Why does an evil teacher force students to learn statistics? - ExamTests 1

Open questions

Question 1

What are qualitative research methods based on?

Question 2

When does falsification occur?

Question 3

What is a continuous variable?

Question 4

What does validity mean?

Question 5

When is a study reliable?

Question 6

What is the difference between correlationial and experimental research?

Question 7

Which three conditions should a causal relationship meet according to Hume?

Question 8

What is the difference between non-systematic and systematic variation?

Question 9

What does a normal distribution look like?

Question 10

What methods can be used to find the center of the frequency distribution?

Question 11

What scores do you need to calculate the distribution of scores?

Question 12

How do you calculate the standard deviation?

Question 13

What does a z-score indicate?

Question 14

What is the importance of randomization?

Question 15

Twenty-one heavy smokers are put on a threadmill. Researchers measured the time until these smokers were not able to run any longer. Time is measured in seconds. The times of the smokers were as follows: 18, 16, 18, 24, 23, 22, 22, 23, 26, 29, 32, 34, 34, 36, 36, 43, 42, 49, 46, 46, 57.

Calculate the mode, median, mean, upper & lower quartile distances, and the interquartile range.

Question 16

What is the difference between an independent and a dependent variable? Describe both concepts.

Question 17

What is the median of the scores 4-6-8-10-18?

Question 18

What is the median of the next row of numbers: 8, 9, 14, 15?

Question 19

The following five concepts are commonly used to summarize characteristics of a statistical variable: minimum, maximum, 1st quartile, 3rd quartile, median. What is the correct order, from smallest to largest?

Question 20

To get an impression of the relationship between the number of cigarettes smoked per day and the time required to jog 2 km, you make a graph from this data for a number of test subjects. Which variable do you put on the x-axis?

Answer indication

Question 1

Qualitative research methods are based on language.

Question 2

Falsification occurs when the collected data contradicts the theory or hypothesis.

Question 3

A continuous variable is a score that can take any value used on the measurement scale.

Question 4

Validity is the degree to which the instrument actually measures what you want to measure.

Question 5

A study is reliable when the instrument gives the same result under the same conditions.

Question 6

In experimental research the variables are manipulated, and in correlational research the relationship between variables is studied. However, correlational studies cannot make statements about cause-and-effect-relationships.

Question 7

According to Hume, a causal relationship must me the following three conditions:The three conditions that a causal relationship must meet according to Hume are:

• Cause and effect closely follow each other in time;

• The cause precedes the consequence;

• The effect never occurs without the cause occurring.

Question 8

The difference between non-systematic and systematic variation is that systematic variation can be explained, and non-systematic variation cannot.

Question 9

A normal distribution is a bell-shaped curve with symmetrical halves.

Question 10

Methods that can be used to find the center of the frequency distribution are the mode, median, and mean.

Question 11

To calculate the distribution of scores you need the highest and lowest scores.

Question 12

The standard deviation is calculated by squaring the difference between the scores and the mean, and take the sum of these outcomes. Next divide it by the sample size - 1 and take the square root of that.

Question 13

A z-score indicates how many standard deviations the score is from the mean.

Question 14

Randomization is important because it ensures that the "confounding" variables are gone. The confounding variables are factors that affect the dependent variable other than the factor you are interested in. For example, when there is a group of people, random allocation of people means that factors such as intelligence, age and race are generally the same so that the results are not affected.

Question 15

It is helpful to start by putting the numbers in the right order: 16, 18, 18, 22, 22, 23, 23, 24, 26, 29, 32, 34, 34, 36, 36, 42, 43 , 46, 46, 49, 57.

• The mode is calculated as follows: the scores with the frequencies in brackets are: 16 (1), 18 (2), 22 (2), 23 (2), 24 (1), 26 (1), 29 (1) , 32 (1), 34 (2), 36 (2), 42 (1), 43 (1), 46 (2), 49 (1), 57 (1). Because of this there are different modes: 18, 22, 23, 34, 36 and 46 seconds all have a frequency of 2. In this case 2 is the highest frequency.
• The median is calculated by the formula: (n + 1) / 2th score. In this case there are 21 scores, so 22/2 = 11th. The 11th score in 32 seconds.
• The average is calculated by taking the sum of all numbers and dividing it by the number of numbers. The average is 32.19 seconds.
• The lower quartile distance: this is the median of the bottom half of the scores. If we split the data at 32 (excluding this score), then there are 10 scores below this value. The median of 10 scores is 11/2 = 5.5th score. Therefore, we take the average of the fifth score and the sixth score. The fifth score is 22, the sixth score is 23; the lower quartile distance is therefore 22.5 seconds.
• The upper quartile distance is calculated by taking the median of the upper part of the scores: 42.5 seconds. The quartile distance is the highest score (57) minus the lowest score (16). This will be 41 seconds.
• The interquartile range is the difference between the highest and lowest quartiles: 42.5 - 22.5 = 20 seconds.

Question 16

An independent variable is a variable that is manipulated by the researcher. This often consists of two or more conditions to which the test subjects are exposed. The dependent variable is the variable that is observed after the independent variable has been manipulated.

Question 17

The median is 8.

Question 18

The median is 11.5. The numbers are already arranged from smallest to largest. There is an even number (8) of numbers. The median is then the mean of the two middle numbers: 9 and 14. This is: 11.5.

Question 19

Minimum, 1st quartile, median, 3rd quartile, maximum. It is obvious to start with minimum and end with maximum. Furthermore, the 1st quartile is the value of the variable for which 25% of the possible values ​​are smaller than this value. The median is the middle value, and the third quartile is the value for which 75% of the possible values ​​are less than this value. This determines the order from smallest to largest.

Question 20

The independent or explanatory variable is placed on the x-axis, and the dependent variable on the y-axis. The number of cigarettes smoked per day is the explanatory variable here: you want to know whether (and possibly how) this number affects the condition of the smokers.

What does statistics consist of? - ExamTests 2

Open questions

Question 1

On what formula are all statistical models based?

Question 2

What is the difference between variables and parameters?

Question 3

What is meant by a hypothetical value of the mean?

