgression is pervasive. This argument says that “If the researcher is completely honest with participants about his activities, they will try to hide actions and attitudes they consider undesirable, and so will be dishonest. Consequently, the researcher must be dishonest to get honest data” (Gans, 1962)
Anything goes (more or less). This principle does not really mean that for researchers “anything goes” what is necessary to conduct their experiments, but that they have some flexibility in ethical decision making. Nowadays, there are not so many scientists who stick to this stance.
Ethical Principles
Ethical principles are divided into four groups:

  • whether there is harm to participants
  • whether there is a lack of informed consent
  • whether there is invasion of the privacy
  • whether deception is involved

Harm to participants may refer to physical and psychological harm, harm of participant’s self-esteem and many others types of harm.
It is responsibility of researcher to take care of absence of any kind of harm to participants. One of the solutions of such problems is a signing of confidentiality agreement with person or company. Again, if researcher publishes his/her findings, the names must be non-identifiable. As a rule, in quantitative research it is easier to anonymize records than in qualitative.
Another problem is that researchers sometimes cannot predict to what extend the harm will be or whether there is any risk of harm. There were some cases when researchers did not except such big harm on their participants.
Lack of informed contents implies that participants are not given the full information about the purpose of the research or degree of risk involved in the experiment. It causes some difficulties, since it is almost impossible to provide a participant with absolutely full information concerning the study, or to inform all participants in the same degree.
Invasion of privacy occurs mainly in covert observations. Invasion of privacy may happen because other participants do not know that they are being studied and give sensitive information which otherwise they would not. While in invasion is mainly expressed in covert research, other researchers often involve confidentiality and anonymity problems.
Deception occurs when researchers represent their experiment to the participants something other than what it is. In order to avoid this, some researchers tell their participants about some details of the study (which, in their opinion, have influence on participants’ answers) at the end, but not the beginning of experiment. Thereafter they give an opportunity to delete participants’ records if they want them to.
Other Ethical Considerations
Besides four abovementioned ethical issues, there are 3 others that are vital for a research.
Data protection. This issue mainly concerns confidential information. According to Data Protection Act, data must be processed fairly, accurate, up-to-date, and not kept longer than necessary.
Reciprocity and trust. In a research there should be mutual benefits for both researcher and participants. If possible, there should be some collaboration or active participation between them.
Affiliation and conflicts of interest. As a rule, every research is financially sponsored by a source that might have different interests in the outcome of the research. Even if it is not sponsored by a person or organization, funds are definitely got from elsewhere (e.g. government) and this source will have interest. Therefore, there is always a risk that some conflict will happen.
The Difficulties of Ethical Decision Making
One of the most difficult things to do while making a research is drawing a line between ethical and unethical behavior. There are a lot of sources about how to conduct interview and researches; however, sometimes these guidelines may cause in participants using these rules against researchers.

 

 

Chapter 6: The Nature of Quantitative Research

Introduction
Quantitative Research involves collection of numerical data, exhibiting a view between theory and research as deductive, natural science approach (especially positivism) and objectivist conception of reality.
The Main Steps in Quantitative Research
The figure below shows how quantitative research looks like. It does not necessary mean that there are no deviations from this schema; sometimes process excludes some of these operations and in rare cases includes additional ones.

 

 
Concepts and Their Measurements
Concepts represent points around which business research is concentrated. If concept is used in quantitative research, it has to be measured. Some examples of measurable concepts are organizational size, productivity, level of technology.
There are three reasons why we should measure a concept:
They allow us to determine fine differences between people and organizations.
It gives us a consistent device to make distinctions. Measures should generate consistent results, but not ones that occur as a result of natural changes.
It provides the basis for more precise estimates of the degree of relationships between concepts.
It is important to see the difference between indicators and measurements. Measurement is an index of something that can be unambiguously counted - in other words, it is a quantity. Indicator is something that already exists as a measure of a concept, and it cannot be calculated.
There are several ways to determine an indicator:

  • through questions
  • through recording individuals’ behavior
  • through official statistics
  • trough an examination of mass media

However, there are problems involved in using one single indicator:

  • It may incorrectly classify many individuals.
  • It may be too general.
  • We cannot make finer distinctions.

Because of abovementioned reasons it is useful to use several dimensions of concepts. Nevertheless, some researchers still prefer to focus on a single indicator.
Reliability
Reliability refers to the consistency of a measure of concept. To determine whether something is reliable or not, we must evaluate three factors; stability, internal reliability and inter-observer consistency.
Stability. To check stability, the most often used method is test-retest. It shows whether variables are stable in time. If we measure a group and after some time measure them again, and there will be little variation, then we say that the research is stable.
Internal reliability. Internal reliability shows whether results of one indicator in an experiment are consistent and coherent with results of another. Sometimes high correlation between variables is a sign of internal reliability. Correlation is a measurement which absolute value is between 0 and 1. Correlation of 0.8 or higher is an indicator of strong internal reliability. Cronbach’s alpha measures a degree of internal reliability, its coefficient is also between 0 and 1. Its value is acceptable if it is 0.8 or higher, in some cases the limit begins from 0.7.
Inter-observer consistency. When there are a lot of observers of an experiment, it is probable that their opinions will vary a lot. If their interpretations are similar, the inter-observer consistency presents.
 
Validity
Validity deals with the question of whether the measurement of a concept really measures it. There are five types of validity: face validity, concurrent validity, predictive validity, construct validity and convergent validity.
Face validity is a measure that replicates the content of concept of research.
To estimate concurrent validity, a researcher uses a criterion (for example, job satisfaction) that is relevant to the concept in the question.
Predictive validity is the same as concurrent one, except for the fact that it measure future criterion.
Construct validity is estimated by deducing hypotheses from a theory that that is relevant to certain concept.
Convergent validity is high if measurement of a concept is the same to the measurement of this concept developed through other methods.
On the whole, validity presumes reliability. If your measure is unreliable, it cannot be valid. The types of validity and reliability are summarized below.
 

 

 
The Main Preoccupations of Quantitative Researchers
In this section for preoccupations of quantitative researchers will be explained: measurement, causality, generalization and replication.
Measurement is the main preoccupation and this chapter has dealt with this term quite often.
Causality means that researchers may not always describe how the things are, but they explain the causes why these things are what they are. When experimental design is used, independent variable is manipulated in order to find out its result on dependent one.
Generalization it is important that researcher can state whether his findings can be generalized, and, if they can be, to what degree.
The research is replicable if it was not conducted in a way such the researcher was bias to this topic. If piece of work cannot be replicated, there is a serious question of whether the work itself is valid. However, replication is not a high-status criterion in natural science; that is the reason why it does not appear in print as often as it could.
The Critique of Quantitative Research
Four critical points are relevant to quantitative research: criticism of quantitative research in general as a research strategy, criticism of epistemological and ontological foundations of quantitative research and criticism of specific methods and research designs with which quantitative research is associated.
Four issues have to be underlined here:

  • Quantitative researchers fail to distinguish people and social institutions from “the world of nature”.
  • The measurement process possesses an artificial and spurious sense of precision and accuracy.
  • The reliance in instruments and procedures hinders the connection between research and everyday life.
  • The analysis of relationships between variable creates a static view of social life that is independent on people’s life.

Is it always Like This?
The difference between ideal type of quantitative research and actual practice always exists. For example, figure that is given in “Introduction” part of this chapter more likely represents guideline for quantitative research rather than strict instructions.
An example of source of this difference is “reverse operationism”. It implies that despite the fact that quantitative research deals with deductive method, sometimes it is possible to generate theory (inductive method) rather than to test it.
Another reason why there is a gap between ideal type and actual practice is that majority researchers do not follow recommended practice. For example, only 3% of articles contain the measurement of validity of certain experiments. It is not necessarily that research is unstable but invalid, but we cannot definitely know that it is not.
A good practice, as a rule, works with random, or probability sampling. However, many researchers use non-probability sampling for their studies, sometimes because of unavailability of random sampling, or cost and time it requires. There are cases when the opportunity of non-probability sampling is too good to miss; that is the reason why researchers do not want to switch to random one.

 

Chapter 7: Sampling

Introduction
Imagine the situation when you have to conduct a survey of 10 000 students of your university. It is almost impossible to send questionnaires to all students, and it is absolutely impossible to conduct interview with all of them. What should you do then? The answer is that you should sample students from total population of the whole university.
Population is universe of units (students, firms, etc.) from which you will choose some for investigation.
Sample is segment of population that is selected for investigation.
Sampling frame is the listing of all units of population.
Representative sample - is sample which reflects population accurately.
Probability sample is a sample which consists of randomly selected units.
Non -probability sample is a sample which consists of units selected for certain criteria, but not randomly.
Sampling error is difference between population and probability sample selected from it.
Non-sampling error is difference between population and non-probability sample selected from it.
Non-response - a source of non-sampling error when participants refuse to provide data.
Census - the enumeration of entire population.
There are three sources of bias in sampling:
If a non-probability or non-sampling method is used. In that case, a possibility arises that sampling will be affected by researcher’s own judgments and opinions. Furthermore, since selection process is not random, some units have greater probability to be included into sample than others. To avoid this, probability sampling method must be used.
If a sampling frame is inadequate, it cannot present the whole population even if probability sampling is used.
If some sample members refuse to participate or cannot be contacted (non-response). The issue here is that those who agree to participate may differ dramatically from those who do not. Sometimes it is possible to determine approximately the dissimilarities and their impact on research. In majority cases, however, the difference of such factors as attitudes or behavior cannot be evaluated.

