What are the research methods in clinical psychology? - Chapter 3

What is research and what is meant with the scientific method?

Results of research should be replicable, meaning that other researchers can produce the exact same findings because collection of results has taken place under controlled conditions. If a scientific explanation gives suggestions of ways by which it can be tested or potentially falsified, it is testable. The scientific method relies on theories, which describe the cause-effect relationship in terms of sets of propositions. From there on a hypothesis can be created, which then explains the phenomenon and furthers investigation. Disconfirmed hypotheses are wrong or need to be changed.

What are the pros and cons of the scientific method?

The National Institute for Health and Clinical Excellence (NICE) is a UK organization that uses scientific evidence to see which treatments can be labeled as evidence-based, treatments that have proven to be efficient. These treatments are then recommended. In contrast to the scientific method, the alternative approach of social constructionism proposes that there are no basic 'truths' because reality is only a social construct. Because many disorders involve individuals developing their own realities, this approach is relevant in clinical psychology, seeing only language itself as means of understanding psychopathology.

What do we want to find out in clinical psychology research?

Goals of clinical psychology research entails description, prediction, control, and explanation (understanding). Description refers to definition and categorization of events that are of relevance in psychopathology. Prediction attempts to explain future behavior under specific conditions. Risk factors indicate which factors might contribute to a greater risk of developing psychopathology at some point in life. We control events and behavior in a way to learn more about the causal relationships that are involved and to find better treatment methods. Finally, we try to arrive at the stage of understanding. By using models of a phenomenon, we describe the interaction of all factors, in terms of a mechanism.

What questions can be answered with research?

Aetiology is a term used for the description of the origins and causes of symptoms of psychopathology. When research is done on a healthy population to gain knowledge about the aetiology of psychopathology, this is known as analogue research. Evaluative research or clinical audit aims at investigating whether a treatment has been effective. This supports the effective use of current knowledge.

Which research designs are used in clinical psychology?

What are correlational designs?

This type of research design allows a researcher to see if there is a relationship between two or more variables. Yet, this methodology does not provide a causal explanation of a relationship. The researcher needs to collect pairs of scores to perform a correlational analysis. Analysis can be done in computer programs such as The Statistical Package for the Social Sciences (SPSS). When computing correlation, the program will give the correlation coefficient r, which goes from +1.00 to -1.00, the former meaning a perfect positive correlation and the latter giving a perfect negative correlation. In a scatterplot the relationship between two variables can be displayed. Because of the differing nature of the relationships of the variables, the line of best fit differs with it. If the outcome of a study has a low probability of occurring by chance,

What are longitudinal studies and prospective designs?

The longitudinal studies, also known as prospective designs, do not obtain measures only at one point in time, but at several points to find out more about the time-relationship between variables. In the cross-sectional design only one sample from one point of time is taken.

What are epidemiological studies?

Epidemiological studies try to yield details about the prevalence of psychological disorders within a specific population over a set period of time. The frequency and distribution have been studied, giving an epidemiological study the form of a large-scale and descriptive survey. Prevalence rates can be described by lifetime prevalence, one-moth prevalence, and point-prevalence. Prevalence is represented by incidence x duration. It is important that the epidemiological study uses a sample that is a true representative of the population. This is hard to achieve as many people do not want to participate in studies.

What are types of experimental designs?

In an experiment, which can determine a causal relationship, a researcher starts with an experimental hypothesis or experimental prediction, which predicts how the experimental manipulation might affect the outcome response. The manipulated variable is called the independent variable and the variable of the outcome is called the dependent variable. Control conditions control for confounding effects, which are effects not produced by the independent variable but by something else. The independent variable is experienced in the experimental group, but not in the control group. Still, there has to be an objective way to measure the dependent variable, so statistical analysis can be applied. Random assignment is used to assign participants to the different experimental conditions. A participant might exhibit demand characteristics, which means that the participant thinks about certain behaviors that might be expected of him or her in the experiment and behaves according to that. The experimenter might also be prone to giving certain cues to a participant because of knowing to which experimental condition he or she was assigned. To prevent this, a double-blind experimental procedure can be used.