Question 4

How is the variance determined?

Question 5

What is the method of least squares?

Question 6

What is the sampling distribution used for?

Question 7

What is the confidence interval?

Question 8

How do you determine the boundaries of the confidence interval?

Question 9

What is Fisher's rule?

Question 10

What is the difference between a one-tailed and a two-tailed test?

Question 11

What is the difference between a type I and a type II error?

Question 12

What is the power of a test?

Question 13

What is the mean and standard deviation of the standard normal distribution?

Answer indication

Question 1

All statistic models are based on the formula: Outcome_i = (model) +error_i

Question 2

The difference between variables and parameters is that parameters are estimated and variables measured.

Question 3

The fact that the mean has a hypothetical value means that it does not actually have to appear in the data.

Question 4

The variance is determined by dividing the sum of the squared measurement errors by the number of degrees of freedom.

Question 5

The method of least squares means that the chosen parameter is always the one that produces the least error.

Question 6

The sampling distribution is used to determine how representative a sample is of the population.

Question 7

The confidence intervals are the boundaries between which you think the true population mean falls.

Question 8

The boundaries of the confidence interval are determined by multiplying the corresponding z-score by the standard error.

Question 9

Fisher's rule states that you only know if there is a real effect, if there is only a small chance that the result will be achieved by accident.

Question 10

The difference between a one-tailed and two-tailed test is that with a one-tailed test, the hypothesis indicates a direction for the effect, while this is not the case with the two-tailed test.

Question 11

The difference between type I and type II error is that the type I error means that there is thought to be an effect in the population when there is not. Type II error means that there is no effect in the population, while this is in fact the case.

Question 12

The power of a test is the degree to which it is able to find an effect.

Question 13

Mean is 0 and standard deviation is 1.

What are the limits of statistical research? - ExamTests 3

Open questions

Question 1

What is meant by effect size?

Question 2

In which three ways can the effect size be calculated?

Question 3

What is meant by the Cohen's d and what is its formula?

Question 4

What is the difference between the null hypothesis and the alternative hypothesis?

Question 5

What is the step-by-step plan for a null hypothesis significance test?

Question 6

What do you need to do to check if the hypothesis is a good explanation for the data?

Question 7

What are the three conditions for causality?

Question 8

What is meant by a meta-analysis?

Question 9

What is meant by the Bayes theory?

Question 10

What is meant by the Bayes factor?

Answer indication

Question 1

The effect size is an objective and standardized measure of the size of an observed effect.

Question 2

The ways to calculate the effect size are: the Cohen's d, the odds ratio and the Pearson correlation coefficient, r.

Question 3

The Cohen's d is the difference between two means divided by the standard deviation of the control group. This can be calculated by the formula: d = (X₁ – X₂) / s.

Question 4

The difference between the null hypothesis and the alternative hypothesis is that the null hypothesis assumes no effect, while the alternative hypothesis does assume an effect.

Question 5

The step-by-step plan for a null hypothesis significance test is that you look at how much variance the data explains in the model. Then you calculate the p-value, see how small that probability is and determine whether the null hypothesis is rejected.

Question 6

To check if the hypothesis is a good explanation for the data, divide the variance explained by the model by the variance not explained by the model, in other words: the effect by the error.

Question 7

The three ways are:

• The variables must covariate together
• The cause must precede the effect
• The influence of other variables must be excluded

Question 8

A meta-analysis is the estimation of the size of an effect in the population by combining effect sizes from different studies testing the same hypothesis.

Question 9

The Bayes theorem can be used to revisit and update the development of a previous hypothesis based on the observed data.

Question 10

A Bayesian factor is the ratio of the probabilities of two competing hypotheses. Usually these are the alternative hypothesis and the null hypothesis. A Bayesian factor greater than 1 means that the observed data is more likely, given the alternative. Values ​​less than 1 suggest the opposite. Values ​​between 1 and 3 reflect evidence for the alternative hypothesis that is "hardly worth mentioning," values ​​between 1 and 3 are evidence that "has substance," and values ​​between 3 and 10 are "strong" evidence.

What does the SPSS statistics environment look like? - ExamTests 4

Open questions

Question 1

What can be done in the data editor?

Question 2

What is the difference between the data view and the variable view?

Question 3

What do you do when you code a variable?

Question 4

What are the three options when entering missing data?

Question 5

What do you see on the output screen of SPSS?

Question 6

Why is it useful to also install a smart reader?

Question 7

What is the SPSS syntax?

Answer indication

Question 1

In the data editor you can enter data and perform statistical calculations with it.

Question 2

The difference between the data view and the variable view is that you enter numbers in the data view and look at the variable view what variables are used.

Question 3

When you code a variable you give a number to certain groups.

Question 4

The three options when specifying missing data are that there is no missing data, that some values ​​are missing, and that the values ​​in a particular area are the missing values.

Question 5

On the output screen of SPSS you can see all graphs and tables and a tree diagram of all the analyzes you have performed.

Question 6

It is useful to install smartreader, because it allows you to open all files with old or new versions of SPSS.

Question 7

SPSS syntax is the language of commands to perform statistical analysis and data manipulation.

In which way can data be explored with graphs? - ExamTests 5

Open questions

Question 1

What is chartjunk?

Question 2

How can you make graphs in SPSS?

Question 3

What is the difference between a simple and a stacked histogram?

Question 4

What is a population pyramid?

Question 5

What are the three different types of box plots?

Question 6

What is the most common way to display means?

Question 7

What is the difference between line charts and bar charts?

Question 8

What is the difference between a simple line and a multiple line?

Question 9

What is a scatter plot good for?

Question 10

Why is the regression line important in the scatter plot?

Question 11

How do you edit a graph in SPSS?

Answer indication

Question 1

A chartjunk is the name for the unnecessary things on a chart.

Question 2

In SPSS you can create charts with the chart builder.

Question 3

The difference between a simple and stacked histogram is that a simple histogram shows the frequencies of a simple variable, while you use a stacked histogram if you want to see different groups.

Question 4

A population pyramid are two graphs with their bottoms against each other, so that the frequency is shown in the horizontal axis.