This schema suggests how to conduct a social research:
 

Sampling Error
The significance of sampling error is crucial for research. For example, imagine we need to conduct a research of sets A and B, both consist of 36 units (72 in total). We want to sample 18 units, equally from both sets (9 per set). If we make no sampling mistake, our sample will look like this:
 

 

The following are examples with (1) little sampling error, (2) some sampling error and (3) significant sampling error.
                                    (2)        
         


(3)
 
Types of Probability Sample
Simple probability sample is the most basic form of probability sample. Taking previous example, we have a population of 10 000 students. Let’s suppose that we have opportunity to interview 250 of them. The probability of each student to be selected is 250/10000 = 1/40. Steps for conducting a simple random sample are:
Define the population (in our case, it is 10 000).
Select comprehensive sample frame.
Define sample size (in our case, 250).
Assign all students a number from 1 till 10 000.
Use a table of random numbers which selects 250 number between 1 and 10 000.
The students whose numbers are chosen, are the sample.
For example, student numbers may be 21, 1334, 6709, 65, 3, 243, and so on.
Systematic sample. Above we determined that out of 40 students only 1 will be studied. Systematic sample suggests that we randomly pick any number between 1 and 40 (say, 21) and then add to this number 40 to know the assigned numbers of other students. Thus, we will get numbers 21, 61,101,141,181,221 and so on.
The principle of stratified-random sampling is that the whole population is divided into groups by certain criteria (for example, our 10 000 student population can be divided according to their faculty). Then, we have to take sample from each group so the number of sample in this group divided by total sample is equal to number of students at this faculty divided by the whole population. For example, out of 10 000 students there are 2 000 learning International Business Administration. To choose how many students to sample from this faculty (given population of 10 000 and 250 sample size), we must solve the equation:
X/250 = 2 000/10 000
X/250 = 1/5
X=50
Thus, we take a sample of 50 students from IBA faculty. However, due to sample errors, this number can be a little more or less than 50, say 45 or 52.
Multi-stage cluster sampling is almost the same as stratified-random one, except for two details. It is easier to explain them on our example. First issue is, in stratified-random sampling, we chose each faculty and then decided upon sample size; now, we group faculties by their size and consider 2 biggest faculties as a whole group, 2 relatively big ones (yet smaller than in group one) - as another, medium ones - as third, 2 relatively small ones - as fourth and 2 smallest ones - as fifth (in this case we assume that there are 10 faculties in the university).
Second difference is that we do not choose sample size according to proportion to population, but simply divide our sample size by number of groups (250/5=50). This number of students will be surveyed at each group of faculties.
Qualities of a probability sample
The strength of a probability sample it makes it possible to make inferences from the sample to the population from it was derived. Findings derived from the sample can be generalized over the population. That is the main difference with equivalence sampling: this method strives findings to be equivalent between samples and population.
Sample Size
One of the most basic considerations about sample size is that its absolute value is important, but not his relative size to population. 200 students sample would have the same validity in both universities with 1 000 students and 20 000 students. The aim for increasing sample size is to decrease probability of sample error and to increase confidence interval. The more is the sample, the more stable are results and the more confident we are about our interpretations about population as a whole.
On the other hand, even though you see the difference in accuracy in either sample of 50 or 100, it is hardly to feel the difference between accuracies of samples of 25 000 and 50 000 (yet the latter will result in less sample error than the former). The reason for that is that after certain point (as a rule, about 1 000), the confidence level is increasing while increasing the sample size, but the rise happens at a decreasing rate. When high costs and time expenditure are involved, it is not feasible to increase sample size from this point on.
 

Another vital issue about sample size is non-response (discussed in the beginning of chapter). For example, if we know that response rate is 80%, and we need a sample size of 250, we would better use sample size of approximately 313, since 313*80%=250. Thus, we will get our 250 responses. Nowadays, the tendency of not responding to a survey increases.
Heterogeneity means that variations inside population are high, homogeneity - that they are low. The more heterogeneous the group is, the bigger sample size should be.
Response rate shows how many percent of all chosen participants agreed to participate. It can be calculated as:
Response rate = (number of usable questionnaires/ (total sample - unsuitable members of the sample)) x 100%.
Types of Non-probability Sampling
Non-probability sampling is any sampling that was not discussed above. There are three types of non-probability sampling - convenience, snowball and quota samples.
Convenience sampling is chosen by a researcher because of its availability - for example, professor samples students of only his university.
Snowball sampling is also used because of its availability. Here, one participant brings surveys to others; they bring it to someone else, and so on. The disadvantage of this method is that the findings cannot be generalized; advantage - it usually results in a very high response rate. This sampling works better in qualitative research than in quantitative in terms of generalization to the whole population.
The purpose of quota sampling is to generate a sample in terms of relative proportion of characteristics of its units.
Disadvantages of quota sampling:
Cannot be representative enough.
People sampled may not be typical.
Possible element of bias presents.
Non-random method of selection makes it impossible to calculate size of population.
Advantages of quota sampling:
Cheaper and quicker to conduct.
No need to wait for people to call back, since all questions are asked directly and immediately.
Talking about the speed of sampling, it is invaluable compared to others.
The figures below sum up types of sampling and of errors involved in sampling.
 

Limits to Generalization
One point to emphasize here is that generalizations of sampling can be limited only to this population from which this sample is taken. If we conduct a survey among several employees of a certain company, we can only generalize our findings to this company, but not to all companies in the same industry or the same market.
Error in Survey Research
Term error is made up of four factors:
Sampling error.
Sampling-related error.
Data collection error.
Data processing error.

 

 

Chapter 8: Structured interviewing

Introduction
In business research interview, the aim is to gather as much information as possible. Think of behaviour, attitudes, norms, values and beliefs. The primarily focus in this chapter will lie on a structured interview, but there are also other kinds of interviews.
Structured interview
A structured interview has several characteristics:
-The most important characteristic is that the replies can be collected in a database.
Therefore it is important to give all the interviewees the exact same questions and all in the same order as instructed.
-Questions are very specific and the interviewee has been given a limited amount of choices
-Typically used in a social survey research
Reducing variability error
In an interview, there is always variation among the answers. There is a true variation, and a variation that is caused by error. To entail a high level of validity, variability caused due to error should be reduced to a minimum. Most common errors are

  1. Poorly formulated questions
  2. The ‘way of asking’
  3. Misunderstanding by the interviewer
  4. Misunderstanding by the interviewee
  5. The method of information-recording
  6. The method of information-processing

Other types of interview

  • Semi structured interview. The interviewer has formulated several questions of a general form, but variation is possible. The questions are less specific and the interviewer usually has some freedom to ask further questions
  • Unstructured interview. This type of interview only contains a list of issues and topics to use in the interview. Lot of variation is possible and questioning is often informal.
  • Qualitative interview encompasses semi-structured and structured interview types.
  • Focused interview. Mainly open questions are asked about specific events relevant to either the interviewer or the interviewee. A variation of a focused interview is the focused group, where specific issues are discussed in a group. Related to the Focused group, is a group interview, where the difference mainly is that members discuss a variety of issues that might by only partly related.
  • Oral history interview is either semi-structured or unstructured. Interviewees are asked to memorize events and reflect on them.
  • Life history interview is very similar to the oral history interview, only a life history interview is unstructured and the goal is to obtain the whole ‘life story’ of the interviewee.

Interview contexts
There are more ways to conduct an interview.

  • Several (not one) respondent
  • Several (not one) interviewer
  • In person or by telephone

There are some advantages of interviewing by telephone:

  • Cheaper, as it saves travelling and processing time
  • Easier to supervise when more interviewers are conducting an interview
  • Respondents are sometimes affected by the appearance of the interviewer, in a telephone survey this bias is reduced to minimum

Disadvantages:

  • People without a phone cannot be called. People in a lower income-class are less likely to own a phone, and therefore the sample is biased.
  • Telephone interviews are often not as lengthy as a face-to-face interview
  • It is believed that response rates in telephone interviews are lower
  • Asking about sensitive issues tend to lead to a higher response in a personal interview
  • Technology has led to lower response rates and less access through conventional landlines. (think of mobile phones, answerphones)
  • Specification of the target group is more difficult, since it is harder to know if the correct person is replying
  • Showing objects visually is impossible
  • It is believed that the quality of telephone interviews tend to be inferior to personal interviews

Computer-assisted interviewing In computer-assisted interview, questions appear on a screen. Advantages:

  • After filter questions, certain answers may be skipped due to the reply of the interviewee, which eliminates asking inappropriate questions.
  • Quicker in processing information

There are two types: CATI, which stands for computer assisted personal interviewing and CATI, computer assisted telephone interviewing.
Conduct interviews
When conducting an interview, there are several important things to take for account:

  • Know the schedule from inside out
  • Give the respondents information about the goal, clear instruction, etc
  • It might be important to quickly build up a relationship with the interviewee
  • Think about the way you ask questions. Think of intonation, word order, etc.
  • Record the answer accurately
  • Ask questions in the right order, unless variation is part of the interview
  • Topic related questions should be asked in the beginning of the interview, and personal information questions should be asked in the end of the interview
  • Be careful with probing(=helping the respondent to understand the question, or asking for more information in an answer). This can influence the respondent and thus create error.
  • When prompting(=suggesting a possible answer in a closed question), it may be better to use show cards where the list of answers is given. This is easier to memorize for the respondent and less dull to read out answers several times, in opposite to reading the answers out loud.

Other approaches to structured interviewing
There are numerous other methods part of a structured or semi-structured interview, namely the critical incident method, PPP method, verbal protocol approach and repertory grid technique.

Critical incident method
When asking interviewees to recall critical incidents (particular types of events or behaviour) to understand their sequence and their effect on the individual, the critical incident method is used. Different kinds of situations that lead to negative or positive attitudes could be derived from the answers of the respondents.
Projective methods, pictorial and photo elicitation(PPP)
A projective technique consists of presenting visual, ambiguous incentives to respondents. The results are interpreted by a researcher to explore the underlying characteristics of the respondent. Think of an ink stain, which takes various forms. What a respondent sees in the stain is interpreted by the researcher to suggest dominant thinking manners. Pictures (or other visual objects) can also be used in a certain context(for instance after a long work day) to compare the results to a group under normal conditions.
Verbal protocol approach
This approach is used to understand problem solving. In this approach, respondents are asked to think out loud so the process of decision making can be monitored.
Repertory grid technique
This technique puts the individual as a scientist, determined to understand his environment in order to forecast and cope with future events. This is the typical procedure:
 First a number of categories which are relevant to the study will be set up. The researcher will ask questions to let the participant express how he or she sees the relationship between these categories. The participant’s picture of context is formed this way. The data will then be related to the elements’ underlying constructs of the participant’s ratio for sorting decision. Last, the participant is asked to rank each element in relation to each construct.
Problems with structured interviewing
Structured interviews have some pitfalls as well.