To avoid ethical difficulties, such as experimenting with humans who already experience difficulties or suffer from mental health disorders, experimenters often use analogue experiments. In analogue experiments, analogue populations are often used. Analogue populations are usually participants without any mental health problems and often consist of a normal sample of healthy, student participants. When using analogue populations, it needs to be considered what makes these populations valid analogues of psychopathology processes. There are three ways in which it can be argued that analogue studies are valid. First, it is often argued that psychopathology is dimensional rather than discrete. This means that psychopathology is often an extreme form of normal behaviours. Then, what one finds out about these behaviours in nonclinical populations will tell us something about the processes that cause the more severe reactions found in clinical populations. Second, one can use experimental manipulations to simulate mild psychopathology symptoms in nonclinical participants, and then study how this psychopathology may be alleviated. Third, nonclinical participants might be selected for an experimental study because they are similar to individuals with psychopathology.

In addition to analogue populations, researchers often use animal studies. Animal models allow researchers to experimentally investigate factors such as the genetics of psychopathology, changes in brain biochemistry associated with specific psychopathologies, and the effects of drugs on psychopathology. The advantage of this is that experimenters have complete control over the organism’s developmental history and permit the use of more intrusive experimental methods.

Another important experimental design is the use of clinical trials. In clinical trials, it can be examined whether a treatment is more effective than no treatment, whether treatment A is more effective than treatment B, or whether a newly developed treatment is more effective than existing treatments. In clinical trials, a placebo effect can arise, where a participant may improve simply because the procedure, they are undergoing leads them to believe they should or might get better. To control for this, researchers often use a placebo control condition.

What are mixed designs?

A mixed design is a type of research that works with participants that are not randomly assigned to groups, but rather assigned consciously to groups. This design can be especially helpful in psychopathology research because researchers can see how a group of participants with one specific psychopathology reacts to a variable in comparison to a group of participants with another psychopathology. Even though this technique is very useful, it does not provide information about the causal relationship.

What are natural experiments?

Observing the effects of naturally occurring events on behavior is another form of research, coined by the term natural experiments. These events can be a natural disaster, accidents, terrorist attacks or poverty and social deprivation.

What are single-case studies?

An individual's psychopathology can be studied by means of a case study, and case formulation can be a form of that used in therapy. Besides using case formulations, clinical practitioners can make use of the single-case experiment, which assesses an individual's behavior at two points, before the experimental manipulation and afterwards.

Case studies were one of the first research designs used, such as by Sigmund Freud with his psychoanalysis approach. Case studies are also a good source to disprove existing theories. Drawbacks of using case studies include little objectivity and control, low external validity, and the difficulty of providing evidence which supports the theory.

In a single-case experiment, the participant acts as both the experimental and control participant. For psychopathologies that are rare this is a good option, as often there are not enough participants for the experimental and control groups. Baseline measures of the behavior are taken before the experimental manipulation, which are then compared with the behavior after the manipulation. The ABA design is often used, which measures changes in behavior directly after the experimental manipulation, and hence controls for confounding factors that could occur within time after the manipulation. The ABAB design adds a second manipulation after the second stage of no intervention. Problematic is the alternation between periods of treatment with non-treatment if treatment provides important benefits for the participant. To solve this problem, the multiple-baseline design can be introduced. Two or more behaviors are measured in this design, while one behaviour is manipulated, and the other behaviors serve as the control condition. Another option is to use multiple participants, where in stage A the baseline measures are taken and then in stage B the manipulation is installed successively for all participants. A limitation of single-case studies is that results are not easily generalizable to other individuals, but this problem can be overcome by having more than one participant in the study.

What are mediators and moderators?

In statistical analyses, mediators are a third variable between the relationship of the independent and dependent variable. For instance, in a correlational study, a mediator can be examined to find out whether this variable is mediating the relationship between two variables. For instance, males weigh more than females, but this is mediated by length. So, a mediator explains the relationship between two variables.

What are meta-analyses and systematic reviews?