Question 5

The three different types of boxplots are simple boxplot, clustered boxplot and 1-D boxplot.

Question 6

The most common way to display means is to use a bar chart.

Question 7

The difference between line graphs and bar graphs is that the information in a line graph is represented by a line.

Question 8

The difference between a simple line and a multiple line is that the simple line shows the means of the scores of different groups, while a multiple line also shows the mean of a particular variable and can consist of multiple lines for multiple variables.

Question 9

A scatter plot is good for looking at the relationship between two variables.

Question 10

The regression line is important in the scatter plot, because this line summarizes the relationship between two variables.

Question 11

You can edit a chart in SPSS by using the chart editor.

What is meant by the bias beast? - ExamTests 6

MC-questions

Question 1

Which of the following is least affected by outliers?

1. The range
2. The mean
3. The median
4. The standard deviation

Question 2

What are assumptions for the use of parametric tests based on the normal distribution?

1. Some properties of the data are normally distributed.
2. The samples to be tested should have approximately the same variances.
3. The data must have at least the level of interval.
4. All options are true.

Question 3

Which of the following is not a transformation that can be used to correct skewed data?

1. Log transformation
2. Square root transformation
3. Reciprocal Transformation
4. Tangent transformation

Open questions

Question 1

Where can bias occur?

Question 2

What are the four big assumptions for parametric tests?

Question 3

Why are outliers an important source of bias?

Question 4

Why is the assumption of linearity very important?

Question 5

What does normality affect?

Question 6

What does the central limit theorem say?

Question 7

What does homoscedasticity mean?

Question 8

What can a researcher look for in a small sample to investigate bias?

Question 9

What is the difference between a Q-Q plot and a P-P plot?

Question 10

How can a researcher view the assumptions of homoscedasticity and linearity at the same time?

Question 11

What is Levene's test used for?

Question 12

How can a bias be reduced?

Question 13

What are robust methods?

Question 14

What does transforming data mean?

Question 15

What is meant by the so-called selection bias?

Answer indication MC-questions

Question 1

C. The median.

Question 2

D. All options are true.

Question 3

D. Tangent transformation.

Answer indication Open questions

Question 1

Bias can occur in the estimation of the parameters, the standard error and the confidence interval, and the test statistics and p-values.

Question 2

The four main assumptions for parametric tests are that the data must be distributed normally, that the variances of the different groups must be approximately equal, that the variables are linearly related and that there is independence.

Question 3

Outliers are an important source of bias because they can drive the average up or down enormously.

Question 4

The assumption of linearity is important because the model is no longer valid if the variables do not show a linear relationship.

Question 5

Normality influences the estimates of parameters, the confidence interval, the null hypothesis significance tests and the error.

Question 6

The central limit theorem says that the distribution is always normal in large samples.

Question 7

Homoscedasticity means that all groups have approximately equal variance.

Question 8

With a small sample, you can look for skew and kurtosis to investigate bias.

Question 9

The difference between a Q-Q plot and a P-P plot is that the Q-Q plot looks at quartiles and the P-P plot at individual scores. A P-P plot shows the cumulative probability of a variable versus the cumulative probability of a given distribution.

Question 10

The assumptions of homoscedasticity and linearity can be viewed simultaneously by using a scatter plot.

Question 11

Levene's test is used to look for the homogeneity of variances.

Question 12

To reduce the bias, you can remove certain extreme scores, replace outliers with the highest score that is not an outlier, analyze your data with robust methods and transform the data.

Question 13

Robust methods are tests that are little affected by violations of assumptions.

Question 14

Transforming data means that you convert all data into different scores in the same way.

Question 15

This means that the way you select your participants can lead to a distorted picture. Consider, for example, internet surveys. You do not automatically select people who do not have internet.

What is meant by a non-parametric test? - ExamTests 7

MC-questions

Question 1

Which symbol represents the test statistic of the Mann-Whitney test?

1. Ws
2. T
3. U
4. H

Question 2

When it is assumed that all assumptions of parametric testing are met, then non-parametric testing:

1. is more conservative.
2. is less likely to accept the alternative hypothesis.
3. has less statistical power.
4. All are correct.

Question 3

A researcher measures the psychological reactions of a group of people when they watch a horror film and compares this to watching an erotic film and a documentary about wild life in Africa. The results are skewed. Which test can the researcher use to analyze the data?

1. Independent analysis of variance
2. Repeated measure analysis of variance
3. Friedman's ANOVA
4. Kruskal-Wallis test

Question 4

A researcher measures the psychological reactions of a group of people when they watch a horror movie and compares this to watching an erotic movie. The results are skewed. Which test can the researcher use to analyze the data?

1. Independent t-test
2. Wilcoxon signed-rank test
3. Dependent t-test
4. Mann-Whitney test

Open questions

Question 1

When is a non-parametric test used?

Question 2

Which tests are used as an alternative to the independent t-tests?

Question 3

How do you research the alternative tests for the independent t-tests in SPSS?

Question 4

What does the second part of the output of the non-parametric independent t-test show?

Question 5

Why is it useful to also calculate the effect size?

Question 6

What is the non-parametric variant of the dependent t-test?

Question 7

What is the non-parametric variant of the one-way independent ANOVA?

Question 8

What happens with the Kruskal-Wallis test?

Question 9

What does the Jonckhere-Terpstra test look at?

Question 10

For what can Friedman's ANOVA be used?

Question 11

What must be reported in non-parametric tests?

Answer indication MC-questions

Question 1

C. U.

Question 2

D. All are correct.

Question 3

C. Friedman's ANOVA.

Question 4

B. Wilcoxon signed-rank test.

Answer indication Open questions

Question 1

A non-parametric test is used if the assumptions cannot be met.

Question 2

The tests used as an alternative to the independent t-tests are the Wilcoxon rank-sum test and the Mann-Whitney test.

Question 3

You do this by clicking on analyze - non parametric tests - independent samples in SPSS.

Question 4

The second part of the output shows the test statistics along with the z-scores.

Question 5

It is useful to calculate the effect size as well, as this is a standard measure. The effect size allows you to compare the results with other studies.