  • The characteristic of interviewers seem to affect interviewees, such as ethnicity
  • Acquiescence (consistently agreeing or disagreeing with a set of questions) may occur. This bias is caused by only providing statements that are ‘negative’. In this way, respondents will respond by consistently disagreeing. Thus, positive statements should also be provided to avoid acquiescence(since respondents will probably agree on ‘positive’ statements)
  • Because of the social desirability bias, people will tend to pick answers that are commonly accepted as a socially desirable answer.
  • Researcher and respondent may not share the same meaning of terms used in the interview, and hence suggest different things in their use of words.

 
 

Chapter 9: Self-completion Questionnaires

Introduction
In many aspects, structured interview is the best way to get secondary data. On the other hand, questionnaires are sometimes better than interview. In this chapter we will determine advantages and disadvantages of questionnaires, how to construct them in a right way and increase response rate.
Self-completion Questionnaire or Postal Questionnaire?
Self-completion questionnaire is sometimes referred to self-administered questionnaire. There are different types of self-completion questionnaires, the most widely used of which is mail (or postal) questionnaire. As its name suggests, this type of questionnaire is sent by mail and participants are asked to bring the filled questionnaire to certain place or send it back to sender. Self-completion questionnaire is a wider definition than postal questionnaire, though the latter one is the most popular form of the former.
Evaluating the Self-completion Questionnaire in Relation to the Structured Interview
There are a lot of similarities between self-completion questionnaire and structured interview. However, the major disadvantage of questionnaire is that there is no interviewer or supervisor, therefore participants must read and fill the questionnaire by themselves. This essential issue is a reason for questionnaires to be more understandable and clear to follow, since no one is assured that random participant is trained in answering surveys.
Comparing to interviews, questionnaires:

  • have more closed questions (as a rule, multiple-choice) than open ones, since closed one are easier to answer and there is less probability that participant will misunderstand instructions
  • are designed in an easy-understandable (but not necessary the best) way for participant so he or she do not miss or misinterpret a question
  • be as short as possible (but still covering many issues in which researcher is interested for) so respondent is not “afraid” of filling a long survey and does not throw it away

Advantages of self-completion questionnaires over structured interview:

  • Cheaper to administer. This advantage is expressed more when the survey involves a big geographical area. To conduct interview, researcher must travel a lot, which will result in high cost and long time. Sending a postal questionnaire to respondent is much cheaper even compared to telephone interview.
  • Quicker to administer. The main advantage of this issue is that questionnaires can be sent in large amount to different addresses simultaneously, while interviews require more trained interviewers for this papoose, and still it wouldn’t be as fast as postal questionnaires.
  • Absence of interviewer effects. This concern has both advantages and disadvantages. Now we will focus on advantages. The greatest importance is that the fact that interviewer does not present while participant forms a questionnaire leads to respondent being less biased towards sensitive questions.
  • No interviewer variability. It simply means that questionnaires do not have interviewers who ask questions in different order and manner.
  • Convenience for respondents. Questionnaires can be completed whenever and at whatever speed participants want.

Disadvantages of self-completion questionnaires over structured interview:

  • Cannot prompt. There is nobody to help respondents when they fill out a questionnaire. Therefore, questionnaires must be constructed in an easy way.
  • Cannot probe. Probing is important when open-ended questions are asked. On the other hand, open questions are not so popular in self-completion questionnaires; therefore this disadvantage is not expressed so much.
  • Cannot ask many questions that are salient to respondent. When questions are not salient to respondents, they are more likely to withdraw or stop answering questions during a survey than during interview. Research shows, however, that when question are salient to participants, they tend to answer them all.
  • Difficulty of asking other kinds of questions. Researcher must avoid asking lot of open questions since respondents may get tired or bored of writing much text.
  • Questionnaire can be read as a whole. Some respondents may firstly read the whole questionnaire and then start to answer the questions. Then, all answers are somehow biased and depend on each other.
  • Do not know who answers. With postal questionnaire you can never be sure whether it is answered by the person you have sent it or he outsourced this task to someone other.
  • Cannot collect additional data. If researcher wants to know some additional information about participant, it is difficult, if not impossible, to ask additional questions. The situation is easier when you send questionnaire to an organization - then you can retrieve more information about the company.
  • Difficult to ask lot of questions. As stated above, long questionnaires can “frighten” participants because of time and effort it requires.
  • Not appropriate for some kinds of respondents. Respondents whose literacy are limited or knowledge of English is low are not able to understand questionnaire and its instructions completely.
  • Greater risk of missing data. If someone fills in a questionnaire, and then decides to skip a question (for whatever reason), it causes in missing data. This problem is particularly expressed in quantitative research.
  • Lower response rates. The last, the most important disadvantage of questionnaires are response rates. It is very easy to refuse to answer the questions and just to throw the questionnaire away. The classification of response rates looks like this:
    over 85%    excellent
    70-85%    very good
    60-70%    acceptable
    50-60%     barely acceptable
    below 50%    not acceptable
    The last disadvantage requires a better look.

There are several strategies to increase respondent rates.

  • Write a good covering letter.
  • Accompany questionnaire by stamped address envelop.
  • Follow up individuals who do not answer by further letters.
  • Keep questionnaire as short as possible.
  • Questionnaires must have clear instructions and attractive layout.
  • Questionnaires must not be extremely bulky.
  • Try making questions salient to respondents.
  • Write a good cover letter.
  • Decrease number of open questions.
  • Provide monetary intensives.

Designing the Self-completion Questionnaire
There are several useful hints how to design a questionnaire. Firstly, do not cramp the presentation. If you have little free space in questionnaire, it seems to respondent that there is a lot of work even if you have relatively small amount of questions.
Secondly, make the appearance clear. Use different fonts for questions, instructions and possible answers.
When designing closed answers, better write them vertically than horizontally. In case of horizontal closed answers it is easier to confuse. Thus, do not use this:
What do you think about this company’s performance? (Tick the appropriate response)
Very good___Good___Fair___Poor___Very poor___

But this:
What do you think about this company’s performance? (Tick the appropriate response)
Very good___
Good___
Fair___
Poor___
Very poor___
Diaries as a Form of Self-completion Questionnaire
A diary is a method of data collection where respondent writes his/her activities during certain interval of time.

Advantages of the diary as a method of data collection:

  • It provides more valid and reliable data than questionnaire.
  • Performs better in questions about sequential behavior.
  • Disadvantages of the diary as a method of data collection:
  • More expensive than interviews.
  • People may think that they have already written enough even if they haven’t.
  • Diarists become less accurate in their notes over time.
  • Problems to note activity in the diary immediately.

Chapter 10: Asking questions

Introduction
The way questions are formulated, can have a great influence on the results. This chapter will elaborate on this.
Open and closed questions
Closed questions are questions where the answer will be either yes or no, indicating in which range one falls in (i.e. salary scale), or ‘ticking a box’. In an open question, the researcher wants to know more; they usually start with Why, What, How etc.
Advantages of open questions are:

  • Respondents can respond in their own words
  • Unusual responses can be derived
  • It gives more information than closed questions
  • Useful in exploring new areas
  • Useful to use in a fixed-choice format

Disadvantages:

  • Answers are time-consuming to process
  • The answers have to be coded
  • Require greater efforts and time from respondents
  • Recording answers differ between interviewers, which may cause variability

Advantages of closed questions are:

  • Easy to administer answers
  • Easier to compare answers
  • Closed questions are clearer
  • Easier to complete for respondents
  • Reduced variability in recording answers

Disadvantages:

  • Loss of potential interesting replies
  • It is difficult not to let answers overlap (such as age ranges)
  • Limited amount of possible answers
  • Respondent may feel irritated when one can’t find an answer that suits them
  • Impersonal

Types of questions
There are various types of questions to employ in an interview:

  • Personal factual questions (Asking for personal information)
  • Factual information about others (personal information about others)
  • Informant factual questions (Questions about characteristics of an entity they have knowledge from)
  • Attitude questions
  • Questions about belief(‘What do you think about..’)
  • Questions about norms and values
  • Knowledge questions (to test the knowledge of the respondents)

Rules for designing questions
There are some general rules of thumbs for question design:

  • Questions should be set up to answer your research question
  • What is it exactly what you want to know?
  • Turn it around: How would you answer it?

There are also specific rules:

  • Avoid ambiguous questions such as ‘often’. Rather ask for absolute frequency
  • Avoid long questions
  • Avoid questions that actually contain two questions
  • Avoid too general questions; rather use a frame of reference
  • Don’t use questions that already point in a specific direction (‘Do you agree with..”)
  • Minimize negatives in a question. Words like ‘not’ are often missed by the reader
  • Use simple, easy to understand language
  • Respondents should have the knowledge you’re interested in
  • Match closed questions with the answers
  • Make sure to have a balance in answers (evenly distributed negative and positive anwers)
  • Include ‘don’t know’ questions, so respondents aren’t forced into an answer. Accompany it with a filter question (a related question only focused on people who don’t know)

Vignette questions
Vignette questions are questions that have the following answer form: Strongly agree-Agree-Undecided-Disagree-Strongly disagree. In particular useful for subjects in a sensitive area, such as ethical questions.
Pre-testing questions
Pre-testing questions has many advantages:

  • Open question can be asked in a pilot to develop fixed-choice answers for an closed question-survey
  • It provides experience to interviewers
  • Identify questions where virtually everyone gives the same answer
  • Identify questions that make respondents feel uncomfortable
  • Identify questions that are not being understood
  • Find out how well the questions flow into each other

Existing questions
Using existing questions have numerous advantages:

  • If validated reliable, you know the quality of the questions
  • Possibility to draw comparisons

Chapter 11: Structured observation

Introduction
A method whereby the behaviour of individuals is systematically observed in a table of different categories is called structured observation. The main advantage is that behaviour can be directly observed, in opposite of survey research.
Problems with survey research:

  • Meaning. Terms can be interpreted differently by people
  • Omission. In an answer, key terms may accidently be omitted
  • Memory. People might have forgotten certain aspects of behaviour in terms of occurance
  • Social desirability effect. Replying in ways to be consistent with their perceptions of desirability of certain answers.
  • Question threat. People may fail to provide a honest answer to a question that appears threatening to them
  • Interview characteristics. Characteristics of the research may influence answers provided
  • Stated and actual behaviour gap. The difference between the stated behaviour by the respondent and the actual behaviour the respondent.

