# Statistics Quiz #1

This exam consists of ten multiple choice questions. For each, circle the letter corresponding to the best answer.

**This is a closed book/laptop test.**You are not permitted to look at any papers, books, or other material during this test.- The standard deviation for a set of scores gets larger when:
- you increase the number of scores
- you add a constant to all of the scores
- you add a few scores that are unusually far from the mean
- the mean gets larger
- all of the above

- The standard error of the mean gets smaller when:
- the size of the samples gets smaller
- the size of the samples gets larger
- the size of the mean gets smaller
- the number of samples gets larger
- the number of samples gets smaller

- The 99% confidence interval based on a single sample mean:
- always contains the population mean
- is always smaller than the 95% confidence interval
- always contains the sample mean
- contains the sample mean 99% of the time
- contains the population mean 95% of the time

- Suppose that you read about a two-group experiment, in which the t value turned out to be 19. Given no other information, which of the following can you be sure of?
- the sample sizes must have been large
- the effect size must have been large
- alpha must have been large
- the p value must have been small
- all of the above

- Imagine that you have read a journal article in which a correlation of -.09 was reported as statistically significant at the .05 level. Which of the following must be true?
- The sample was very large.
- The correlation in the population was large in magnitude.
- The relationship between the two variables was curvilinear.
- A Type I error was committed.
- A calculation error was made.

- Which of the following would tend to increase the power of a matched t-test?
- producing a greater separation of the population means
- increasing the size of the two matched samples
- increasing the (positive) correlation between the two sets of scores
- using a larger alpha (e.g., .1)
- all of the above

- Multiple comparison methods, like Fisher's protected t-tests and Tukey's HSD test, are used to:
- keep the rate of Type I errors in a complex experiment from getting too high
- keep the rate of Type II errors in a complex experiment from getting too high
- improve the power for detecting effects in a complex experiment
- obtain significant results when the initial complex tests fail to reject the null hypothesis
- all of the above

- If you get a significant F value at the .05 level when performing a three-group ANOVA, you can be sure that:
- Each pair of sample means will differ significantly from each other.
- Each sample mean will differ significantly from the other two.
- The critical F value is larger than the calculated F value.
- A Type I error could not have been made.
- A Type II error could not have been made.

- If the F you calculated for a five-group ANOVA turned out to be exactly zero, which of the following must be true?
- all of the sample means are the same
- all of the sample variances are the same
- all of the sample variances are zero
- all of the population means are the same
- you made a mistake -- this is not possible

- Imagine a 2 X 2 ANOVA in which drug vs. placebo is one factor and gender is the other. Which of the conditions regarding the simple main effects of the drug (i.e., the effect of the drug for each gender separately) will guarantee that the interaction will be significant?
- The drug effect is significant for one of the genders, but not the other.
- The drug effect is significant and in the same direction for both genders.
- The drug effect goes in opposite directions for the two genders.
- The drug effect for one of the genders differs significantly from the drug effect for the other one.
- Any of the above.