All RegNOVA Exam Questions 2004

  1. Describe the mean and standard deviation of a distribution of data.  What do they tell us about a distribution of data?
  2. What does a z score of 3 mean?
  3. What is the principle of maximum likelihood?  If my data were 1, 2, 3, why would I prefer an estimated population mean of 2 over 1?
  4. In scientific work, we should quantify the magnitude of error of our study results, that is, we should say by how much we are likely to be wrong.  How does the standard error of the mean help us to do this?
  5. What does the Central Limit Theorem tell us?  Why is it helpful for setting confidence intervals for the mean?
  6. Why is the number 1.96 so widely used in setting confidence intervals?
  7. Draw a diagram to illustrate the rejection region.  Show what happens to the rejection region when you change alpha.
  8. Draw a diagram to illustrate both the null hypothesis and an alternative hypothesis.  Indicate the regions of the curve corresponding to the alternative hypothesis that show beta and power.
  9. What is the standard error of the difference in means for the independent samples t-test (that is, conceptually, how is it related to a sampling distribution and what the heck do you do with it)?  What are the things that influence its magnitude?
  10. Describe a concrete situation in which you might want to use a single sample t-test.  Why is the single sample t-test the right test to use in this situation?

 

  1. According to the General Linear Model (ANOVA model), what is an error?  How is an error computed in ANOVA?
  2. Use a concrete example of a fixed-effects and a random-effects experiment to explain the difference in the two types of models (fixed- vs. random-effects).  Note: you don’t have to write equations or calculations for this, just state the primary conceptual differences and illustrate concretely.
  3. What are the three main assumptions of ANOVA?  Briefly describe each.
  4. What is the null hypothesis in ANOVA?
  5. What are planned comparisons and post hoc tests?  When do we use each?
  6. What are the main issues in the choice of post hoc tests?  That is, why might somebody prefer one post hoc test to another?
  7. Describe a concrete example of a two-factor ANOVA study in which you expect to find a significant interaction.  Explain why the interaction should be significant.
  8. What affects the power of the test of the correlation coefficient spit out by the computer when the null is that ?  (Hint:  what determines the sampling variance of r?)
  9. What does the sign of the correlation coefficient tell us?  What does the absolute value of the correlation coefficient tell us?
  10. Describe a concrete, correlational study in which you would experience range restriction.  In that study, what would be the effect of range restriction on the study’s outcome?

 

  1. What is a sampling distribution?  Why is it important in testing hypotheses about values of parameters (e.g., whether a b weight has a certain value)?
  2. Draw a diagram to illustrate both the null hypothesis and an alternative hypothesis.  Indicate the regions of the curve corresponding to the alternative hypothesis that show beta and power.
  3. Describe a concrete example in which it would make sense to compute a dependent samples t-test.  Explain why the dependent t is the appropriate test to use for your example.
  4. Describe one advantage and one disadvantage of a repeated measures ANOVA design.
  5. How do changes in the slope and intercept affect (move) the regression line? Draw a diagram to illustrate your answer.

 

  1. Suppose I use a test battery to predict freshman GPA for college students.  My R-square value is .10.  What does this value of R-square mean?  Supposing that the value is statistically significant, what is its interpretation?

 

  1. What is collinearity?  Why is it a problem?  Describe one collinearity diagnostic and how it might be used.

 

 

 

  1. What might you do if you were worried that the assumptions of linearity and homoscedasticity were not met for data you wanted to analyze.  That is, how would you go about deciding whether linearity and homoscedasticity were problems?
  2. Why is the squared semipartial correlation always less than or equal to the squared partial?
  3. Suppose you have data relating scores on extroversion to sales performance (dollars) for a group of male and female furniture sales people; your independent variables are a) scores on extroversion (a personality variable scaled M=50 and SD=10) and b) sex (m = 1 and f = 2). How would you analyze these data?  Describe the steps and how you would interpret the results.