Question 6

The non-parametric variant of the dependent t-test is the Wilcoxon signed-rank test.

Question 7

The non-parametric variant of the one-way independent ANOVA is the Kruskal-Wallis test.

Question 8

In the K-W test, a coded variable is first created from the independent variable. These can be tested for normality and homogeneity. To look at the difference between groups, these assumptions must be tested per group.

Question 9

The Jonckhere-Terpstra test looks at the pattern of the medians of groups that you want to compare with each other.

Question 10

Friedman's ANOVA can be used when the same subjects have been used in more than two different conditions.

Question 11

For non-parametric tests, the chi-square statistic, number of degrees of freedom, significance and effect size and medians of the follow-up tests should be recorded.

What is meant by the correlation between variables? - ExamTests 8

Open questions

Question 1

What is a positive relationship between two variables?

Question 2

What does the covariance indicate?

Question 3

What is the Pearson correlation coefficient?

Question 4

What is a bivariate correlation?

Question 5

How do you examine the significance of r?

Question 6

What is a bootstrap confidence interval?

Question 7

Why does correlation not automatically involve causality?

Question 8

What is the coefficient of determination?

Question 9

What is the Spearman correlation coefficient?

Question 10

When is Kendall's tau used?

Question 11

When do you use the biserial correlation coefficient?

Question 12

What is the partial correlation?

Question 13

What is the difference between semi-partial correlation and partial correlation?

Question 14

What do you use to compare correlations?

Question 15

What do you have to do to calculate the effect size?

Question 16

Which three properties of the relationship between X and Y are measured with a correlation?

Question 17

Variable x and y have a r² of ​​0.15. Is this a large, a small or an average correlation?

Question 18

What is the difference between the Pearson correlation and the multiple correlation R?

Answer indication

Question 1

A positive relationship means that an increase in one variable is associated with an increase in the other variable.

Question 2

The covariance indicates whether variables are related and whether this relationship is positive or negative.

Question 3

Pearson correlation coefficient is the standardized covariance, which is also used to measure effect size.

Question 4

A bivariate correlation is a correlation between two variables.

Question 5

You can examine the significance of the test by using a t-test with N-2 degrees of freedom.

Question 6

A bootstrap confidence interval is a confidence interval which is also accurate if the distribution is not normal.

Question 7

There is no automatic causality in correlation, because other variables also influence the correlation. Additionally, the correlation says nothing about which variable causes the change in the other variable.

Question 8

The coefficient of determination is a measure of how much variance the coded variables share.

Question 9

Spearman correlation coefficient is the non-parametric variant of the Pearson correlation.

Question 10

Kendall’s tau is used when you have a small sample with many tied scores.

Question 11

You use the biserial correlation coefficient if the variable is continuously dichotomous.

Question 12

Partial correlation is the relationship between two variables in which the effects of another variable are held constant.

Question 13

The difference between semi-partial and partial correlation is that semi-partial correlation controls for the effect that a third variable has on one of the variables in the correlation, and not on both variables as with the partial correlation.

Question 14

To compare correlations, use z-scores and t-tests.

Question 15

Correlations are effect sizes, so to calculate the effect size you do not need further steps.

Question 16

The direction of a relationship, the form of a relationship and the degree of a relationship.

Question 17

This is an average correlation.

Question 18

The multiple correlation always has a value between 0 and 1 and therefore cannot be negative. The Pearson correlation can run from -1 to 1.

What are the main characteristics of the linear model? - ExamTests 9

Open questions

Question 1

What does a researcher want to analyze when he or she performs a regression analysis?

Question 2

In regression, is the model linear or not?

Question 3

What is meant by the term residuals?

Question 4

What is the meaning of R2?

Question 5

Does a good model have a high F ratio or a low F ratio?

Question 6

What does a regression coefficient of 0 mean?

Question 7

If a researcher has a prediction about the importance of the predictors; which input method does the researcher use in SPSS?

Question 8

What is a problem with the input forward method?

Question 9

What is a problem with the step-by-step input?

Question 10

What is meant by an outlier?

Question 11

How does a researcher check if there is an outlier in x-space? In the y space? In the xy space?

Question 12

What requirements must the measurement levels of a regression analysis meet?

Question 13

What is multicollinearity? And which three problems arise with increased multicollinearity?

Question 14

What is homoscedasticity?

Question 15

What does a researcher test with the Durbin-Watson test?

Question 16

What is cross validation? And by which two methods can this be done?

Question 17

What does the confidence interval indicate?

Question 18

What is explained variance?

Question 19

What are dummy variables?

Question 20

What does a mediating variable do?

Question 21

What are the assumptions for multiple regression?

Answer indication

Question 1

A researcher wants to predict an outcome variable Y by means of a single predictor (simple regression) or by several predictors (multiple regression).

Question 2

Linear, because with a good model a relationship is found between the variables.

Question 3

A residue is the vertical distance between the actual data and the regression line. These differences can be both positive and negative.

Question 4

This is the amount of variance explained by the model in relation to the total variance. It represents the percentage of variance in the outcome that can be explained by the model.

Question 5

A high F ratio. This means that the improvement in prediction is large (MSm) and the difference between the model and the observed data is small (MSr).

Question 6

This means that a change in the predictor variable results in no change in the outcome variable.

Question 7

Hierarchical regression.

Question 8

There may be supressor effects. This means that a predictor is significant, but only if another predictor is kept constant. As a result, when using the forward method, there is a greater chance of a type II error (eliminating a predictor while it was significant).

Question 9

There may be a chance of overfitting (too many variables with only a small variance) or underfitting (not adding important variables).

Question 10

An outlier is a measurement that differs substantially from the average trend of the data.

Question 11

Respectively with mahalanobis distance, standardized residues and with the Cook's distance.

Question 12

The dependent variable must at least be ratio or interval measurement level and the independent variable of ratio or interval or be recoded into a dummy.

Question 13

This occurs when there is a strong correlation between two or more predictors. The problems that arise are: unreliable b's, limits the size of r and more difficult to distinguish in importance between predictors.