Observing behaviour
Structured observation may avoid problems that might occur with survey research.
The coding scheme is the core of any structured observation study. It specifies the categories of behaviour to be observed and how this behaviour should be designated to the categories. Structured data can be collected by several records. To clarify, here is an example:
 A researcher observes a manager for one week as they were at work as usual; taking phone calls, meetings, reporting and receiving visitors. The categories where: meetings, desk work, phone calls and tours. Data was collected by using three records:

  • Chronology record( describing activity outlines while noting time, duration etc)
  • Mail records(every incoming or outgoing mail and corresponding activity)
  • Contact record(describing each verbal contact)

From all this data several features of the job could be derived.
The observation schedule
A crucial step in a structured observation is a developing a schedule for the recording of the observations. There are several important characteristics:

  • It should be clear what or who is exactly to be observed.
  • The forms of any category should be both mututally exclusive and mutually inclusive
  • The recording system should be easy to operate
  • A certain amount of interpretation is required. For instance, it is hard to recognize the difference between an unscheduled meeting and a discussion which takes place in a hallway. There should be clear guidelines to the observer and a certain amount of experience to distinguish these two, since the difference is that in the latter the participants are less likely to be seated, which affects observation results.

Strategies for observing behaviour
Numerous strategies have been developed:

  • Observers can start recording after an incident has happened, to follow the consequences of the incident
  • Recording in terms of short periods. It is similar to the use of a sample. Small, representative time periods can be selected to be observed instead of one, longer period.
  • Recording in terms of long periods. It is the same approach as recording in terms of short periods, although in this strategy there are less selected time periods and the time period itself lasts longer.
  • Time sampling is the opposite of long periods-recording. The period could be only 5 minutes long, but the time between noting behaviour could be short as twenty seconds.

Sampling
Often it is hard to observe the whole population researchers are interested in, so samples are taken.

  • Sampling people Ideally, a researcher wants a random sample. If a sample is not random, the results will be biased. If in a job satisfaction research-sample low-paid employees set the overtone, the outcome may not represent the job satisfaction for higher-paid employees.
  • Sampling in terms of time Key example is here that participants should not always be observed at the same time of the day
  • Ad libitum sampling The observer records notes whatever is happening at the time
  • Focal sampling When an example of behaviour occurs that is interesting in terms of a schedule, this will be noted by the observer. This happens in a set period of time. This is the most used form of sampling.
  • Behaviour sampling An entire group is observed. When somebody is involved in a particular kind of behaviour, this will be recorded.
    Reliability and validity

As in all research methods, every method has issues in reliability and validity.
Structured observation’s reliability has the following features:

  • When two observers are employed, an inconsistency in coding behaviour might occur
  • The application of the observation schedule over time might be inconsistent.In a structured observation it is necessary to observe people in different occasions (other moments of time), since activities vary over time. If not properly applied, the results will be biased.

Validity
Measurement validity encompasses the following question: Is a measure measuring what it is supposed to measure? The following issues might occur:

  • The measure is not reflecting the concept it has been designed to measure
  • Mistakes in implementation of the measure in the research process (i.e not following the procedure)
  • The reactive effect. This relates to the change of behaviour of participants because they know they are being observed.

Other forms of structured observation

  • Field stimulation. This observation form focuses on the consequences that an intervention by the researcher has on the participants.
  • Organizational stimulation. Participants will be observed in an artificial setting which represents a situation. From this setting, individual and group behaviour is observed

Criticism on structured observation

  • A potentially irrelevant framework in the setting might be used to observe. The risk of this problem is larger when the setting is one about which is little known.
  • Structured observation does not allow the observer immediately to understand the meaning of behaviour, since it only observes behaviour and not the intentions behind it.
  • Data tends to be fragmented into categories. Certain activities might be observed, but are put in a category and thus not specifically displayed in the results.
  • Structured observations ignores the context of behaviour

Advantages of structured observation

  • When overt behaviour is the main focus of analysis, and meaning is less important, structured observation is more accurate than a questionnaire.
  • Structured observation does not provide reasons of behaviour. However, combined with a method that does provide reasons for behaviour, it is of great use.

Chapter 12: Content analysis

Introduction
Content analysis is a technique to put content into numbers in terms of predetermined categories in a replicable and systematic manner. The main use of content analysis is to examine mass media items. An example: in 2000 and in 2001 press articles about a company are gathered. In both years the amount of positive news is counted. Then the years will be compared in terms of positive news. Content analysis is often used in relation to research questions. When a research question is formulated, a content analysis is executed to find and examine information to answer the questions.
What are research questions?
Research questions often revolve around What, Who, Where, and How questions.
Selecting a sample

  • Sampling media. Which mass media might be chosen?
  • Sampling dates. From which years are we going to investigate? When using newspaper, which dates (editions) are used? The Monday or the Friday edition?

What is to be counted?
There are several kinds of units of analysis that are commonly used. The most important thing is, however, is what you actually want to count with respect to your research questions. It is also important to look behind the article if it is feasible to use.
There are several actors in this:

  • Who is the producer of the item, what is the background of the person?
  • To whom is the item focused?
  • If used, who are the alternative voices?
  • Of which context does the item derive?

Counting of analysis can be done by several techniques:

  • Subjects and themes. How many times is a subject encountered in two weeks of following a newspaper?
  • Disposition. How many times in an article did the writer (or reader) had a negative disposition?
  • Images. How frequently are images used and which types of images?

Coding
It is essential in content analysis to develop a coding manual and a coding schedule.
A coding schedule is in essence relating a specific dimension that is being coded. For instance, the code of question “What is your gender” is coded as i. The second question “What is your age” is coded as ii, for example
A coding manual is a report of instructions to coders that specifies the categories that will be used to classify a text. How the text is classified depends on written rules.
Dangers for coding schemes

  • There should not be overlap in the dimensions
  • The supplied categories in the dimensions should also not overlap
  • All possible categories should be available to the coders
  • Coders should not have any discretion in allocating the codes to the elements of analysis
  • The elements of analysis should be clear

Furthermore, coding must be consistent between coders, known as inter-coder reliability. Also over time, coders should be consistent: intra-coder reliability
Advantages of content analysis
Content analysis offers a method for the cultural study of organizations. It can access the organizational values because they can be derived from organizational documents. By measuring frequency of values, researchers are able to rate their importance.
There are several other advantages:

  • Content analysis is very transparent, and therefore easy to build upon by follow-up studies. Furthermore, it is seen as an objective method due to its transparency.
  • A long-time period study is possible within content analysis, and it allows the researcher to monitor changes over time.
  • Content analysis is a nonreactive method, since the researcher will not have to be taken in for account because it does not involve participants.
  • It is very flexible
  • Information of hard-accessibly social groups is can be generated from content analysis.

Disadvantages

  • The quality of the analysis is dependable of the quality of documents used
  • Interpretation is part of the coding manuals, which usually can attract biases
  • When the aim is to ascribe hidden content rather than manifest content, the potential of an invalid conclusion is enlarged.
  • It is difficult to find answers for “Why” questions.
  • Because content analysis is emphasised upon measuring, the accent will lie on what is actually measurable instead of theoretically important.

Chapter 13: Secondary analysis and official statistics

Introduction
Secondary analysis is the analysis of data done by another researcher that was not involved in the collection of data. This chapter involves two kinds of secondary analysis:

  1. Other researcher’s data
  2. Data that has been collected by other organizations in the same field of business

Other researcher’s data
Advantages:

  • It saves a lot of time and money to use available, good-quality data instead of executing research yourselves. Because of this, more time is available for the analysis.
  • Most data used for secondary analysis is of high quality (when derived from a regulated source) because the samples are representative, often countrywide and data itself is gathered by highly experienced researchers (such as government statistics)
  • Lengthy analysis is possible. Think of studies that are conducted every year, in this way trends can be monitored of a long period of time
  • When the source of data are large samples, these can be broken into subsamples to study them in particular.(subset analysis)
  • When databases of several countries are available, these can be used for cross-cultural analysis
  • Secondary analysis may yield new interpretations, since there are various ways to interpret data.
  • Data of business research is often only used for the primary goal; secondary analysis can make fuller use of the data.

