In: Statistics and Probability
Please note that for all problems in this course, the standard cut-off (alpha) for a test of significance will be .05, and you always report the exact power unless SPSS output states p=.000 (you’d report p<.001). Also, remember that we divide the p value in half when reporting one-tailed tests with 1 – 2 groups.
Problem Set 1: Two-way ANOVA (8 pts) Research Scenario An Industrial/Organizational psychologist conducted a study examining differences in sex (women and men) and primary mode of communication with superiors (face, email, indirect) on perceived likelihood of receiving a raise in the next 6 months. Perceived likelihood was measured in percent likelihood of expecting a raise (e.g., 0 indicates they absolutely do not expect a raise in the next 6 months). Twenty-eight participants completed the study. Their results are in the table below. Conduct a two-way ANOVA to determine whether perceived likelihood of receiving a raise in the next 6 months is affected by sex and/or primary mode of communication. Remember to name and define your variables (your two independent variables and your one dependent variable) under the “Variable View,” then return to the “Data View” to enter and analyze the data. (Note the data set is small to ease your burden – use a two-way ANOVA regardless!)
|
SPSS output:
Tests of Between-Subjects Effects |
|||||
Dependent Variable:RAISE |
|||||
Source |
Type III Sum of Squares |
df |
Mean Square |
F |
Sig. |
Corrected Model |
21252.631a |
5 |
4250.526 |
8.994 |
.000 |
Intercept |
96217.338 |
1 |
96217.338 |
203.594 |
.000 |
SEX |
11870.330 |
1 |
11870.330 |
25.117 |
.000 |
COMMUNICATION_MODE |
7228.342 |
2 |
3614.171 |
7.648 |
.003 |
SEX * COMMUNICATION_MODE |
1035.940 |
2 |
517.970 |
1.096 |
.352 |
Error |
10397.083 |
22 |
472.595 |
||
Total |
140276.000 |
28 |
|||
Corrected Total |
31649.714 |
27 |
|||
a. R Squared = .671 (Adjusted R Squared = .597) |
Multiple Comparisons |
||||||
RAISE Tukey HSD |
||||||
(I) COMMUNICATION_MODE |
(J) COMMUNICATION_MODE |
Mean Difference (I-J) |
Std. Error |
Sig. |
95% Confidence Interval |
|
Lower Bound |
Upper Bound |
|||||
FACE |
|
29.8545* |
9.49856 |
.013 |
5.9936 |
53.7155 |
INDIRECT |
41.5429* |
10.71322 |
.002 |
14.6306 |
68.4551 |
|
|
FACE |
-29.8545* |
9.49856 |
.013 |
-53.7155 |
-5.9936 |
INDIRECT |
11.6883 |
10.51079 |
.517 |
-14.7155 |
38.0921 |
|
INDIRECT |
FACE |
-41.5429* |
10.71322 |
.002 |
-68.4551 |
-14.6306 |
|
-11.6883 |
10.51079 |
.517 |
-38.0921 |
14.7155 |
|
Based on observed means. The error term is Mean Square(Error) = 472.595. |
||||||
*. The mean difference is significant at the .05 level. |
Graphs:
Results
The null hypothesis, ho: the main effect of sex is not significant. The alternative hypothesis, h1: the main effect of sex is significant. WIth F=25.117, P<5%, I REJECT ho and conclude that the main effect of sex is significant.
The null hypothesis, ho: the main effect of COMMUNICATION_MODE is not significant. The alternative hypothesis, h1: the main effect of COMMUNICATION_MODE is significant. WIth F=7.648, P<5%, I REJECT ho and conclude that the main effect of COMMUNICATION_MODE is significant.
The null hypothesis, ho: interaction effect is not significant. The alternative hypothesis, h1: the interaction effect is significant. WIth F=1.096, P>5%, I fail to REJECT ho and conclude that the interaction effect IS NOT significant.
From posr Hoc analysis, with p<5%, there is significant difference in the mean raise between (face, indirect) and (email and face).