Question 14

The variance of the residuals must be constant at each level of the predictor variable, if this variance is unequal, we speak of heteroscedasticity.

Question 15

With this, a researcher tests whether there are independent errors, the rule of thumb is that the value must be between 1 and 3.

Question 16

If it is not clear how accurately the sample describes the population, the extent to which the model predicts the outcome in other samples can be examined. This can be done by means of the adjusted R2 and data splitting.

Question 17

The confidence interval indicates the interval where the true value of b resides in the population.

Question 18

That part of the variance that is determined and caused by the independent variables and not by external factors.

Question 19

This is a variable that is of nominal measurement level with more than 2 categories that has been converted to a dichotomous item with only 0 or 1 as a score. Dummy variables are multiple variables coded with only zeros and ones.

Question 20

The effect of the predictor on the dependent variable proceeds via a third variable, which can be partial or complete.

Question 21

The effect of the predictor on the dependent variable then depends on a third variable.

How can two means be compared? - ExamTests 10

Open questions

Question 1

What are the ways to compare two groups with two averages?

Question 2

What does a researcher use a t-test for?

Question 3

When does a researcher use the dependent t-test?

Question 4

What does the law of the sum variance say?

Question 5

What is the standard measurement error?

Question 6

When does an effect not come by chance?

Question 7

Is the systematic variance greater or less than the non-systematic variance if an experimental condition has an effect?

Question 8

Which scores should be normally distributed in a dependent t-test?

Question 9

What is Levene's test used for?

Question 10

Why is Cohen's d sometimes preferred as a measure of effect size?

Question 11

Why is a researcher more likely to find a significant effect in repeated measurement design?

Answer indication

Question 1

Ways to compare two groups with two means are by exposing two groups of participants to different manipulations or by exposing a group of participants several times to different manipulations.

Question 2

You use a t-test to see whether the difference between the group means significantly deviate from 0.

Question 3

The dependent t-test is used in two experimental conditions in which the same subjects participate in both conditions.

Question 4

The law of the variance sum says that the variance of the difference between two independent variables is equal to the sum of the variances.

Question 5

The standard error of measurement is the standard deviation of the sample distribution.

Question 6

An effect is not a coincidence when the mean difference between the samples and population is large and the standard measurement error is small.

Question 7

The systematic variance is greater than the non-systematic variance if an experimental condition has an effect.

Question 8

In a dependent t-test, the differences between scores must be normally distributed.

Question 9

Levene's test is used to see if the variances are different and if there is no homoscedasticity.

Question 10

Cohen's d is sometimes preferred as a measure of effect size, because it has an effect size that is independent of the design.

Question 11

A researcher is more likely to find a significant effect with a repeated measurement design, because the non-systematic variance is much smaller than with an intermediate group design.

What is meant by moderation, mediation and multicategory predictors? - ExamTests 11

Open questions

Question 1

What affects the moderator?

Question 2

What does centering mean?

Question 3

What does a researcher have to do to find the effect of the moderator?

Question 4

When is it called a mediation?

Question 5

How come there is a direct and an indirect effect in mediation?

Question 6

Which three regression models is mediation based on?

Question 7

How can the effect size be calculated in mediation?

Question 8

What is a moderating effect?

Answer indication

Question 1

The moderator influences the relationship between a predictor and the outcome.

Question 2

Centering means that you transform a variable into deviations around a certain point.

Question 3

To find the effect of the moderator you have to perform a simple slope analysis.

Question 4

Mediation means that the relationship between a predictor variable and the outcome is explained by the relationship with a third variable, the mediator.

Question 5

The direct effect is the relationship between the predictor and the outcome, the indirect effect is the effect of the predictor on the outcome via the mediator.

Question 6

Mediation is based on regression that:

1. predicts the outcome from the predictor
2. the mediator predicts from the predictor
3. predicts the outcome from both the predictor and the mediator

Question 7

In mediation you can calculate the effect size by looking at the combined effect of a and b, by looking at the magnitude of the indirect effect compared to the total effect of the predictor, or by calculating R2.

Question 8

A moderating effect refers to the effect of a moderator that influences the relationship between the predictor variable and the outcome variable.

How are different independent means compared with each other? - ExamTests 12

Open questions

Question 1

When does a researcher perform an ANOVA?

Question 2

Why are t-tests not used for this?

Question 3

What are contrasts?

Question 4

What is meant by a significant F ratio?

Question 5

What is meant by a type I error? And what is meant by a type II error?

Question 6

What is the power of a test?

Question 7

Which procedures does a researcher use to see which groups actually differ from each other?

Question 8

What test is performed if assumptions are violated and ANOVA cannot be performed?

Question 9

What is the value of the type I error if three tests are compared with each other by a t-test?

Answer indication

Question 1

When researchers want to compare more than two conditions, the ANOVA is used.

Question 2

Because separate t-tests increase the chance of a type I error. This is called chance capitalization.

Question 3

Contrasts are alternative coding schemes.

Question 4

A significant F ratio shows that the means of the groups are different.

Question 5

A type I error occurs when an effect is detected while it is not present. In this case, H0 is incorrectly rejected. A type II error occurs when an effect is not detected while it is present. In this case, H1 is incorrectly rejected.

Question 6

The ability of a test to detect an effect of a certain size (.80 is good).

Question 7

Post-hoc procedures.

Question 8

The Kruskall-Wallis test.

Question 9

The type I error per test is 0.95 (5% chance of incorrectly rejecting the null hypothesis). So the answer is: 0.95 x 0.95 x 0.95 = 0.857 so 1 - 0.857 = 14.3% is the correct percentage.

What is meant by an ANCOVA? - ExamTests 13

Open questions

Question 1

When does an investigator decide to conduct an ANCOVA?

Question 2

What are covariates?

Question 3

What are the two reasons for including a covariate in the analysis?

Question 4

ANCOVA has the same assumptions as ANOVA but what are the two additions?

Question 5

At ANCOVA we have the partial eta squared. What is this?

Question 6

When can the omega squared (ω²) be used in ANCOVA?

Question 7

How can the effect size of the contrasts be calculated with ANCOVA?

