This is the culmination of what you’ve done this semester. I am trying to test your ability to formulate a question, select a test, check its assumptions, visualize your data, run the test, and appropriately report and interpret what you have done – to think like a psychologist about statistics.
This is an individual assignment. You may not skip it. It will be due at the end of our last class during the final exam period.
Throughout the semester, you have been developing an understanding of hypothesis testing and effect size measurement. You have been delving into multiple descriiptive and inferential statistics: when it is appropriate to use each one (including how to check the assumptions necessary to use them confidently), how to compute them by hand, and – visualizing your data first – how to run them in SPSS.
This end-of-the-semester assignment is about putting your statistical skills to use to explore and understand data on your own. During the first of two classes during which you may work on the analyses, I will share datasets with you.
Our penultimate (second-to-last) class will be devoted to the analyses. I will show you how to get the data into SPSS, will show you some issues to resolve in the data (if needed), and I will be around to offer you advice and encouragement as you complete the assignment. It will be due at the end of our last class: there will likely be time to finish the assignment in that last class, if needed.
You will use SPSS to complete this assignment. The variables you choose are up to you. My interest here is in seeing that you can choose relevant variables and then run, write up APA-style, and appropriately interpret a range of important statistical procedures using those variables.
You will complete multiple tests. For each test you run, answer these questions: number your answers:
1) Name the variables involved. Briefly indicate why you chose them. For each, include the SPPS name of the variable, the label of the variable, and what the variable measures if that is not clear from the name and label.
2) State your hypothesis (or hypotheses) in plain English. Be sure to include the null and the alternative (but don’t use those words – that’s not plain English!)
3) Indicate what statistical test you’ve chosen and why (The why is very important – why is this the correct test to use for your data?) Also indicate what you had to do to the data to prepare it for your analysis (if anything)
4) State your null and alternate hypotheses in mathematical terms (DO use those words here!)
5) If applicable, indicate what assumptions need to be met to use the test you have chosen and make clear arguments or show SPSS output (or both) to indicate whether those assumptions are met.
6) Provide basic descriiptive statistics for the variables involved: means and standard deviations.
7) Visualize your data (I agree that chart builder is annoying): include a graph or graphs of your results. If a graph is inappropriate, include a table. HINT – I will be looking for error bars when needed in graphs!
8) Include relevant SPSS output for the test. Include effect size and the amount of variance explained if possible.
9) Write up the results in APA style following Gravetter’s guidelines from the appropriate “In the Literature” portion of your textbook.
10) Explain in plain English what the results you obtained mean, as you would to a motivated and curious NCC Psychology student who knows nothing about statistics. In your homework, this last part has been difficult. Be sure to phrase this as you would to someone who is not a student in the class. Use your own words!
Divide your work into four parts, one part per test, using headers in this way:
Your answers to the 10 questions go here, numbered, and you should be sure to include an test-specific information I ask for in the next section of these instructions.
The same, but for the second test.
… and so on.
Each part (each test) has additional instructions.
Part One is about the t-test.
Run either an independent samples t-test or a related samples t-test with accompanying an r-squared and Cohen’s d. Report a 95 percent confidence interval in terms of raw scores.
Part Two is about ANOVA
Run an ANOVA with at least one post-hoc test – justify your choice of that post-hoc test. It is up to you whether you run a One-Way or Two-Way ANOVA.
Part Three is about correlation
Any of the kinds covered in Chapter 14 will suffice. Make sure it is very clear why you have chosen that kind of correlation over the others: address each of the others to say why it is not the best option. Be sure to include and interpret a coefficient of determination! Test the significance of the correlation.
For reasons covered in Zoom, I am not emphasizing regression this semester.
Part Four is about the Goodness-of-Fit Chi-square test or the Chi-Square Test for Independence
Choose one type of Chi-Square test to use. Be certain to include a table.
You will have 4 sections, each with 10 questions, for 40 questions in all. (I have eliminated two sections in the hopes of making this activity the right size for a meaningful end-of-class demonstration of your abilities. I will give you 0.5, 0.25, or 0 points for each question on a holistic basis – full credit, half credit, or no credit. This will lead to up to 20 points.
This is a 30-point assignment, so I will give you 10 additional points if you simply address all the questions. If you do not answer 1-4 of them, I will give you 5 of those 10 points. If you do not address 5 or more of them, I will give you 0 of the 10 points.