For this assignment, you'll
be using some data that I made up. In this mythical study, 30 people are
interviewed as they leave the movie theater. Ten people are walking out
of the original Star Wars movie (technically episode 4, "A New Hope").
Another ten are interviewed as they leave the second Star Wars movie filmed
("The Empire Strikes Back" / Episode 5). And 10 people are interviewed
as they leave "Return of the Jedi" / Episode 6. Each person
answers several questions on a 1-7 scale where 1 = "not very much"
or "very bad" and 7 = "very much" or "very good."
The questions they are asked are:
- How much did you like the
- What was the overall quality
of the movie?
- How good were the special
- How much bad acting was
in the movie?
- How much did the movie remind
you of a Western?
Their data is available in
the following SPSS file:
Your job is to:
- Create a new Google
Doc and name it "PSY 220 Stats 4 [Your last name]"
- Recode the "episode"
variable into a number, and then attach text labels to those numbers.
Note: When recoding, be sure to type the old value (e.g., "New
Hope") exactly like it appears in the data file. Any mis-spellings
will prevent SPSS from performing the recode.
- Reverse-score any variables
that seem like they would produce answers in the opposite direction
of the other variables [Hint: If you're not sure which one(s)
to reverse, use Analyze->Correlate->Bivariate and check the correlations
among all the variables. A correlation between -0.4 and -1 suggests
that one item is the reverse of another. For tips on interpreting correlation
- Compute the Cronbach alpha
reliability of the 5-item scale. Remove any variables that are dragging
down your alpha. Report the Cronbach alpha for your final scale.
- Create a new variable that
is the mean of the items left in your scale.
- Use that mean as the dependent
variable in a one-way ANOVA with the episodes as the levels of the independent
variable. Report the results in APA style. [Note: if
you obtain a significance value in SPSS that reads ".000",
you cannot technically say that p=0. Instead, you would report
- Report the results of Tukey
post-hoc comparisons (in APA style), including confidence intervals.
- Put the means into an APA-style
Table. It might be easiest to copy and paste the table
from the previous page and then modify it.
- Report the results of an
independent t-test that compares the New Hope episode to the
Empire episode in terms of how much it reminded people of a Western.
- Share your Google Doc with
email@example.com and log out of Google Docs.