1. If you are running a statistical test, but you don't know what kind of variables you have, which of these is equivalent to the null hypothesis? Note that only one answer is correct.

• The value for r is zero
• All the means are the same
• This is not a useful predictor
• The value of r-squared is zero
• The p-value is smaller than alpha
• There are very few degrees of freedom

2) Why do we like large sample sizes? Choose all that apply.

• Makes p-values smaller (assuming we should reject the null)
• Makes SE smaller (in cases where we calculate SE)
• Might be necessary to make the chi-square test legal
• Makes r-squared better
• Less likely to be biased
• Makes confidence intervals smaller

3) I look at 1000 St Thomas graduates, to see if there's a relationship between what month they were born in (12 choices) and what major they chose (25 choices). How many df would this test have?

• Your answer

Which of these are reasons to choose not to do a regression analysis? Choose all that apply.

• Scatterplot looks curve
• P-value is high
• Some serious outliers
• Too many data points
• Negative r value
• One of the variables is ZIP code

How do we prove that the null hypothesis is correct?

• We get a p-value below alpha
• We get a p-value above alpha
• We can't

Which of these measures are resistant?

• Mean
• Median
• Variance
• IQR
• t

Which is larger, a 95% confidence interval, or a 99% confidence interval on the same sample?

• 95%
• 99%
• They are the same size
• It depends on another factor

Which software did you use in your section of the lab, and do you feel like you learned that software?

• Your answer

I want to know who spends more money on Halloween, people in the city or people in the suburbs. What statistical test should I run?

• Regression
• ANOVA
• Test of Association
• Goodness of Fit
• Two sample t for Means

There are four different statistical software packages used in 220. Are they equally likely to be chosen by students? Which test can we do to examine this question?

• Regression
• ANOVA
• Test of Association
• Goodness of Fit
• Two sample t for Means

I do an ANOVA to look for a relationship between your birth month and your height (in inches). What is your value for x? Please answer like we did on the practice exams.

5 points

• Your answer

If you run a two-proportion test, and the p-value is 0.28, what is your conclusion? Include the statement of the appropriate hypothesis.

4 points

• Your answer

If I'm looking for whether there is a relationship between a college student's major and their height, what statistical test should we do?

• Regression
• ANOVA
• Test of Association
• Goodness of Fit
• Two sample t for Means

I did a two-sample t-test. Would the one-sided or the two-sided test have a higher p-value?

• One-sided
• Two-sided
• It's a tie
• Need more information.

What's the difference between an outlier and an influential point on a scatterplot?

4 points

• Your answer

Which of these cases does not have a model that we used in class? These describe the variables involved.

• Numerical predicting numerical
• Numerical predicting categorical
• Categorical predicting numerical
• Categorical predicting categorical

Which of these are test statistics that can never be negative?

• z
• t
• F
• p
• r-squared
• chi-square
• alpha

If we make a model using multiple predictors, the most useful of those predictors is the one with the largest...

• p-value
• SS
• MS
• df
• F

What is always in a confidence interval?

• The population mean
• The population parameter
• p-hat
• The sample statistic

Have we talked about reasonable models where both p-value and r-squared could be lower than 0.05?

• Yes
• No

How would you analysis 1-sided two-sample t question? Give an example.

• Your answer

If we calculate a z value of 1.72 when doing a two-sided test, what is the p-value?

• Your answer

If you run a regression comparing two variables, then run it again with a larger sample, do you expect r^2 to get bigger? Explain why or why not?

• Your answer

True or False: The p-value for a statistical test (that we have done in class) can never be zero.

• True
• False