Psychology
StephDiaz
Wee6Day2.pptxWeek 6, Day 2
PSY 3215 U01B 1215, Lab
Lab Overview Statistics review for final paper
Part I: The Independent Samples T-test
Part II: Levene’s Test
Part III: Effect Size
Important point
The test for your final paper (Experimental Paper II) will be the
Independent Samples T-test
This has been confirmed with Dr Bonnin
ParT I
Setting up, Entering & Running the Independent Samples T-Test
Independent T-test
The content in this lecture assumes you have reviewed the lecture from July 21st, Wednesday
IF YOU HAVE NOT REVIEWED THAT LECTURE, YOU MAY GET LOST
Independent T-test
Entering data in SPSS for an Independent T-test is easy
Just make sure to mark your groups
What do I mean by mark?
Independent T-test
To see, go to the data for your paper on canvas
Modules Paper Data Experimental Paper II – Data
Download it and open it up!
Independent T-test
Now, remember:
The second tab (PRE-POST DATA) is the pre-data you will use if you are using two groups
Be sure to read the data; the left data is for results that go down, the right data is for results that go up
Independent T-test
Since this data is considered pre-data, this would be marked as #1 in spss
Mark the first few rows and drag the column all the way to the bottom
Independent T-test
Now, you will repeat this process on the next page, but instead of a 1, you will mark this data as a 2
Congratulations, you have now marked your data with their groups (1 & 2). This can work, even if you’re using pre & post measures.
These marks help SPSS to keep track of the data as being two different kinds
Independent T-test
SPSS needs this help because we are going to stack the groups when we enter them in SPSS
That means that we will enter the DVs in the same column in SPSS; this is different from the paired-samples t-test where we had two different columns
Independent T-test
To begin, go to SPSS, create a new dataset and go to variable view
Now, create three variables
These variables need to be able to track:
1) The subject number
2) The DV measure
3)The IV group (pre-post/control-treatment)
Independent T-test
Now, copy and paste the first data we marked into SPSS
Copy & Paste
Independent T-test
Now, go to the second group you marked and copy and paste that on to the bottom of your SPSS Data
You now have one variable for the DV but have it designated group 1 or 2
Independent T-test
Congratulations! You have stacked data!
With your data in this state, you can now do independent sample t-tests
You can do the same thing for control/experimental data from the first tab in the excel sheet; just add some columns between them, then mark the columns before you copy and paste
Independent T-test
To follow along, let’s look at some different stacked data
Go get the In-Class Practice Data from Canvas
Download it and open it in SPSS
Independent T-test
It’s time to run a t-test!
Go to Analyze Compare Means Independent-Samples T Test…
Independent T-test
To run this, put your DV in the Test Variable Box
For this data, that’s the Ice Cream rating
Then, put your IV in the Grouping Variable Box
We’ll use groups 1 & 2; Chocolate & Vanilla
You’ll only have 2 groups to pick from, so don’t worry about picking the right groups. Just put them both in
Independent T-test
Congratulations! You’ve got your results!
Part II
Levene’s Test
Levene’s Test
Now, one thing that’s very important to ask is: Are the variances between my two groups even?
This is important to understand the character of your sample
It’s also important when you obtain an effect size measure
Levene’s Test
Do understand how to figure this out, go look at the result we obtained from the ice cream data t-test comparison between chocolate & vanilla
Levene’s Test
A significant value says:
The variance between our two groups are significantly different!
This means that we:
A) Have unequal variances
B) Need to use the t-test results from the bottom row
Levene’s Test
Levene’s Test is that easy!
All you have to do is check the sig. value for the Levene’s test
Part III
Effect Size
Effect Size
Now that you have your Levene’s Test results, let’s review at how to calculate your effect size
Remember these breakpoints:
d < 0.2 - Tiny effect size
d > 0.2 & < 0.5 – Small effect size
d > 0.5 & < 0.8 – Medium effect size
d > 0.8 – Large effect size
Effect Size
The formula for effect size:
Effect Size
For between-subjects design, you need:
Means, SD’s, and sample sizes for each group (212 for both groups)
For within-subjects design, you need:
Means, SD’s, and sample size (212 for pre- and post-)
Bonus: Part IV
Showing SPSS Results in Word or other Documents
Part IV
You often want to show off the tables you create in SPSS; this can be done in a variety of ways
First, go to SPSS and chose the “Copy Special” option
Part IV
You have several options
RTF & Image are the most used
Experiment with each to find which one best suites your needs
