Psychology
Week 6, Day 1
PSY 3215 U01B 1215, Lab
Lab Overview
Part I: Advanced Analyses in SPSS
Part II: Odds & Ends for your Final Paper
Part I Advanced Analysis in SPSS
Part I Setting up
In this section, we’re going to be covering how you can do advanced analyses in SPSS; looking at two or more independent variables would fall under this example
To begin, let’s go back to the in-class practice data that was used for the second half of last week’s day 2 lecture
Canvas Modules Example Data In-Class Practice Data
Part I Pruning data
It’s very common to eliminate a portion of your data in a study to focus on a smaller subset. For example, let’s say you wanted to eliminate all the ratings of Strawberry ice cream from your data so you can just focus on comparing Chocolate vs Vanilla…
To start this, in SPSS go to…
Data Select Cases
Part I Selecting specific data
Now, what you want to do is select only the data you want to work with.
In this case, that’s the chocolate and the vanilla data
Go to the Select Box, then hit the
If button from “If condition is satisfied”
Part I Selecting specific data
Now, be sure to enter exactly:
(flavor = 1)|(flavor = 2)
Then hit Continue at the bottom of the box
What you have just said is:
Only select data if the flavor variable is 1 or 2 (in our case, vanilla or chocolate)
Part I Selecting Specific Data
Now that you have the right select condition, make sure the output is set to “Filter out unselected cases” and then hit Ok
Your data will no long analyze any strawberry ice cream measures
Part I Basic uses of selected data
Now with your data selected specially, you can do things like look at the mean rating for ice cream, but without strawberry
Remember, to find the mean variable, go to Analyze Descriptive Statistics Descriptives
From there, you would select the rating and look at the descrptives
Part I More advanced analyses
However, now that we’ve pruned out the strawberry data, we now have a chance to look at ratings based on a 2 x 2 design
In addition to flavor, there is another independent variable in the data: store
In this data, ice cream was sold at one of two stores
1: Uptown
2: Downtown
It can be useful to see if there’s a different in the rating based on flavor, but what if we also wanted to see if there was a different in rating based on store (looking at 2 IVs, each with 2 levels/groups)
Part I More advanced analyses
To start this, go to:
Analyze General Linear Model Univariate
This will let us look at a single variable (single – uni; variable – variate; univariate!) in the context of multiple independent variables
Part I Univariate Analysis
To set this up properly, first:
Move the dependent variable (rating) to the Dependent Variable box
Next, put your two IVs (flavor & store) in the Fixed Factor(s) box
Independent variables are sometimes called fixed factors because they are a fixed/unchanging part of your study; something that is predetermined
Part I Univariate analysis – The details
Now, you’re going to want to check to see if our DV (rating) is affected by either of our 2 IVs
However, when looking at 2 or more IVs with two or more groups, something interesting can happen
You can get what’s called an interaction effect; something where both the IV 1 & IV 2 together are having a special influence on the DV measure
For our example, this might mean the rating for uptown chocolate ice cream is better than either the effect from uptown or the effect from chocolate ice cream!
Part I Univariate analysis – The details
To look for this kind of effect, go to the EM Means box option on the side
While here, make sure you have all three options moved to the “Display Means for” box, and select the “Compare Main Effects” box
Part I Univariate analysis – The details
Hit Continue Ok, then go to look at the output window
From here, we can see that both the flavor & the store variables had a significant influence on how much people rated enjoying their ice cream!
Do you think there was any kind of special interaction between the flavors & the store?
Part II Misc Notes
Part II - Misc notes Final paper Content
Abstract
Remember: Purpose of Research, Study Design, Sample Size & Unique Characteristics, Statistical Analyses
Our statistical analyses is the paired-samples t-test that was covered in last week’s lab
Appendix
Demographics & Consent
Start each one on it’s own page and make sure to label it!
Part II - Misc notes Final paper Content
Overall Writing Quality
Make sure to check your tense – the rubric for Experimental Paper I can help with this
Make sure to use scientific and objective terms; no slang, vernacular or abbreviated words (even if the abbreviated word can pass as a regular word!)