EDCO745 Statistic2

EDCO 745
REGRESSION DATA SCREENING OUTPUT ASSIGNMENT INSTRUCTIONS
OVERVIEW
Data screening is fundamental prior to any analysis. The basics of data screening are outlined in
your Warner text. This is the third time you will do data screening.
INSTRUCTIONS
1. Using data based upon the topic, instruments, and variables selected in your models you have
developed in Modules 5 and 6, turn in your data screening output for regression analysis.
Note: You only have to turn in output for data screening of the variables you use in your analysis.
Output should include:
• Variable Selections
• Frequencies
• Crosstabs (if necessary)
• Boxplots (if necessary)
• Scatterplot
Note: Your assignment will be checked for originality via the Turnitin plagiarism tool.
EDCO 745
REGRESSION DATA: OUTPUT AND WRITE-UP ASSIGNMENT INSTRUCTIONS
OVERVIEW
Writing a results section is the foundation of the peer-reviewed research journal article. When
you write a results section, you are producing the building block to the discovery of new
knowledge.
INSTRUCTIONS
Provide a results section of a multiple regression analysis using Hayes Process Macro written in
the current APA format
• Include all relevant tables and figures
• Attach your SPSS output to the assignment as well
• Be sure to include a title page and a reference page
You will find examples of Hayes Process writeups in a wide range of research articles.
** Refer to Model 4 Write-up Document (file in Week 6 folder) for more information on
this assignment. Also find writeups of Hayes in the following articles**
Volk, F., Floyd, C. G., Bohannon, K. E., Cole, S. M., McNichol, K. M., Schott, E. A., Williams,
Z. D. R. (2019) The Moderating role of the tendency to blame others in the development
of perceived addiction, shame, and depression in pornography users. Sexual Addiction &
Compulsivity, 26(3-4), 239-261. https://doi.org/10.1080/10720162.2019.1670301
Volk, F., Thomas, J., Sosin, L., Jacob, V., & Moen, C. (2016). Religiosity, developmental
context, and sexual shame in pornography users: A Serial mediation model. Sexual
Addiction & Compulsivity, 23(2-3), 244–259.
https://doi.org/10.1080/10720162.2016.1151391
Note: Your assignment will be checked for originality via the Turnitin plagiarism tool.
RESULTS OF MODERATOR ANALYSIS
Regression Data: Output and Write-up
Sparky Flame
School of Behavioral Sciences, Liberty University
Author Note
Sparky Flame
I have no known conflict of interest to disclose.
Correspondence concerning this article should be addressed to
Sparky Flame
Email: sparky.flame@liberty.edu
1
RESULTS OF MODERATOR ANALYSIS
2
Results of Simple Moderator Analysis
A simple moderator analysis using Hayes (2018) model 1 was completed using the
variables gender, a dichotomous (categorical) variable with two groups of male and female, as
the moderator (W), Stress, as measured by DASS-Stress, a continuous-interval level of
measurement variable, as the predictor (X) and Depression, as measured by DASS-Depression, a
continuous-interval level of measurement variable, as the criterion (Y) from the EDCO 745
course dataset. The data were screened and tests of assumptions for multiple regression were
completed, as demonstrated in the Week 6 assignment, Regression Data: Screening Output
Assignment. Although there was a minor deviation from normality for the residuals, the data
were not transformed. Violation of the normality of residuals does not lead to issues with
significance testing, particularly with larger sample sizes, and does not suggest any discrepancies
with model coefficients should be anticipated. Instead, this violation may be an indication that a
better regression model may be needed (Cohen et al., 2003). The tests of assumptions
demonstrated the data are suitable for conducting the moderator analysis.
The research question of to what extent, if any, does gender moderate the predictive
effect between Stress (X) and Depression (Y), was addressed by testing the null hypothesis of
Gender does not significantly moderate the predictive effect between Stress (X) and Depression
(Y). The relationships examined are presented in Figure 1.
Figure 1
Conceptual Model of Simple Moderator Analysis
Gender
Stress
Depression
RESULTS OF MODERATOR ANALYSIS
3
A moderator analysis using PROCESS (v.4.0; Hayes, 2018) was conducted to test the
null hypothesis. There was a significant predictive effect between Stress and Depression, t =
19.9224, p < .001, b = 0.9353. The result of the interaction effect between gender and Stress upon Depression was not statistically significant, t = -0.8145, p = .4155, b = -0.0251. See Tables 1 and 2. The results failed to reject the null hypotheses, which stated that gender does not significantly moderate the predictive effect between Stress (X) and Depression (Y). The results suggest that although there is a significant predictive effect between Stress and Depression, the extent of the effect between Stress and Depression is not dependent upon one’s gender. Table 1 Model Summary for Moderator Analysis R .8571 R-squared .7345 MSE 40.0287 F 1136.3851 df1 3.0000 df2 1232.0000 p .0000 LLCI -0.1526 0.8432 -2.0661 -0.0856 ULCI 0.3884 1.0274 0.4884 0.0354 Table 2 Model Coefficients Coeff Constant 1.8529 DASSSTRS 0.9353 Gender -0.7888 Int_1 -0.0251 Note. Y variable is DASSDEP SE 1.0222 0.0469 0.6510 0.0308 t 1.8127 19.9224 -1.2116 -0.8145 p .0701 .0000 .2259 .4155