Description
Submit a synthesis of statistical findings derived from multiple regression analysis that follows the Week 6 Assignment Template. Your synthesis must include the following:
- An APA Results section for the multiple regression test [see an example in Lesson 34 of the Green and Salkind (2017) text].
- Only the critical elements of your SPSS output:
- A properly formatted research question
- A properly formatted H10 (null) and H1a (alternate) hypothesis
- A descriptive statistics narrative and properly formatted descriptive statistics table
- A properly formatted scatterplot graph
- A properly formatted inferential APA Results Section to include a properly formatted Normal Probability Plot (P-P) of the Regression Standardized Residual and the scatterplot of the standardized residuals
- An Appendix including the SPSS output generated for descriptive and inferential statistics
- An explanation of the differences and similarities of bivariate regression analysis and multiple regression analyses
Note: You will cut and paste the appropriate SPSS output into the Appendix. The SPSS output is not in APA format, so you will need to type the information from the SPSS output to the appropriate sections of the APA table. You must use the Week 6 Assignment Template to complete this Assignment. Also, refer to the Week 6 Assignment Rubric for specific grading elements and criteria. Your Instructor will use this rubric to assess your work.
To prepare for this Assignment, review Lesson 16A and Lessons 31–35 in your Green and Salkind (2017) text, the Week 6 Assignment Exemplar and Week 6 Assignment Template documents, as well as the tutorial videos provided in this week’s Resources. Consider how a multiple regression analysis will allow you to answer your research questions effectively.
Explanation & Answer
Let me know if you have questions!
1
DDBA 8307 Week 6 Assignment – Multiple Regression
Your Name
DDBA 8307-6
Your Instructor's Name
2
Multiple Linear Regression
Numerous studies have been dedicated to determining which factors contribute to
a heightened job performance. Such questions as whether salary, experience, or the
experience with the colleagues actually contribute to a heightened work performance,
will be attempted to be answered in this analysis. Fifty-one employees of a major
corporation were randomly chosen, while data on each of the variables was collected for
each individual. Job performance will be assessed using a five-point scale, administered
to each employee's immediate supervisor. The other variables, such as years of
experience, salary, and the number of friends among colleagues, will be self-reported in
survey format.
The use of a multiple linear regression is justified in this analysis, since all of the
variables are either ratio, interval, or ordinal (Liebscher, 2012). The independent variable
(performance rating) will be measured on the ordinal level. The dependent variables –
salary (in 1000's of (S)), experience (in years), and the number of friends in the
workplace are also ratio-level. The necessary assumptions have been tested using the
appropriate assessment measures. The scatterplots of the dependent variable against each
of the predictors indicated a roughly linear relationship. Normal probability plots were
created to confirm the normality of residuals. The correlations between the independent
variables were examined to test whether multicollinearity existed among the predictors.
The multicollinearity issue came up, and will be addressed in further research on this
topic. Finally, the residuals were plotted to test the homoscedascity assumption (Slinker
& Glantz, 2007).
3
Research Question
Do the years of experience, salary, and the number of friends at work significantly
influence employee performance rating?
Hypotheses
H0: The years of experience, salary, and the number of friends at work do not
significantly influence employee performance rating.
H1: The years of experience, salary, and the number of friends at work
significantly influence employee performance rating.
Results
Descriptive Statistics
A sample of 51 workers were surveyed, and their results analyzed. The
assumptions of outliers, normality, linear relationships, and evenly scattered and
independent residuals were assessed with no major violations were detected. Table 1
displays descriptive statistics for all variables.
Table 1. Descriptive Measures and Bootstrapping
Variable
M
SD
Bootstrapped 95% CI (M)
5.14
2.33
[4.45, 5.76]
61.61
17.22
[56.75, 66.24]
Number of Friends
2.41
1.46
[2.00, 2.80]
Performance Rating
2.73
1.56
[2.25, 3.14]
Years of Experience
Salary
Note: N = 51.
Descriptive Measures for All Variables
Variable
M
SD
Years of Experience
5.14
2.33
Salary
61.61
17.22
4
Number of Friends
2.41
1.46
Performance Rating
2.73
1.56
Tests of Assumptions
The appropriate assumptions for multiple linear regression (see paragraph 2) were
addressed. Bootstrapping, with use of 1,000 samples, was employed to help detect
violations.
Multicollinearity. Multicollinearity was assessed via testing the correlations
among the predictors. Although significant correlations between the variables were
found (Table 2), we decided to nevertheless include these variables in the analysis, as
they were qualitatively important to the assessment. Further analyses must be conducted
in order to account for the multicollinearity of these variables.
Table 2
Correlations (r) among The Independent Variables
Experience
Salary
Experience
1.00
.947
Number of
Friends
.841
Salary
.947
1.00
.867
Number of
Friends
Note. N = 51.
.841
.867
1.00
Variable
Outliers, normal distributions, linear relationships, even scatter and
independenc...
Review
Review
Europe(Continental)