Exams 2003

Exam 1

 

  1. What is a probability?  Give an example to illustrate the concept.
  2. Define the mean, median and mode.
  3. Define bias as a statistical concept.  Why is bias important?
  4. Describe the sampling distribution of the mean when data are drawn from a population that is normally distributed.
  5. What two parameters drive the normal curve?
  6. Suppose we sampled 100 people at random from USF and measured their interest in statistics.  The mean turns out to be 4 on a 10-point scale and the SD is 1.  Where do you think the population (USF) mean is likely to be (compute an approximate 95% confidence interval; show your work).
  7. What are the null and alternative hypotheses?  How do we use them? 
  8. Suppose .  Further suppose N =25, the distribution of data is normal and .  Draw a diagram that specifies the rejection region.
  9. Describe the alpha and beta errors in statistical significance testing.
  10. Suppose I’m doing a study for an orange juice company.  They want to know if their juice tastes better or worse than the leading brand (say Tropicana).  So I’m going to do a little blind taste-testing study to find out.  What might I do to increase or boost the power of my study?

 

Exam 2

 

  1. What is a treatment effect in ANOVA?  Write the equation and explain terms.
  2. What is the F distribution?  What is its use in ANOVA?
  3. What is a mean square?  What does it estimate?
  4. Suppose we have 3 groups and 15 people per group.  Our SS between is 20 and our SS within is 42.  Construct an F table to display the results and compute F for the ANOVA.  Although you don’t have the critical value of F, does the obtained F appear to be significant?
  5. Describe why you might want to estimate the power of an F test for an ANOVA design before you collect any data.
  6. Describe a concrete example (name & describe your variables) of a randomized block design.  You should have 1 factor as the blocking factor and the other factor as the main factor of interest.  Explain why the blocking factor reasonable.
  7. Describe the sampling distribution of the correlation coefficient.  Mention bias, the variance of the sampling distribution, and the tails (skew) of the sampling distribution of r.
  8. Describe a situation in which it would make sense to use this test for the equality of correlations: [Hint: this is a test of the equality of independent correlations.]

 

 

 

 

 

9.  How do we find the slope and intercept for the regression line with a single independent variable? (Either formula for the slope is acceptable.)

10.  In regression, what is a residual?  What is the relation between the residuals and scores on the predictor? [Hint:  recall the equation for the linear model.]

 

 

Exam 3

 

  1. Sketch and label a sampling distribution for a statistic.  What is a sampling distribution?  How does it relate to a significance test?
  2. In statistics, what are the decisions errors alpha and beta?  Briefly define each.
  3. Write a linear equation for analysis of covariance in which there is one continuous independent variable and one categorical variable. Describe the meaning of each of the b-weights in the model.
  4. Describe concrete example (names of variables, null and alternative hypotheses in terms of the research question) in which it would make sense to use an independent samples t-test.
  5. What are the assumptions of homogeneity of variance and independence of error in ANOVA?  What problem(s) might arise if the assumptions turn out to be false?
  6. Describe a concrete example (names of variables, research question) in which it would make sense to use a test for independent correlations.
  7. In what sense does the partial correlation ‘hold constant’ some third variable?  How is statistical control (rather than experimental control) achieved?
  8. What is the difference between a partial and a semipartial correlation?  Draw a diagram to illustrate your answer.
  9. Describe cross-validation in multiple regression. How do you do it and what does it accomplish?
  10. What is collinearity?  Why is it a problem for multiple regression?  Describe 2 possible remedies if you find collinearity to be a problem in your data.