Limitations

  • It takes time to become familiar with the structure of data
  • Data can be very complex
  • If data is not derived from a regulated source, the quality might be a problem
  • Key variables that the secondary analyst needs might be missing, because others gathered the data for their own purpose

Official statistics
Official statistics are data that has been gathered by the government, such as the level of unemployment in the UK. The use of official statistics for business research purposes is subject to controversy. There are some advantages:

  • It saves time and money, since the data is already collected
  • The reactivity problem will be minimal, because the data is not derived from questions to people
  • It has cross-sectional, cross-cultural and long-period research opportunities

Limitations

  • It involves interpretation of data
  • Agencies differ in the criteria used to record information (people working in the informal sector are recorded as unemployed, even though they have a job)

Reliability and validity
Reliability is affected by the change in definitions and policies regarding the recorded phenomena. If the informal sector is suddenly also defined as official employment, it affects the reliability and validity of employment data. Unemployment suddenly decreases, even though according to previous standards if would have stayed the same rate.
Ecological fallacy regards the error of assuming that individual members of a group have the same characteristics of that group as a whole. When this occurs, for example findings about an industry are wrongly concluded to every individual firm in that industry. It is comparable to the social term of stereotyping. Data can be misleading as well. When research official statistics shows that strikes occur more often in the manufacturing industry than in nursing, it does not show the reasons of the strikes. The variations might be rate of payment, working conditions, etc.
Official statistics as a form of unobtrusive measure
Unobtrusive measures are ‘any method of observation that directly removes the observer from the set of interactions or events being studied’ (Denzin 1970). Four types are mainly distinguished:

  1. Psychical traces left behind by a group such as graffiti and clutter
  2. Archive materials, collected by (non-)governmental organizations. Examples are diaries, mass media and historical records. Official statistics belong to this category.
  3. Simple observation, where the observer has no control over the behaviour and plays an unobserved, passive role in the research situation.
  4. Contrived observation is similar to simple observation, but the observer changes the setting in some way or utilizing hidden hardware such as video cameras.

The main advantage of unobtrusive measurements is that the reactivity bias is reduced to a minimum.
 

Chapter 14: Quantitative data analysis

Introduction
This chapter is concerned with the analysing quantitative data.  An important first remark is that you should think about how you are going to analyse the data before designing the research instruments. There are two reasons for this:

  1. Not just any technique can be applied to any variable
  2. The kind of technique that can be used will be constrained by the size and nature of the sample

Types of variable
Different types of variables are classified in four main types:

  • Interval/ratio variables. These are variables measured between certain intervals that are categorized. The time (minutes) people spend jogging is an interval variable. The distance between the categories is one minute, since the time is measured by minutes.
  • Ordinal variables are in essence the same as interval variables, only the distances between the categories don’t have to be equal throughout the questions. Think of the frequency when people go jogging, the answers could be:
    Once a week
    4-6 days
    2-3 days
    Once a week
    Less than once a week
    In this case, the distance between the categories is not the same, since the answer ‘4-6 days’ can either be a one, two or three day difference between the category ‘once a week’.
  • Nominal variables. These are variables that cannot be rank ordered. The reason why (relaxation, stamina, health) people go jogging cannot be ‘more’ than another.
  • Dichotomous variables. These variables have only two categories, male or female for instance.

 


14.2 Tool to categorize a variable
Univariate analysis
Univariate analysis means analysing one variable at a time. There are several approaches to execute this, the most common will be mentioned.

  • Frequency table. It provides the count of people and percentage belonging to the categories. A frequency table can showed as an interval variable, but also as an ordinal variable.
  • Diagrams. These are the main utilised methods. Think of pie charts, histograms and bar charts
  • Distribution of values. Examples are arithmetic mean, mode, and the median of values
  • Measures of distribution. That could include the range between the smallest and the largest value, the standard deviation (=average amount of variation around the mean) and a boxplot (see figure).


Q1 represents the lowest value, and Q4 the largest(without taking the outlier in account). 50% of the values fall in between Q2 and Q3. The dots above are called outliers, a value that differs so much from the rest of the values that is falls outside the boxplot.
Bivariate analysis, relationships/causality.
Bivariate analysis is about analysing two variables at a time to discover whether the two are related. An important remark here is that the analysis uncovers relationships but not causality, it cannot be inferred that one variable causes another. Even though a causal influence may look evident, it might work in the other way, therefore this point is stressed.
Several methods of bivariate analysis will be provided:

  • Contingency tables look like frequency tables, but it allows analysing two variables at the same time. I.e. gender might be included in the frequency of reasons why people go jogging. In this way, it is examined which reasons are most common, but now also shows the reason per gender. Hence, the relationship between gender and motivation for jogging can be analysed.
  • Pearson’s r method examines relationships between interval/ratio variables. The coefficient of relationship varies between -1 (strong negative relationship) to 1 (strong relationship) A coefficient of 0 means no relationship. A positive relationship means that when variable 1 increases, variable 2 also increases and vice versa. Therefore the coefficient might be negative.
  • Spearman’s rho method. It is similar to Pearson’s r method, only this method can be used to explore the relationship between ordinal and interval/ratio variables. See figure.

 

 


 

  • Phi and Cramér’s V is a method used to analyse the relationship between dichotomous variables and nominal variables. The Phi coefficient uncovers relationships between dichotomous variables. Like the previous methods, the results will vary between -1 and 1. Cramér’s V method is to be used as a method to discover the relationships between nominal variables. It can only take a positive value, thus can only give the strength of a relationship, not if it is negative or positive. This method is usually reported with a continency table.
  • A relationship can also be examined by comparing the means of two variables, such as the mean of the time spent jogging and the mean of frequency of jogging. This process is regularly accompanied by a test, called eta. This association will always be a positive number. Furthermore it is very flexible as it can process nominal and interval/ratio methods relationships.

Multivariate analysis
Multivariate analysis is the analysis of three or more variables.
Spurious relationships
When a relationship between two variables is being created by the impact of a third variable, the relationship is spurious. I.e. When job satisfaction and employee commitment have a positive relationship because of the leadership style, this is called a spurious relationship.
Intervening variable
Two variables that have a positive relationship do not have to be caused by a third variable. An intervening variable might explain the relationship between the two variables. It suggests that the relationship is an indirect one, thus one variable occurs via another variable.
Moderation
When a relationship holds for one variable, but not for another one (gender, for instance) it is said that the relationship is moderated by gender.
Statistical significance
There is no possible way of concluding whether results of a sample can be applied to the whole population. However, you can give an indication of how confidence you are in your findings. That is statistical significance. The level of statistical significance is the level of risk you are willing to take that there you wrongly inferred a relationship between two variables.
 
Errors
Two types of errors might occur at statistical significance:

  1. Type I error: risk that you’ve rejected the null hypothesis while it should be confirmed
  2. Type II error: risk of confirming the null hypothesis while it should be rejected

Correlation and statistical significance
Whether a coefficient of correlation is significant is affected by two factors:

  1. The size of the coefficient. I.e. If the coefficient is 0.62 and p: 0.1, there could be a chance of 10 out of 100 that there is no correlation. Therefore, the level of risk is not acceptable.
  2. The size of the sample. The larger the sample, the more likely the correlation coefficient is statistically significant.

Chapter 16: The Nature of Qualitative Research

Introduction
Qualitative research, unlike quantitative, deals with words but not numbers. Three other aspects must be underlined about qualitative method:
Inductive approach of relationship between theory and research, i.e. theory is generated through the data.
Epistemological position described as interpretivist.
Ontological position described as constructionist.
The term “qualitative research” is not straightforward. Sometimes it is defined simply as an approach to business research in which quantitative data is not collected or generated.
In other cases, it is described by comparing to quantitative research. The problem issue here is that in the end we find out what quantitative research is not.
Third option to describe this term is by the use of 4 traditions:

  • Naturalism - understands social reality in its own terms
  • Ethnomethodology - understands how social order is created
  • Emotionalism - concerns with inner reality of people
  • Postmodernism - explains different ways how social reality is constructed

There are five qualitative research methods:

  • Ethnography/ participant observation. Researcher observes and listens to participants through period of time to understand his point of view and social culture.
  • Qualitative interviewing. There are two types of qualitative interviews - semi-structured and unstructured. The third type of interview - structured - deals mainly with quantitative method.
  • Focus groups. A sort of interview which consists of open-questions about particular topic or event (which researcher is interested in) to be answered by group of respondents.
  • Language-based approaches to the collection of qualitative data, for example discourse or conversational analysis.
  • The collection and analysis of texts and documents.


 
Qualitative research, as a rule, is based on inductive approach, which means that data is collected and analyzed for theory generation.

The Main Steps in Qualitative Research
The picture below depicts a guideline for qualitative research.

 

Theory and Research
As stated before, theory generates through data collection and analysis. In rare cases, however, the reverse connection between theory and research is appropriate (the same issue that was covered in quantitative research). Such approach is being used lately to greater extend.
When just have started to collect data, the researcher may think of a question that probably this data collection does not lead anywhere and he might stay with nothing in the end. However, as the case flows and data is constantly collected, such hesitations, normally, disappear.
Concepts in Qualitative Research
There are two most common concepts - definitive and sensitizing.
Definitive concepts are the ones that become fixed after elaboration of relevant indicators.
Sensitizing concepts provide common sense of reference and guideline.
Researchers suggest that for qualitative research sensitizing concepts should be used and definitive ones are absolutely inappropriate.

Reliability and Validity in Qualitative Research
As in quantitative research, reliability and validity are also applicable for qualitative methods. However, many sciences have agreed that the meanings of these terms must be altered.
Validity and reliability can be divided into external and internal.

Some researchers suggest that different alternative criteria should be used in qualitative research than in quantitative.
Credibility, parallels internal validity
Transferability, parallels external validity
Dependability, parallels reliability
Confirmability, parallels objectivity
Respondent validation (also called member validation) is a process where researcher tells his participants about purpose and findings of his study. Disadvantages of this approach are that respondent validation may cause defensive reactions from participant and censorship and it is not certain whether or not participants validate the research.
Triangulation is a method where more than one source of data is used in studies of phenomena. Triangulation can operate both within and across research strategies. Despite triangulation is mainly associated with quantitative research, it can also refer to qualitative one. Besides, connection between quantitative and qualitative research often allows access to different levels of reality.
Even though complete objectivity cannot exist in a research, confirmability of qualitative research is high when researchers’ opinions and views play as little role as possible in the project.
Besides credibility, transferability, dependability and confirmability there are fifth factor that should be investigated - authenticity.
This criterion can be divided into five:

  • Fairness - Does the research fairly represent different viewpoints among members?
  • Ontological authenticity - Does research help members to understand environment?
  • Educative authenticity - Does research make members to appreciate perspectives of other members?
  • Catalytic authenticity - Has research changed members’ circumstances?
  • Tactical authenticity - Has the research empowered members to take necessary steps?