Answer indication

Question 1

If an ANOVA contains a continuous variable that also has an effect on the dependent variable, it can be included in the analysis as a covariate.

Question 2

Continuous variables that also predict the outcome variable.

Question 3

With this you can be the first to reduce the intragroup variance, by including the covariate we can explain more of the variance and less unexplained variance remains. In addition, confounds can be eliminated (third variable).

Question 4

The two additions are:

• independent relationship between the covariate and the dependent variable
• homogeneity of regression coefficients

Question 5

The partial eta squared is the amount of variance in the dependent variable that is shared by the independent variable but not explained by the covariate.

Question 6

ω² can be used with equal group sizes.

Question 7

The effect size of the contrasts can be calculated with a t-test.

What is meant by a factorial design? - ExamTests 14

Open questions

Question 1

When do you perform a factorial ANOVA?

Question 2

What is another name for a factorial ANOVA?

Question 3

What three types of factorial designs are there?

Question 4

What is the difference between these designs?

Question 5

What does the residual sum show?

Question 6

Do the same assumptions apply to the factor ANOVA as the other linear models?

Question 7

How does SPSS calculate the effect size and in what two other ways can the effect size be calculated?

Question 8

How should the effects be reported?

Answer indication

Question 1

You perform the factorial ANOVA when you look at two or more independent variables.

Question 2

A two-way ANOVA.

Question 3

Independent factorial design, repeated measures factorial design and mixed design.

Question 4

Independent factors concern an intermediate group design, repeated measures an inner group design and mixed design makes a combination of this.

Question 5

The residual sum shows the unexplained variance.

Question 6

Yes, the same assumptions apply to the factor ANOVA as to the other linear models.

Question 7

SPSS gives partial eta squared as the effect size, but it is more sensible to use ω² or the simple effects analysis r.

Question 8

When reporting the results, the same should be displayed as with a regular ANOVA. The F ratio and the number of degrees of freedom must be stated in each case, for the two main effects and for the interaction effect.

What is meant by a repeated-measures design? - ExamTests 15

MC-questions

Question 1

What is a disadvantage of repeated measurements?

1. The assumption that the different groups are independent cannot be satisfied.
2. The assumption that the different conditions are independent cannot be satisfied.
3. The assumption that the different groups are dependent cannot be satisfied.
4. The assumption that the different groups are dependent cannot be satisfied.

Question 2

Which assumption for repeated measurements is made by the decline of the assumption made in question 1?

1. The assumption of composite symmetry.
2. The assumption of sphericity.
3. The assumption of normality.
4. The assumption of homogeneity.

Question 3

Which test in SPSS tests whether the variances between the pairs of conditions (assumption of sphericity) are equal?

1. The Greenhouse-Geisser test.
2. The Pillai-Bartlett test.
3. The Mauchly's test.
4. The Huynh-Feldt test.

Question 4

What options are there when the sphericity assumption is not met?

1. Performing a MANOVA.
2. Carrying out a Greenhouse-Geisser correction.
3. Perform a Huynh-Feldt correction.
4. Conduct an ANCOVA.

Question 5

Which statement is correct?

I. For repeated measurements, the effect size ω2 can be used, it can be calculated in the same way as for the independent ANOVA.

II. The calculation of the total variance (SSt) is done in the same way with repeated measurements as with a one-way independent ANOVA.

1. Statement I and II are correct.
2. Statement I is correct, statement II is incorrect.
3. Statement I is incorrect, statement II is correct.
4. Both statements are incorrect.

Open questions

Question 1

What does a repeated measures design entail?

Question 2

Which assumption is only violated by this?

Question 3

What is sphericity? Which test in SPSS do you use to check this assumption?

Answer indication MC-questions

Question 1

B. The assumption that the different conditions are independent cannot be satisfied. A subject undergoes all conditions and thus the conditions are related to each other.

Question 2

B. The assumption of sphericity. It means that the relationship between pairs of experimental conditions is the same. This is calculated by the variances of treatment level pairs.

Question 3

C. The Mauchly's test. When this test shows a significant result, the assumption of sphericity is not met.

Question 4

D. Conducting ANCOVA. The Greenhouse-Geisser provides the correction of the variances in degrees of freedom. When this correction gives a high value, a Huynh-Feldt correction can be used.

Question 5

C. The effect size for repeated measurements cannot be calculated in the same way as with an independent ANOVA.

Answer indication Open questions

Question 1

In repeated measures, the participants participate in all conditions of the study. This allows checking for individual differences.

Question 2

The assumption of independent groups is thereby violated because there is a relationship between the scores of the different conditions.

Question 3

This is the assumption that the relationships between the pairs of the experimental conditions are equal. This is only important with at least three conditions. In SPSS you use the Mauchly's test to check this assumption.

What are mixed designs? - ExamTests 16

Open questions

Question 1

What does a mixed design entail?

Question 2

What is a repeated measurement and what is a disadvantage?

Question 3

What is the assumption of sphericity?

Question 4

What is the difference between sphericity and compound symmetry?

Question 5

What is a good substitute for the post hoc test?

Question 6

What kind of effect size do we measure with repeated measurements?

Question 7

How do you report the results of a repeated measurement?

Question 8

What is the difference between a main effect and an interaction?

Answer indication

Question 1

With a mixed design, you look at both between and within group variables, so there are always more than two IVs. Participants in different groups are examined (within group variance) and the differences between two groups (between group variance).

Question 2

A disadvantage of a repeated measurement design is that the assumption that the different conditions are independent cannot be met.

Question 3

That the relationship between pairs of experimental conditions is the same and the dependence between experimental conditions is about the same.

Question 4

Sphericity is a bit more general than compound symmetry. This means that both variances between the conditions are the same and that the covariances between the pairs are also the same. Sphericity is less strict than composite symmetry, it concerns the difference between the conditions, those variances must be approximately equal: Greenhouse-Geisser correction, Huynh-Feldt correction or the MANOVA.

Question 5

The Bonferroni test.

Question 6

ω².

Question 7

Repeated measurements report the same data as an independent ANOVA. For repeated measurements, the number of corrected degrees of freedom must also be stated if the assumption of sphericity is not met. The multivariate tests can also be mentioned.