A meta-analysis attempts to provide an objective review of existing studies by using statistical methods. A particular finding across a number of studies is assessed in its strength, detecting possible trends by comparing effect size of the different studies. Effect size measures the magnitude of the effect in an objective and standardized way, whereby the problem of different numbers of participants, forms of measurement, and procedures of the studies is overcome. Meta-analyses therefore allow a comparison between different studies that deal with similar research topics. But meta-analysis has its limitations too, such as that published studies usually provide significant results, thereby overestimating mean effect-size. Also, meta-analysis does not control the quality of the studies they include, with effect size being influenced by the quality that the research was conducted with. A second method of assessing multiple studies is with the use of a systematic review. A systematic review is a literature review where a clearly formulated question is attempted to be answered, which is done with systematic and explicit methods for identifying, selecting and appraising relevant research.

What are qualitative methods?

Next to the quantitative methods that draw conclusions from studies using statistical inference, qualitative methods represent a growing body of methods used in clinical psychology research. They place emphasis on the raw material for research, with analysis being verbal and not statistical. Because the raw data are descriptions the participants make of their feelings, experiences and thoughts, the researcher can gain insight into the daily life of individuals with a certain psychopathology. The qualitative methods are usually open-ended and can also be a good precursor to studies using quantitative methods. Also, qualitative methods allow in-depth study of individuals and interesting things can be discovered which were not initially looked for.

In qualitative studies participants are usually deliberately assigned to groups, such as those with the same psychopathology. In the first phase participants are undertaken a semi-structured and open-ended interview. Participants can respond to specific but also general questions. The researcher then makes sense of the data, noting down and relating a participant's re-occurring issues in the responses back to the original research question. The grounded theory presents an approach to organize the gathered information into units. Abstract theoretical insights are formed from identified consistent themes within the data, which then help in forming a refined research question. Qualitative and quantitative methods can be combined very well. Often, research in a new area will start with qualitative data and then go over to or develop quantitative methods from there.

What are ethical issues in clinical psychology?

Ethical committees’ control for basic ethical standards in research to protect the participant. The three main ethical issues fall under these categories: informed consent, causing distress or withholding benefits, and privacy and confidentiality.

What is informed consent?

The informed consent is requested from participants prior to the study by means of an informed consent form which includes information about the purpose, procedure, duration, confidentiality, whether the participant is paid or participates voluntarily, and that they can leave the study any time without prejudice. The participant should also be given the opportunity to ask questions. Deception is sometimes used in the informed consent to prevent the results from being influenced due to revelation of all details of the study to the participant. This means for example that not all information is revealed of what the study will be about. At the end of the study, they will then receive a debriefing, an explanation of any deception or withheld information. At the end it comes down to ensuring that the participation in the study is truly voluntary. A problem with informed consent is that some participants might not be able to make rational decisions about their consent, like children or individuals suffering from mental problems. The consent of children is often obtained by a parent or guardian.

What about causing distress or withholding benefits?

The researcher needs to be attentive to any indications of stress that might come from the participant, and if this is the case, the study should be terminated or otherwise continued at a point when the participant feels well. A participant should never feel worse after having done a study. If the participant feels distressed after participation, the experimenter should provide a way for the participant to deal with these feelings (e.g., relaxation tape). Also, the ethical question stands whether one should withhold effective treatment in the no treatment control condition from a participant who would be in need for it (such as a new psychotherapy). A good way how to avoid this ethical conflict is to use participants who are on a waiting list for this treatment anyway, who would serve as the no treatment condition called waiting-controls.

Confidentiality and privacy

Privacy and confidentiality are rights that every participant in psychological research has. Confidentiality ensures participants that their data will be treated confidentially, and privacy enables participants to withhold some information that they do not want to provide, such as age or sexual orientation. The informed consent should always indicate who will access the data of the study if it collects personal information from the participant. Confidentiality is not absolute, like in case of illegal or immoral activities being revealed in data collection. For individuals who are in danger of harming themselves, the individual can be provided with reference to appropriate support after the study.

What can be concluded about research methods in clinical psychology?

Research methods in clinical psychology for understanding psychopathology allow the description of symptoms, the understanding of causes and the assessment of efficacy of interventions and treatment services. Many research methods are based on the scientific approach, which is often criticized for not fitting the investigation of important aspects of psychopathology, such as the phenomenology.