Another important criterion in qualitative research is relevance, especially when it deals contribution of the topic to the literature.
On the whole, different researchers vary in their definitions of all these standards while conducting their studies.

The Main Preoccupations of Qualitative Research
At the level of epistemology qualitative research is influenced by interpretivism.
Objects of analysis of natural science cannot attribute meaning to the environment, whereas objects of social science - people - can.
Qualitative researchers, compared to quantitative, like to deal with description of their experiments. It does not mean that they are focused on describing the research, but generally they provide explanations of why things are what they are - element of causality. One of the main reasons for that is that these details emphasize importance of understanding of contextual behavior.
Qualitative research views social life in terms of its processes. Frequently researchers try to show how things change over time. The main instrument for this purpose is ethnography and participant observations. However, it is also achievable by interviewing. Qualitative researchers prefer less structured form of interviews.
Another form of qualitative method is life history approach. However, this method is not used widely in the field of management and business.
The Critique of Qualitative Research
There are several critiques made by quantitative researchers about qualitative research:
Qualitative research is too subjective - it relies too much on researcher’s often unsystematic views.
It is difficult to replicate qualitative data - since it is unstructured and there is no strict rules of conducting, almost impossible to replicate.
Problem of generalization is highly expressed - there is a very serious concern that one or two cases are not likely to represent all cases.
Qualitative research lacks transparency - sometimes studies are unclear because author does not mention which method he/she used for the study.

Some Contrasts between Quantitative and Qualitative Research
The differences between quantitative and qualitative research can be summarized into the following table:

 

 

Differences between quantitative and qualitative research
Quantitative    Qualitative
Numbers    Words
Point of view of researcher    Point of view of participant
Researcher distant    Researcher close
Theory testing    Theory emergent
Static    Process
Structured    Unstructured
Generalization    Contextual understanding
Hard, reliable data    Rich, deep data
Macro    Micro
Behaviour    Meaning
Artificial settings    Natural settings
Similarities between quantitative and qualitative research

  • Both strive for data reduction to understand the collected data
  • Both strive to answer research questions
  • Both strive to related data analysis to the research literature
  • Both strive for variation to explore the factors that are related to the variation
  • Both strive to eliminate deliberate distortion
  • Both debate for the importance of transparency, allowing other to judge their work
  • Both address the occurrence of error
  • Both argue that research methods should fit the research questions

Researcher-subject Relationships
As stated before, in qualitative research the observer’s opinions and views play not the least roles.
Action research is defined as follows:
Experiments are on real problems in the organizations and are designed to assist with solution.
Iterative process of problem identification, planning, action and evaluation.
Leads to re-education and changing patterns of thinking and acting.
Action research plays a big role in connecting participant and observer. Its outputs are more interesting and understandable for audience, in particular practitioners.
Cognitive mapping is a method (complementary to action research) where maps are used as problem-solving devices.
Cognitive mapping generally consists of several pieces of text and arrows which show the connection between them.  An example of cognitive map is shown below:
 Image icon 1.jpg
Feminism is not suitable for quantitative research because:

  • It suppresses voices of women by ignoring them.
  • It turns women into object for studying.
  • Control is viewed as masculine approach.

On the contrary, qualitative research allows:

  • women’s voices to be heard
  • exploitation to be reduced
  • women not to be treated as objects
  • the emancipator goals of feminism to be realized

Collaborative and participatory research assumes that research should be driven by practical outcomes rather than by theoretical understanding. It can be seen as a form of respondent validation. It is a two-way process where researcher becomes involved in participant’s world and respondent is interested in observer’s experiment.
 
 

Chapter 17: Ethnography and Participant Observations

Introduction
The term ethnography is tightly connected with participant observation. Ethnography is a study where researcher (ethnographer) conducts his research based on behavior of his participants and sometimes involving several interviews with them about research topic.

Organizational Ethnography
Ethnography has certain degree of freedom for those who conduct such research in management and business field.
It is suggested that rules, strategies and meanings within works settings are different from those in social life. Ethnographer’s constant interaction with participants helps understand the world through their eyes.  
Access
One of the main issues to deal with in this chapter is access to organizations’ databases and culture. The below mentioned hints may help you to collect data if you conduct such research:

  • Use friends, contacts and colleagues to help you to gain access.
  • Try to get support from someone from organizations who will act as your champion.
  • Try to get access through top managers/ directors.
  • Offer something in return.
  • Provide a clear explanation of your aims and methods.
  • Be prepared to negotiate.
  • Be reasonably honest about amount of time the research will take from participant.

However, there is another interesting strategy that needs a closer look. This strategy is “hanging around”. Lately, it became very popular among researchers. All you need is to be in a right time in a right place with people who are directly involved with organization you study to hear what is necessary and try to take some note about that as soon as possible.
As described in previous chapters, covert observation is the one where researcher plays a role of participant. Covert observation is a form of ethnography, but ethnographers prefer overt to covert observations.
Advantages of covert observations:

  • There is no problem with access.
  • Reactivity is not a problem.

Disadvantages of covert observations:

  • The problem of taking notes.
  • The problem of not being able to use other methods.
  • Anxiety of ethnographer.
  • Ethical problems.

Roles for Ethnographers
Ethnographers can play four roles in their experiment:
Complete participant - researcher is a member of the social setting he studies and others do not know his true identity. The most common for - covert observations.
Participant-as-observer - the same as “complete participant”, with the difference that participants know that they are being studied.
Observer-as-participant - ethnographer conducts interviews majority of time, very seldom is involved in social settings.
Complete observer - researcher interacts with participants only during interviews.
 
Image icon 2.jpg
Image icon 3.jpg
Sampling
As a rule, the sampling of ethnographic research is a mixture of convenience sampling and snowball sampling (discussed in chapter 7).
However, theoretical sampling is used quite often as well (again, see chapter 7). The figure below shows general steps of theoretical sampling.

Image icon 4.jpg
 

Fieldnotes
There are some general principles about making fieldnotes:

  • Write down brief notes as soon as possible when you hear something interesting.
  • Write up full fieldnotes at the end of the same day the latest.
  • You may use tape recorder.
  • Notes must be vivid and clear.
  • You need to take copious notes.

For ethnographers, the main instruments for research are notepad and pen. Sometimes tape recorder is useful as well, sometimes it is not.
Visual ethnography makes use not only of fieldnotes, but also photographs, graphics, schemas and drawings.
There are 3 types of notes:
Mental notes - used when you don’t want participants to see that you take notes.
Jotted notes (also called scratch notes) - very brief notes that then will be used to elaborate an idea.
Full fieldnotes - detailed notes must be done as soon as possible.
The End
A vital point in ethnography and participant observation is to know where to stop. Unfortunately, because of its unstructured nature and absence of hypothesis to be tested, this matter is not a straightforward one and research lacks sense of end-point.
Generally, a professional researcher must just feel or know that he has already had enough. In worst cases, it becomes obvious from participants, which are tired or bored with study and have no more willingness to cooperate. Ethnographers must try to avoid the latter.
Besides, ethnographers must not forget about ethics issues and anonymize their observations if participants do not want to be identified.
Feminist ethnography
Feminist ethnography has more focus on the female aspect of research. Feminist research could be useful because it exposes the nature of management in terms of gender. Since organizations tend to be interpreted by a masculine perspective, female gender reinforcement helps to see things in a different light. Transparency plays an important role: transparency in the research process and in the dealing with the women she studies.
Visual ethnography
Using visual tools enjoys growing interest in social science. It has many advantages:

  • Organizational processes can be better understood
  • It can capture data that is unable to capture in an interview
  • Among departments it can be shown how the other departments’ work aspects looks like
  • Using an new way of data validation from respondents
  • involving staff in the debate of redesigning organizational processes
  • Source of data, not o accompanied with text

There is a distinction between extant and research-driven materials. Extant materials are materials that are already available within the company; think of annual reports and advertisements. Research driven materials are materials especially produced for the purpose of research.
The difference between visual ethnography and other research methods focusing on visual data is that visual materials are mainly used for the participants to interpret, and to build up a profile of the participants.

 

Chapter 18: Interviewing in Qualitative Research

Introduction
The interview is the most widely used method of qualitative research. There are three types of interviews: unstructured, semi-structured and structured. First two are mainly used in qualitative research, the third one - in quantitative.
What makes unstructured and semi-structured interviews attractive is their flexibility. Interviews themselves and their transcriptions are quite time consuming, however, this is rewarded by the fact that interview give more detailed description and more full information than any other type of research.
Difference between the Structured Interview and Qualitative Research Interviews
Unstructured and semi-structured interviews are also referred to as qualitative research interviews.
There are several vital differences between quantitative and qualitative interviews:

  • Qualitative interviews tend to be less structured than quantitative. In quantitative research, the researcher has clearly specified information to obtain. Therefore, deviations from beforehand prepared questions are very uncommon. Qualitative research, in its turn, is about more generalized type of questions and interviewer tries to get as much information as possible.
  • In qualitative research, observers are more interested in participants’ point of view and opinions.
  • In qualitative interview, there is more probability of deviation from the main research question, unlike quantitative one.
  • In qualitative research, interviewer can depart from any point of interview he wants, in quantitative he has to stick to the schedule.
  • The abovementioned factors result in qualitative interviews to be more flexible than quantitative.
  • In qualitative interviews, researcher needs detailed answers to the questions; while in quantitative interviews questions need to be coded and processed as quickly as possible.
  • Qualitative interviews can be conducted several times on the same topic. Quantitative, in contrast, are made, as a rule, only once (except longitudinal research).