Question 8

A main effect is a unique effect of an independent variable on the dependent variable. It can be an effect of a standalone strategy or charisma. An interaction looks at the combined effect of two or more variables.

What is meant by a MANOVA? - ExamTests 17

MC-questions

Question 1

When can a MANOVA be used?

1. When there are multiple independent variables.
2. If there are multiple dependent variables.
3. When large groups are used.
4. When dependent groups are used.

Question 2

Which statement is correct?
I. The disadvantage of using multiple ANOVAs is a greater measurement error and a greater chance of type I errors.
II. The disadvantage of using multiple ANOVAs is that the relationship between the dependent variables is not considered.

1. Statements I and II are correct.
2. Statement I is correct, statement II is incorrect.
3. Statement I is incorrect, statement II is correct.
4. Both statements are incorrect.

Question 3

When is a matrix called an identity matrix?

1. When the diagonal components together are 0 and the non-diagonal components together are 1.
2. When the vertical components together are 1 and the non-vertical components together are 0.
3. When the vertical components together are 0 and the non-vertical components together are 1.
4. When the diagonal components together are 1 and the non-diagonal components together are 0.

Question 4

What is compared when calculating a MANOVA?

1. The systematic measurement error is compared to the non-systematic measurement error of several dependent variables.
2. The systematic measurement error is compared with the non-systematic measurement error of the independent variable (s).
3. The systematic variance is compared to the non-systematic variance of several dependent variables.
4. The systematic variance is compared to the non-systematic variance of the independent variable(s).

Question 5

Which term belongs to the following definition: "the total value of the measurement errors between two variables"?

1. Sum of squares.
2. Discriminant Function.
3. Cross product.
4. Pillai-Bartlett trace.

Question 6

What is not a characteristic of a varia?

1. The number of variables is always less than the number of dependent variables.
2. The variations are orthogonal.
3. The variates are correlated.
4. The variates are uncorrelated.

Question 7

How can no effect size be calculated?

1. Pillai-Bartlett trace.
2. Wilks' lambda.
3. Roy's largest root.
4. Box's test.

Question 8

When is it not smart to do a MANOVA?

1. When a set of dependent variables are very highly correlated.
2. If you want to investigate multiple categories of the independent variable.
3. When the dependent variables are averagely correlated.
4. If you want to identify which dependent variables cause the most group difference.

Open questions

Question 1

What is a MANOVA?

Question 2

Why do we perform a MANOVA and not single ANOVAs?

Question 3

What is the advantage of performing a MANOVA?

Question 4

Which additional assumptions apply to the MANOVA?

Answer indication MC-questions

Question 1

B. When there are multiple dependent variables. It does not matter whether there are one or more independent variables.

Question 2

A. Statements I and II are correct.

Question 3

D. When the diagonal components together are 1 and the non-diagonal components together are 0.

Question 4

C. The systematic variance is compared to the non-systematic variance of several dependent variables. A MANOVA can only be used in a situation with several dependent variables. The result of this comparison is a matrix of many variances and covariances.

Question 5

C. Cross product. There are different types of cross products: total cross product (CPt), model cross product (CPm) and residue cross product (CPr). The cross product model looks at how the relationship between two dependent variables has been affected by the experimental manipulation. The residual cross product looks at how the relationship of the two dependent variables is affected by individual differences.

Question 6

C. The variates are correlated. Variates are orthogonal, which means that they are not correlated.

Question 7

D. Box's test. In the other three ways, the eigenvalues ​​are always used to determine the effect sizes.

Question 8

A. When a set of dependent variables are very highly correlated.

Answer indication Open questions

Question 1

A MANOVA is a test in which we look at the effect of several dependent variables.

Question 2

If we were to perform individual ANOVAs, we would speak of probability capitalization (increase of the type I error).

Question 3

The advantage is that possible relationships between dependent variables can be demonstrated and this gives the MANOVA more power to detect an effect. However, the MANOVA can only be used if there is a theoretical foundation for it.

Question 4

The added assumptions are that the SPs are normally distributed within the groups and that the correlation between the SPs is the same in all groups.

What are the main characteristics of an exploratory factor analysis? - ExamTests 18

MC-questions

Question 1

A Cronbach's alpha of 0.7 / 0.8 or higher always shows reliability.

1. Yes
2. No
3. You cannot say anything about this because there is too little information available
4. You cannot say anything about this because you need a different formula

Question 2

Which factor analyzes below are descriptive methods and cannot be generalized to the population?

1. Principal component analysis
2. Image covariance analysis
3. Kaiser's Alpha Factoring
4. Maximum likelihood method

Question 3

Varimax rotations can be used when:

1. One researcher thinks the underlying factors are correlated.
2. A researcher thinks the underlying factors are non-orthogonal.
3. A researcher thinks the underlying factors are independent.
4. Kaiser's criteria are met.

Question 4

A Cronbach's alpha of 0.85 for a questionnaire means that:

1. The questionnaire is valid.
2. The questionnaire has a high reliability.
3. The questionnaire has too few items.
4. The questionnaire displays different scores when administered to the same people at two different times.

Question 5

What is meant by a latent variable?

1. A variable that cannot be measured directly.
2. It is another name for a factor
3. These variables represent cluster variables that are highly correlated with each other.
4. All of the above answers are correct.

Open questions

Question 1

What does the square of a factor load tell us?

Question 2

What is meant by the term cross-product?

Answer indication MC-questions

Question 1

B. No, because a large always items also increases the Cronbach's alpha, this does not have to say anything about reliability.

Question 2

A and B are descriptive methods (only applicable to the sample. C and D are techniques to generalize to the population.

Question 3

C. A researcher thinks the underlying factors are independent.

Question 4

B. The questionnaire has good reliability.

Question 5

D. All of the above answers are correct.

Answer indication Open questions

Question 1

The factor load says something about the coordinate of a variable with respect to two (or more) factors.

Question 2

A cross-product represents the total value for a combined error between two variables. It represents a non-standardized estimate of the total correlation between two variables.

