1. Unless otherwise stated do not copy and paste SPSS output.
2. Whenever using hand calculations show all work either within the problem or as an appendix. Partial credit will not be allotted otherwise.
I. Rearrange the data in a table similar to Table 13.2, showing the variables, marginal means, cell means and grand mean. You don’t need to show the raw data in the table.
II. Use Excel, Word, or R (i.e., not using SPSS) to construct on the same plot the cell means and marginal means by putting Parity on the X axis. Ensure lines are meaningfully labeled (i.e. low-birthweight<18).
III. Use Excel, Word, or R (i.e. not using SPSS) to construct on the same plot the cell means and marginal means by putting Birth Weight on the X axis. Ensure lines are meaningfully labeled (i.e. primiparous).
IV. Run the analysis in SPSS. I advise that you identify the values of your groups appropriately. You may copy and paste or screen shot the SPSS output. By selecting the appropriate table, identify and label the grand, marginal, and cell means using the notation as in your notes.
V. Manually compute the SS for the main effects and the interactions. Make sure to show your work.
VI. Using data only from the ANOVA table, compute the standard deviation across all 60 observations. Be sure to show your work.
VII. Create a ANOVA table to summarize your results in a format similar to table 13.2 (c) in your book, and include information about effect size and power.
VIII. Run at least three post-hoc tests if necessary and
interpret your results
. Discuss findings for each test AND any differences/similarities between tests.
IX. By taking into account the various pieces of information from your analysis, what statistical conclusions can you draw from the study. Interpret those findings.
Compute for the effect for Parity
(use the formula in chapter 11) and explain what it means.
Que 3.
Refer to exercise 13.4 in your text (Do in SPSS). Make sure to report the appropriate statistics. What conclusions can you draw for this analysis?
(See instructions for running simple effects on the next sheets, questions four and five follow instructions – do not miss them!)
Instructions for running simple effects in SPSS
Illustrative example based on data in Table 13.2 (7 & 8th Edition)
I labeled the DV as Recall and my IV’s are Age and Condition. I advise you to get the data online and rename the variables and follow the instructions below to get the simple effects.
Simple Effects of Condition in SPSS
1. Start as you would run an ANOVA follow the screen shots to run the ANOVA from your PPT slides.
To obtain the two sets of simple effects due to Condition, one for Old and one for Young, click Options. Under the Estimated Marginal Means in the box labeled Factor(s) and Factor Interactions, move the OVERALL, Age and Condition and Age by Condition effects into the box labeled Display Means for. Click Continue. Click Paste TAB located at the bottom of window. The syntax window will open up.
The following syntax will appear with the following commands:
===============================================
UNIANOVA Recall BY Age Condition
/METHOD=SSTYPE(3)
/INTERCEPT=INCLUDE
/EMMEANS=TABLES(OVERALL)
/EMMEANS=TABLES(Age)
/EMMEANS=TABLES(Condition)
/EMMEANS=TABLES(Age*Condition)
/CRITERIA=ALPHA(.05)
/DESIGN=Age Condition Age*Condition
==================================================
To obtain the omnibus (simple effects) test for Condition for each Age level, type “ COMPARE (Condition)” without the quotes in the command line indicated below. The changes are noted in bold
==================================================
UNIANOVA Recall BY Age Condition
/METHOD=SSTYPE(3)
/INTERCEPT=INCLUDE
/EMMEANS=TABLES(OVERALL)
/EMMEANS=TABLES(Age)
/EMMEANS=TABLES(Condition)
/EMMEANS=TABLES(Age*Condition) COMPARE (Condition)
/CRITERIA=ALPHA(.05)
/DESIGN=Age Condition Age*Condition.
To run the analysis, right Click and select Run, All. Alternatively, you can click on the right-pointing (Green arrow head) icon just below “Utilities” on the main menu bar.
