Correlations
What are the three characteristics of correlations? And what do each mean?
Form (linear, curvilinear, no relation), direction (positive or negative), and magnitude (strength of relationship)
What is the key difference between correlation and regression?
Correlation examines relationships without directionality, while regression predicts an outcome variable from a predictor variable
What does an independent samples t-test compare?
The means of two independent groups.
Bonus (100 pts): what test do you use if you have more than two groups?
A scientist investigates whether there is a relationship between air pollution levels and asthma rates in different cities. What statistical test should be used? State the IV(s) and DV as well.
Correlation
IV(s): pollution levels, asthma rates
DV: none
When are Lexi's office hours?
T/TH, 12:30 - 1:30
A researcher finds two correlations: r = -.81 and r = .56. Which is stronger?
r = -.81
What would the null and alternative hypotheses be for these two variables using a one-tailed test: sunglasses sales and outside temperature?
Null: no relationship between sunglasses sales and outside temperature.
Alt: positive relationship between sunglasses sales and outside temperature.
What would the null and alternative hypotheses be for these two variables using a one-tailed test: age and stress?
Null: no difference in stress between [younger] and [older] individuals.
Alt: [older] individuals will report [more] stress than [younger] individuals.
A company wants to predict employee job satisfaction using both salary and the number of hours worked per week. What statistical test should be used? State the IV(s) and DV as well.
Multiple linear regression
IV(s): salary, hours worked
DV: job satisfaction
What would Lexi's first act as president be?
Make LED headlights illegal.
If a correlation between stress and GPA is r = -.56, p = .003, what does this mean?
A significant, moderate-to-strong negative correlation
What do residuals in a regression model represent?
The difference between observed and predicted values of Y; how far away the data points are from the line of best fit
If t = 2.48 and p = .008, what does the t-value tell us about our groups and do we reject or fail to reject the null hypothesis?
T-value tells us that group 1 has a larger mean than group 2; reject the null because p < .05
A psychologist is interested in predicting a person's happiness level based on their number of social interactions per day. What statistical test should be used? State the IV(s) and DV as well.
Simple linear regression
IV(s): social interaction
DV: happiness
Please correct this APA write-up. Hint, there are 3 mistakes.
An independent samples t-test was conducted to compare memory score between individuals with Alzheimer's (n = 15) and a control group (n = 15). There was not a significant difference in memory score between the groups, t(30) = -3.98, p = .031, d = .73. Those with Alzheimer's (M = 13.85, SD = 4.98) scored lower on the memory test than the control group (M = 28.14, SD = 3.71).
What is range restriction, and how does it impact correlations?
It reduces variability in X or Y, decreasing the correlation strength
What is R-squared (R²) in regression?
The proportion of variance in Y explained by X
If a 95% CI does not contain zero and p < .05, what does this mean?
The group means are significantly different, and we can confidently reject the null hypothesis.
A professor wants to know if final exam scores differ between students who attended review sessions and those who did not. What statistical test should be used? State the IV(s) and DV as well.
Independent samples t-test
IV(s): attendance (attended vs did not attend)
DV: mean of final exam scores
What are the steps in SPSS to run a multiple linear regression?
1. Check normality using a histogram (-2 to +2; bonus 100 pts)
2. Visualize linear regression for all IVs using scatterplots (bonus 100 pts)
3. Analyze > Regression > Linear: DV in dependent box, all IVs in independent box
4. Statistics: estimates, model fit, CIs, and descriptives
When should Spearman’s rho (⍴) be used instead of Pearson’s r?
When data are skewed, bimodal, or non-normally distributed
Give an example of a real-world multiple regression scenario.
I'll be the judge, muwahaha.
What test is used to check the assumption of homogeneity of variance and how do you interpret it?
Levene’s test.
If Levene's is significant, use 'equal variance not assumed'.
If Levene's is not significant, use 'equal variance assumed'.
A sports analyst wants to determine whether a basketball player’s shooting accuracy can be predicted by their number of hours of practice per week. What statistical test should be used? State the IV(s) and DV as well.
Simple linear regression
IV(s): practice hours
DV: shooting accuracy
Write an APA write-up using this information: (I'm not paying for Jeopardy so I will pull up an image -- please hold).
A simple linear regression was calculated to predict how many books needed to order based on the number of students enrolled. A significant regression equation was found, F(1,10) = 42.48, p < .001, R2 = .81. Specifically, the more times the more students enrolled in a course, the more books needed to be ordered, β = .90, t(10) = 6.52, p < .001. These results indicate that we need to order almost one book for every student.