How can categorical variables be analyzed? - ExamTests 19

Open questions

Question 1

What is categorical data?

Question 2

What can you calculate with Pearson's chi-square test?

Question 3

What can you use with small samples?

Question 4

What do you do when you look at the likelihood ratio?

Question 5

The likelihood of what error does Yates's continuity correction reduce?

Question 6

What is a loglinear analysis?

Question 7

When do we speak of a saturated model?

Question 8

When do you keep the simple new model in log-linear analysis?

Question 9

What does the assumption of independence mean with chi-square?

Question 10

What is the second assumption of chi-square?

Question 11

What are the four options if the assumptions are violated in loglinear analysis?

Question 12

What does a researcher report when he or she uses a chi-square?

Question 13

What does a researcher report on loglinear analysis?

Answer indication

Question 1

Categorical data are data where the outcome variable consists of different categories.

Question 2

With Pearson's chi-square test you can see if there is a relationship between two categorical variables.

Question 3

You can use Fisher's exact test for small sample sizes.

Question 4

The likelihood ratio means that you create a model where the probability of obtaining the observed data is maximal, and you compare this model with the probability of obtaining the observed data if the null hypothesis is true.

Question 5

Yates' continuity correction reduces the risk of type I error.

Question 6

Loglinear analysis is an analysis for when you have more than 2 categorical variables.

Question 7

A saturated model is when the standard deviations are all 0.

Question 8

In loglinear analysis you keep the simple new model if the simpler model is not very different from the complex one.

Question 9

The assumption of independence in chi-square means that a person cannot fall into multiple categories.

Question 10

The second assumption is that with a 2x2 contingency table, the expected frequencies in each cell must be greater than 5.

Question 11

The four options are:

• voiding a variable
• cancelling one of the categories
• collect more data
• accepting loss of power

Question 12

At chi-square you report the statistic, the significance and the number of degrees of freedom. You also display the contingency table.

Question 13

In loglinear analysis you state the likelihood ratio statistic, change of the chi-square and possibly z-scores with corresponding confidence intervals.

What is meant by a loglinear analysis? - ExamTests 20

Open questions

Question 1

What is logistic regression?

Question 2

When does multinomial logistic regression occur?

Question 3

How can you make the non-linear relationship linear?

Question 4

What is the R-statistic?

Question 5

What is the Wald statistic?

Question 6

What does the odds ratio represent?

Question 7

When do suppressive effects occur?

Question 8

What does striving for parsimony mean?

Question 9

What is the assumption of linearity that must be met?

Question 10

When does complete separation occur?

Question 11

When is there overdisperision?

Question 12

What do you report with logistic regression?

Question 13

What do you have to do to test the assumption of linearity in logistic regression?

Question 14

How can you test multicollinearity?

Question 15

What do you use in multinomial regression to predict group membership of more than two categories?

Answer indication

Question 1

Logistic regression is multiple regression where the outcome variable is categorical and the predictor variables are continuous or categorical.

Question 2

Multinomial logistic regression is when there are more than two categories.

Question 3

You make the linear relationship by using a logarithmic transformation.

Question 4

The R statistic is the partial correlation between the outcome variable and each of the predictor variables.

Question 5

The Wald statistic is the z-statistic used to determine the contribution of the predictors to the model.

Question 6

The odds ratio indicates the change in weather risk that comes from the change in the predictor.

Question 7

Suppressive effects occur when a predictor has a significant effect but only when another variable is held constant.

Question 8

Striving for parsimony means that a simple explanation of a phenomenon is preferable to a complicated explanation.

Question 9

The assumption implies that there must be a linear relationship between the continuous predictors and the logit of the outcome variable.

Question 10

Complete separation occurs when the outcome variable is perfectly predicted by a predictor or a combination of predictors.

Question 11

Overdispersion occurs when the observed variable is larger than the expected variable from a logistic regression model.

Question 12

With logistic regression you report the b-values, the standard measurement errors and the significance.

Question 13

You run the logistic regression analysis again, but add predictors that are the interaction of each predictor and the log of itself.

Question 14

To test multicollinearity, use the linear regression analysis.

Question 15

In multinomial logistic regression, you also use logistic regression to predict group membership of more than two categories.

What is meant by multilevel linear models? - ExamTests 21

Open questions

Question 1

What are the benefits of a multilevel linear model?

Question 2

What is the difference between a fixed effect and a random effect?

Question 3

What is distinguished between in multilevel models?

Question 4

When is there usually also a random intercept?

Question 5

What is the difference between a multilevel linear model and regression?

Question 6

Why is it advisable to start with a model with only fixed parameters and add random coefficients if necessary?

Question 7

What are the four covariance structures?

Question 8

What is a special assumption for a multilevel model?

Question 9

What does centering a variable mean?

Question 10

What are polynomials?

Question 11

What is important in growth curves?

Question 12

Why are there no fixed guidelines for reporting a multilevel model?

Answer indication

Question 1

The advantages of a multilevel linear model are that it can be used to measure variation between slope coefficients, that the data does not have to meet the assumption of independence, and that the model can be used well if there are missing values.

Question 2

A fixed effect is an effect where all possible conditions in which a researcher is interested are present in the experiment. A random effect is when the experiment contains only a random sample of possible conditions.

Question 3

In multilevel models, a distinction is made between fixed coefficients and random coefficients.

Question 4

Usually, there is also a random intercept if there is a random slope.

Question 5

In regression there is a fixed intercept and regression coefficient.

Question 6

This is useful because in this way you can compare the fit of the new model with the basic model.

Question 7

The four covariance structures are the variance component structure, the diagonal structure, the AR (1) and the unstructured structure.

Question 8

For a multilevel model it also applies that the random intercepts and slopes must be normally distributed over the model.

Question 9

Centering a variable means that the variable is transformed into deviations from a fixed point, usually the mean or group mean.

Question 10

Polynomials are growth curves, or lines with a trend.

Question 11

Important for growth curves is that you can apply polynominals to one less than the number of time points you have, and that a polynomial is a simple power function.

Question 12

There are no fixed guidelines, because multilevel models can take many different forms.