Unstructured and Semi-structured interviewing
In unstructured interview, researchers prepare several questions (approximately 3-5) about research topic. Then, interviewee simply talk about issues that he considers worth to be mentioned.
In semi-structured interview, researchers have a list of questions that they follow up. However, they serve a purpose of guideline rather than something that has necessarily to be found out.
In both cases, interviews are flexible. The decision of choosing one of them or another is caused by following:

  • Degree to which researchers needs an understanding of participant’s world. The higher this degree is, the more unstructured the interview has to be.
  • If researcher begins his study more focused, then semi-structured interview is appropriate, if it is about getting general data - unstructured.
  • If there are several observers of one research, it is more likely that semi-structured interview will be implemented.
  • In multiple-case studies, in order to compare cases, semi-structured interviews are preferred.
  • There are two special forms of interviews - life history and oral history interviews.

It is important to request an interview in such a way that respondent agrees to it. It can be done either by mail or by telephone.
Another problem that interview deal with is about interviewing employees of a company. Sometimes it is not easy for middle- and low-level managers to find time for interview during their work hours. When employees are paid on hourly basis, it becomes even less possible. However, sometimes managers demonstrate willingness to participate in research, especially when research topic is particularly salient for them.
Interview guide does not necessary has to consist of written sentences. It can be pictures, figures, tables, keywords. While preparing it, one should ask himself a question what information he must possess in order to answer his research questions.
Some important issues in making a guideline are:

  • formulating interview  guideline in a way that it helps to fully exploit the research question
  • questions must be comprehensive and relevant to people that are to be interviewed
  • questions do not to be leading
  • some questions must be about interviewee - some of them general, others - detailed; this might be important while interpreting the answers

Other important details that to be checked before interview:

  • Make sure you are familiar with settings of where interviewee lives or works
  • Have a good microphone and tape recorder. If you make notes, you are more likely to make mistakes. A lot of interviews did not work out because of absence of good microphone.
  • Try to conduct interview in quite settings.
  • Prepare yourself for the interview as perfectly as you can.
  • After interview, you must make some notes not referring to topic question, such as:
  • how the interview went
  • where the interview took place
  • your other felling towards interview
  • description of settings, such as place and crowdedness

 Image icon 5.jpg

Questions the researchers ask during interviews can be divided into following groups:

  • Introducing questions.(I.e. asking about a start of event X)
  • Follow-up questions.(Asking to explain more about an answer)
  • Probing questions.(Asking to explain more about an answer on a direct question)
  • Specifying questions.(Asking for the sequel of an event)
  • Direct questions. (Straight to the point questions)
  • Indirect questions.(To gain the own view of the individual)
  • Structuring questions. (I.e. I would like to conclude now)
  • Silence (no questions are asked to make a small pause).
  • Interpreting questions.(Questions to confirm the ideas the interviewer has)

Quite often people are interested not only in what respondents answer but also in the way they answer. Some people can refuse to answer knowing they are being tape-recorded. In such situation, conduct interview anyways and try to make as many notes as possible.
On the other hand, in case when respondents agreed on microphone, the most valuable information is sometimes obtained after tape-recorder is switched off. Researchers must memorize as much as possible of this information.
It is wrong to assume that longer interviews always result in better results than shorter one. However, the longer time you have the interview, the more questions you can ask. In some cases, fortunately, very rare, interview becomes so short that it is not worth time and cost spent on it.
The problem of transcribing interviews is that it is a very time consuming process. One must decide whether to do it alone or to have an assistant for help. However, it is important that assistant is experienced at this work; otherwise it may lead to false interpretation of information which may result in serious consequences.
 
Image icon 6.jpg
Sampling
Majority of issues concerning sampling in ethnographic research are still valid in qualitative interviewing. Furthermore, you can put certain borders for characteristic of your sample, for example geographical area in order to avoid long distance costs and time, or put an age limit to avoid the influence of differences between generations.
Snowball sampling is sometimes used to group units for which no frames are made.
Sometimes, a probability sampling is used. However, differences between probability and non-probability sampling in qualitative research are much less expressed than in quantitative.
The main specialty of theoretical sample is that it determines whether or not more data needs to be collected.

Feminist Research and Interviewing in Qualitative Research
The qualitative research, unlike quantitative one, has big potential for feminist approach. For example, some characteristic of usual survey interview:
It is a one-way process. Researcher only gets information from his interviewee.
Researcher does not offer anything in reward for information obtained. Especially researchers conducting structured interview give almost zero information back to participants.
The element of power presents.
Due to abovementioned issues, structured interviews are inconsistent with feminism if interviewee is a woman.

Therefore, women conduct interviews which have following characteristic:

  • a high level of rapport between researcher and participant
  • a high degree of reciprocity
  • the perspective of the women being interviewed
  • a non-hierarchal relationships

Qualitative Interviewing versus Participant Observation
This section compares interviewing to participant observation.
Advantages of participant observations:

  • Seeing through others’ eyes. There is more contact between researcher and participant, and sometimes common activities are involved.
  • Learning the native language. To understand foreign country’s culture, its language has to be learned. Furthermore, not only official language is necessary, but also slangs and professional terms.
  • The taken for granted. Interviews cannot cover some data that can be covered by participant observation. As a rule, interviews rely on verbal behavior, and some things that are taken for granted cannot be expressed as much as during observations.
  • Deviant and hidden activities. Generally employees do not report hidden activities of its firm in personal interviews. This can only be detected by personal observation.
  • Sensitivity to context. Since participant observation deals with observations in many segments, it is easier to draw the full map of context.
  • Encountering the unexpected and flexibility. A characteristic that is common for participant observation and much unstructured interview.
  • Naturalistic emphasis. In participant observation researchers confront participants in their natural settings.

Advantages of qualitative interviews:

  • Issues resistant to observation. Some issues cannot be observed in natural situations, therefore the only way to find out about such matters are to ask directly during an interview.
  • Reconstruction of events. Unlike participant observations, interviews may make applicant to think about past and to explain its effect on current situation.
  • Ethical considerations. Some matters that can be observed without participant’s knowledge (especially during covert research) raise questions about ethics. Interviews, on other hand, rarely deal with such problems.
  • Reactive effects. Participant observations suffer from problem that presence of observer make others’ behavior more unnatural and disturbs the environment.
  • Less intrusive in people’s lives. Observations take require more participants’ time for being studied compared to interviews.
  • Longitudinal research easier. Observations may last months or even years, but they require constant observations. Sometimes it is too costly, not mentioning of researcher’s time consumption. Interviews last several hours maximum and the next one may be arranged in several years after first one (if it is necessary for longitudinal research).
  • Greater breadth of coverage. Undoubtedly, interviews cover much more spheres than do participant observations.
  • Specific focus. If researchers holds onto one certain question (generally, semi-structured interview), then the topics described are more focused than they could be in participant observations

To sum up:

 

Chapter 22: Qualitative Data Analysis

Introduction
One of the major difficulties in qualitative research is that it produces vast database in a relatively short period of time. The researcher must be careful while including data into their research since a lot of unnecessary information will make the research of less significance.
Unlike the quantitative research, in qualitative one there are very few accepted rules for collecting the data.
Three vital issues will be covered in this chapter:

  • General strategies of qualitative research, namely grounded theory and analytic induction.
  • Basic operations in qualitative data analysis involve steps and problems that deal with coding.
  • Narrative analysis is a field of management and business that nowadays is very popular among researchers.

General Strategies of Qualitative Data Analysis
General strategy of analysis of qualitative data is simply a framework or guideline of how data analysis should be conducted. Grounded theory and analytic induction do not only refer to data analysis, but also to its collection.
Analytical induction is a method of data analysis in which researcher seeks explanation of a hypothesis through studying cases. In process, if he finds a case that does not confirm the hypothesis, he either reformulates hypothesis or redefines theoretical explanation to exclude the deviant case. This process continues until no deviant cases are found and hypothesis is confirmed.
This method is very rigorous since one deviant case is enough to show that hypothesis explanation is wrong.
The main disadvantage of this approach is that it does not suggest and it is impossible to determine how many cases must be studied before hypothesis is accepted.

Grounded theory is a theory that was derived from data, systematically gathered and analyzed during research process. In this method data collection, theory and analysis have close relationships.
The figure below shows processes and outcomes of grounded theory.

 

 Grounded theory and ools of grounded theory
Grounded theory is a theory ‘that was derived from data, systematically gathered and analysed through the research process. Data collection, analysis and stand-in theory are in close relationship to another.’(Strauss and Corbin 1998:12) Tools of grounded theory are:

  • Theoretical sampling
  • Coding
  • Theoretical saturation
  • Constant comparison

Open coding is process of breaking down, examining, comparing and categorizing research data.
Axial coding is a procedure where data is put back together again after open coding, but this time already categorized.
Selective coding is the process of selecting category and filing in categories that need further investigation and development.
Products of grounded theory are:

  • Concepts
  • Categories
  • Properties
  • Hypotheses
  • Theory

Grounded theory is supposed to be good at:

  • Capturing complexity
  • Linking with practice
  • Facilitating theoretical work in substantive areas that have not been researched by others.
  • Putting life into well-established fields.

More on Coding
Professional coding consists of following recommendations:

  • Code as soon as possible.
  • Read through your initial set of transcripts, fieldnotes, documents, etc.
  • Read through your data again.
  • Review your codes.
  • Consider more theoretical ideas in relation to codes and data.
  • Remember that any item or piece of data can and should be coded in more than one way.
  • Do not worry to be generating many codes.
  • Keep coding in perspective.

Chapter 24: Breaking down the quantitative/qualitative divide

Quantitative research and constructionism
Quantitative research can be conducted in a constructionist way. This is best described by the outcome of an experiment. Researchers used a quantitative research by using data from magazine and newspaper articles. Students were asked to describe a CEO by reading articles of his company. It turned out that the images portrayed changed as the performance of the company varied, and so confirming the constructionism.
Epistomological and ontological considerations

  • In terms of epistemological and ontological reasoning, there are dissimilarities between quantitative and qualitative research. However, qualitative and quantitative research tends to be associated with ontological and epistemological points of view, while the relations are not perfect.
  • It has been suggested that when choosing a research method in a self-completion questionnaire, it more or less automatically selects a natural science model and an objectivist view. The same principal applies when using participation observation: it tends to imply a interpretivism and constructionism view.
  • When using the influential ‘four-paradigm’ framework it is expected to give a clear relation between the paradigm embraced and the utilised research method. This is not the case, since there is more backlash between research methods in epistemological and ontological terms.
  • Furthermore, research in business and management is often conducted by mixed methods in the same study.