Output for simple effects of Condition
Your output will contain a lot of information. But included in the output will be: Estimated marginal means for Age, Condition and Age by Condition Interaction. It will also contain Pairwise Comparisons and Univariate Tests. The section on the Univariate Tests is where you have the SS of the simple effects of Condition at the two levels of Age. I have pasted the results below
|
Univariate Tests
|
|
Dependent Variable:Recall
|
|
Age
|
Sum of Squares
|
df
|
Mean Square
|
F
|
Sig.
|
|
Old
|
Contrast
|
351.520
|
4
|
87.880
|
10.950
|
.000
|
|
|
Error
|
722.300
|
90
|
8.026
|
|
|
|
Young
|
Contrast
|
1353.720
|
4
|
338.430
|
42.169
|
.000
|
|
|
Error
|
722.300
|
90
|
8.026
|
|
|
|
Each F tests the simple effects of Condition within each level combination of the other effects shown. These tests are based on the linearly independent pairwise comparisons among the estimated marginal means.
|
The top row in the table shows the results for the simple effect of the “Condition” factor within the “Old” group
Similarly, the bottom row refers to the simple effect of the “Condition” factor within the “Young” group
##############################################################################
Simple Effects of Age in SPSS
If you are interested in the Simple Effects of Age at any (or all) the five levels of Condition, you replace the COMPARE (Condition) by COMPARE (Age)
\=============================================================
UNIANOVA Recall BY Age Condition
/METHOD=SSTYPE(3)
/INTERCEPT=INCLUDE
/EMMEANS=TABLES(OVERALL)
/EMMEANS=TABLES(Age)
/EMMEANS=TABLES(Condition)
/EMMEANS=TABLES(Age*Condition) COMPARE (Age)
/CRITERIA=ALPHA(.05)
/DESIGN=Age Condition Age*Condition.
The Univariate Tests will give you the simple effect of Age at each of the five levels of Condition
|
Univariate Tests
Dependent Variable: Recall
|
|
Condition
|
Sum of Squares
|
df
|
Mean Square
|
F
|
Sig.
|
Partial Eta Squared
|
Noncent. Parameter
|
Observed Powera
|
|
1
|
Contrast
|
1.250
|
1
|
1.250
|
.156
|
.694
|
.002
|
.156
|
.068
|
|
|
Error
|
722.300
|
90
|
8.026
|
|
|
|
|
|
|
2
|
Contrast
|
2.450
|
1
|
2.450
|
.305
|
.582
|
.003
|
.305
|
.085
|
|
|
Error
|
722.300
|
90
|
8.026
|
|
|
|
|
|
|
3
|
Contrast
|
72.200
|
1
|
72.200
|
8.996
|
.003
|
.091
|
8.996
|
.843
|
|
|
Error
|
722.300
|
90
|
8.026
|
|
|
|
|
|
|
4
|
Contrast
|
88.200
|
1
|
88.200
|
10.990
|
.001
|
.109
|
10.990
|
.907
|
|
|
Error
|
722.300
|
90
|
8.026
|
|
|
|
|
|
|
5
|
Contrast
|
266.450
|
1
|
266.450
|
33.200
|
.000
|
.269
|
33.200
|
1.000
|
|
|
Error
|
722.300
|
90
|
8.026
|
|
|
|
|
|
|
Each F tests the simple effects of Age within each level combination of the other effects shown. These tests are based on the linearly independent pairwise comparisons among the estimated marginal means.
|
|
a. Computed using alpha = .05
|
##############################################################################
More syntax – for additional information
Note you could ask for more stuff like descriptives, effects size etc by clicking where appropriate. Remember to click Paste to get the syntax. It may look like below.
======================================================
UNIANOVA Recall BY Age Condition
/METHOD=SSTYPE(3)
/INTERCEPT=INCLUDE
/POSTHOC=Age Condition(TUKEY BONFERRONI)
/PLOT=PROFILE(Condition*Age Age*Condition)
/EMMEANS=TABLES(OVERALL)
/EMMEANS=TABLES(Age)
/EMMEANS=TABLES(Condition)
/EMMEANS=TABLES(Age*Condition)
/PRINT=OPOWER ETASQ HOMOGENEITY DESCRIPTIVE
/CRITERIA=ALPHA(.05)
/DESIGN=Age Condition Age*Condition
However, because you do not have the COMPARE statement you will not get the simple effects when you run the syntax.
To run the program you need to Click Run, All or you can click on the right-pointing arrow head in the menu bar.
Que 4.
Create a dataset for 2 x 2 balanced design, with n=5 or greater per cell, that has two main effects but no interaction. Show the raw data, the cell means and marginal means and a plot of the cell means.
Que 5.
Compute the power for the interaction effect in Que 1 using GPower. Include a screen shot (not a pic from your phone) of the charts and data that was generated. Interpret the result.
2
ˆ
w
ashitjha
hw_3.docx
0