Problems with the quantitative/qualitative contrast
Behaviour versus meaning

  • Quantitative researchers at least try to address meaning, although quantitative research may feel that they not really gain access to meaning due to imposed, pre-formulated attitude scales. Notable to say is that many quantitative research-related techniques have been shown to relate weakly to actual behaviour.
  • On the opposite, qualitative research tends to examinate behaviour in context such as norms, values and cultures.
  • Since both quantitative and qualitative researchers are interested in meaning and behaviour, and only differ in method of research, the behaviour/meaning contrast should not be exaggerated.

Theory test in research versus emergent from data

  • Quantitative research tends to categorized by theory-testing approach (hypothesis>test>analysis), while studies based on survey inclines a more exploratory view.
    This point of view fails to recognise the contributions of theories and new departures.
  • Furthermore, the proposition that quantitative research is only about theory-testing fall short in the acknowledgement that analysing and interpreting requires creativity.

Numbers versus Words
To expose the general aspects of phenomena, some quantification of findings can be used. For example: counting how many times specific characteristics are mentioned in a paper. Letting participants filling in the paper is qualitative research, but counting how many times specific characteristics are mentioned, is defined as quantification.
Qualitative researchers employ quasi-quantification by using terms like ‘many’, ‘often’ and ‘some’.
Artificial versus Natural
Qualitative research is often called a more natural way of research, since it studies people in their social worlds and context. There are several points of critics on this:

  • People are still taken away from activities they would normally do instead of participating in a interview. Ethnographic research would better suit in this, as this type of research observes people in their natural environment
  • There is very little known about how people feel and react about being interviewed
  • The naturalism of focus groups is often assumed, instead of demonstrated. The reason behind this lies in the participants’ view in the nature of their participation, which is unclear.
  • Interference by a researcher can raise the ‘artificialness’ of the research.

Qualitative analysis of quantitative data
More and more researchers are interested to test quantitative research by using methods that are associated with qualitative research. The main reason to apply this is to boost reliability of the research. In this way, the use of statistics means that business research has an appearance of a natural science and thus can count on greater creditibilty. Ethnostatistics  is an example of this:
-Ethnostatistics is ‘the study of the construction, interpretation and display of statistics in quantitative social research’(Gephart1988:9)
Quantitative analysis of qualitative data
The best example of quantitative analysis of qualitative data would be meta-ethnography;
Meta-ethnography follows the settings of qualitative research –interpretivism, sensitive for social context-, but it ignores factors of contex, to find relationships between variables that would be undiscovered when follow ethnographic research. There are some key issues:

  • Thoroughly search to find suitable literature for inclusion
  • Having considerable knowledge of the subject is key, embracing coding rules and pilot testing the coding schedule
  • the reliability of coding should be sufficient

Quantitative analysis of qualitative data has several advantages:

  • a quantitative researchers then has the possibility to execute investigations in multiple organizations
  • more data of more depth can be used
  • hypotheses of proved theories can be tested
  • A quantitative analysis of documentary accounts creates a link between observations and statistical methods.

Quantification of qualitative research
Criteria to identify of themes in qualitative data are often unclear, and therefore noting the frequency of certain incidents, words, etc might be a possible solution to denote a theme. This is called a Thematic analysis.
Quasi-quantification in qualitative research
Qualitative researchers tend to make  a hint to quantity by making use of words of frequency, such as often, sometimes and rarely.
Fighting the criticism of anecdotalism
Qualitative researchers are often criticised because publications on which it is based are often anecdotal. By doing this, it fails to provide a sense of how much influence the anecdotal items(such as snippets from interviews) are supposed to indicate. Quantification may help to fight this criticism, for instance counting how many times certain leadership styles were cited in numerous interviews.
 

Chapter 25: Mixed method research

Introduction
Although qualitative and quantitative research differs, they can also be combined. Combining the best of both worlds can give a better research outcome. This is called multi-strategy research, but preferably known as mixed method research. Mixed method research is the use of both quantitative and qualitative research methods.

The arguments against mixed methods research
There are two types of arguments: embedded methods argument and the paradigm argument. The embedded methods argument:
The decision how to collect data is not only the collect data itself, but also a commitment to position it belongs. For instance using a participant observation is harmonious with interpretivism, and not to positivism. It is argued when using mixed method research, these views collide with how social reality should be studied.
The paradigm argument
The paradigm argument argues that when combining a quantitative method such as a questionnaire with a qualitative method such as a participation observation, the research is only integrated superficially as paradigms are incompatible with each other. They are incompatible, because epistemological assumptions, values and methods are bound in a paradigm, and are not interchangeable.

Two versions of the debate about quantitative and qualitative research
Epistemological version:
Quantitative and qualitative research is established in separate epistemological standards, so in their nature they are different and thus unable to be used in a mixed method research
Technical version:
Quantitative and qualitative are related to each other by several epistemological and ontological assumptions, but these relationships are flexible. A research method is seen as capable of being used in another role.
The rise of mixed methods research
Mixed methods research has become more popular over the years. During 1994-2003 it increased by 200%. This method tends to be slightly more popular among business studies than social studies. This shows that mixed methods research has gained credibility in the field of business studies.
Approaches to mixed methods research
This part will explain different ways how to execute mixed method research.

  • Triangulation. This approach uses a method of one research strategy, and checks the results with a method belonging to another research strategy. A difficulty might be when the results are inconsistent, because then the question arises which conclusion has the right end.
  • Qualitative guiding of quantitative research. There are several ways how this can be done:
    • Providing hypotheses. Out of qualitative studies, a hypothesis can be derived. Subsequently, this can be used as a guideline for a quantitative research strategy which can test the hypothesis.
    • Aiding measurement. The design of an questionnaire can be developed from thorough knowledge gained by qualitative research.
  • Quantitative guiding of qualitative research (Selection)
  • From a larger population, the results of a questionnaire can be used to form a smaller but representative sample to be used for qualitative in-depth interviews.
  • Filling in the gaps. Sometimes a researcher cannot solely trust upon either quantitative or qualitative research method and therefore uses findings from the other research strategy.
  • Static and processual features. Qualitative research tends to be more focussed upon the process of social life, whereas quantitative research is more static. This originates from the nature of research: when observing (qualitative), the process can be easily recognized and described, while a questionnaire (quantitative) only notes that the process occurs.
  • Research issues and participants’ perspectives. Researchers sometimes want to gain two different kinds of data. Qualitative data will help them to gather the perspective of the participants, and quantitative data will help them to investigate specific subjects they are interested in. In this way, the focus lies on the participant’s meaning plus researching certain issues and therefore a more unstructured approach for the qualitative part can be used, and a more structured approach for the quantitative part of the study.
  • The problem of generality. Some quantification of results gained by qualitative research can be used to clarify the generality of the phenomena described. Bear in mind that quantification should reflect research participants’ own understanding of their social world.
  • Qualitative interpretation of the relationship between variables.In quantitative studies it is generally speaking hard to explain the relationship between variables. Qualitative methods can help to define these relationships. In example, quantitative research may show that businesses with fewer personnel perform better, but it doesn’t explain why they do. Here qualitative methods can help explaining this.
  • Studying different aspects of a phenomenon. This design of mixed methods research encourages to question which kind of research questions are best answered by a quantitative method, and which are best answered by a qualitative method. This is best explained by an example: a quantitative method is best used how employees use their time, and the understanding part (why are they using their time in that particular way?) is best investigated by a qualitative method.
  • Solving a puzzle. Outcomes of a research are not easy to anticipate. If survey responses are too low to use as the only source to rely upon, interviews and observations can be implemented to provide an alternative source.  It may occur that quantitative data is there, but fails to give a distinct and clear outcome, and then qualitative research can be used to fulfil the research.

Reflections on mixed methods research
There are several reasons why mixed methods research has become far more common:

  • a change in reasoning. Research methods are more seen as techniques of data collection or analysis, that are not impeded by epistemological an ontological nature.
  • a change in attitude by feminist researchers who had been resistant to quantitative research.

Mixed methods research better than mono-method research?
It is not true that mixed methods research is automatically superior to mono method research for several reasons:

  1. Mixed methods research has to be designed and executed correctly. This also applies to mono method research.
  2. Research questions should be appropriate to the research area you are interested in.
  3. It is likely to consume more time and money than relying on one method.
  4. It should be clear why one has chosen for a mixed method research and what was meant to achieve in the project.
  5. Mixed method research does not consist of separate components. The evidence should be brought together and not separately as quantitative and qualitative findings.
  6. Allocate details to all components sufficiently, so that none is over-detailed and none is superficial.
  7. Researchers should have the right skills to carry out mixed method research.

 

 

Join World Supporter
Join World Supporter
Log in or create your free account

Why create an account?

  • Your WorldSupporter account gives you access to all functionalities of the platform
  • Once you are logged in, you can:
    • Save pages to your favorites
    • Give feedback or share contributions
    • participate in discussions
    • share your own contributions through the 7 WorldSupporter tools
Follow the author: Vintage Supporter
Promotions
verzekering studeren in het buitenland

Ga jij binnenkort studeren in het buitenland?
Regel je zorg- en reisverzekering via JoHo!

Access level of this page
  • Public
  • WorldSupporters only
  • JoHo members
  • Private
Statistics
[totalcount]
Comments, Compliments & Kudos

Add new contribution

CAPTCHA
This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.
Image CAPTCHA
Enter the characters shown in the image.