Description
In this Assignment, you will access a secondary dataset and construct a research question that can be answered with a moderation or mediation multiple regression analysis.
During this Module, you ideally have used the Support Forum to ask, answer, and otherwise address any questions you had regarding moderation and mediation. In this Assignment, you apply what you have learned to answer a research question using multiple regression, including testing for possible moderation or mediation. Remember that you must still address all the assumptions for multiple regression.
TO PREPARE
Review the module Learning Resources (including media), especially those related to multiple regression, moderation, and mediation analysis methods.
Import the SPSS dataset provided by your instructor.
Review the BRFSS Dataset documentation to familiarize yourself with the variables in the dataset.
ASSIGNMENT
Using the provided dataset, compose a research question that can be answered through a multiple regression analysis. Based on the research question you created, choose either moderation or mediation in your multiple regression analysis technique.
If you chose mediation:
Fit a multiple regression model, testing whether a mediating variable partly or completely mediates the effect of an initial risk factor variable on an outcome variable. Think about whether the model will meet assumptions (or not).
Fit the model, testing for mediation between two key variables.
Analyze the output, determining whether mediation was significant, and interpret that result.
Reflect on possible implications for social change.
If you chose moderation:
Fit a multiple regression model, testing whether a moderating variable partly or completely moderates the effect of an initial risk factor variable on an outcome variable. Think about whether the model will meet assumptions (or not).
Fit the model, testing for moderation between two key variables.
Analyze the output, determining whether moderation was significant, and interpret that result.
Reflect on possible implications for social change.
For both types of analyses:
In two to three pages, excluding the title page and appendix, write an analysis that includes the following:
A description of the research question, hypotheses, and variables, including the type of variable (independent, dependent, mediating/moderating) and level of measurement (categorical, ordinal, continuous).
A clear description and visual display of the statistical analysis in APA Publication Manual, 7th edition (APA) style, including regression tables (e.g., model summary, ANOVA, coefficients) and figures (e.g., regression line) as appropriate.
An analysis of the output, determination about whether the output was significant, and interpretation of the statistical results.
A reflection on possible implications for social change.
The analysis should be in APA format, including a title page, references, and an appendix, which includes your SPSS data output.
Note: For an APA-compliant write-up of these types of analyses, refer to
Warner, R. M. (2021). Mediation. In Applied statistics II: Multivariable and multivariate techniques (3rd ed., pp. 289–308).
The APA-compliant write-up is found on pages 305–306.
OR
Warner, R. M. (2021). Moderation: Interaction in multiple regression. In Applied statistics II: Multivariable and multivariate techniques (3rd ed., pp. 215–253).
The APA-compliant write-up is found on pages 242–243.
DATA
Centers for Disease Control and Prevention. (2021, July 7). Behavioral risk factor surveillance system: Overview: BRFSS 2020https://www.cdc.gov/brfss/annual_data/2020/pdf/overview-2020-508.pdf
Centers for Disease Control and Prevention. (2021, August 6). LLCP 2020 codebook report. Overall version data weightedwith _LLCPWT. Behavioral Risk Factor Surveillance System
https://www.cdc.gov/brfss/annual_data/2020/pdf/cod…
Ferraro, A. (2020). What statistical test should I useWalden University Blackboard. https://waldenu.instructure.com
Document: Guided Sample Dataset: Statistics AnxietyNote: The Statistics Anxiety dataset is not for use in your Assignment. It is used with the Required Media for the Practice: SPSS (Optional) activity.
Crowson, M. (2019, July 27).Multilevel modeling using SPSSJuly, 2019) [Video]. YouTube. https://youtu.be/RU1ps6jaheI
Hageman, K., Kim, A., Sanchez, T., & Bertolli, J. (2015). Chapter 12 | Survey design and implementationIn G. Guest & E. E. Namey (Eds.), Public health research methods (pp. 341–378) SAGE Publications. https://doi.org/10.4135/9781483398839.n12
Walden Office of Research and Doctoral Services. (n.d.). Quantitative, SPSS, and statistics tutoringhttps://academicguides.waldenu.edu/research-center…
Warner, R. M. (2021). Dummy predictor variables in multiple regression. In Applied statistics II: Multivariable and multivariate techniques (3rd ed., pp. 187–214). SAGE Publications.
Warner, R. M. (2021). Multiple regression with multiple predictors. In Applied statistics II: Multivariable and multivariate techniques (3rd ed., pp. 133–186). SAGE Publications.
Download BRFSS_PHYSICAL1(1)-2.sav (98.4 KB)
Unformatted Attachment Preview
Mediation
Mediation
Program Transcript
[MUSIC PLAYING]
MATTHEW JONES: This week, we’re talking about mediation. And just like with
moderation, there are a number of different theoretical and conceptual viewpoints
on how mediation analysis should be carried out. This week in SPSS, I’m just
going to show you one approach of many. And this is often referred to as the
Baron and Kenny approach to mediation. This week, we’re going to be looking at
the question of whether motivation mediates the relationship between anxiety
and final exam score.
Students entering a statistics class might have a high level of anxiety, but that
anxiety might also increase motivation, which in turn might affect final exam
scores. But we can go to SPSS and test this out with our simulated data set. So
using the Baron and Kenny approach, it’s really a four step approach in
performing a multiple regression analysis. So I have to do a series of simple
regression analyzes first.
And the first regression analysis I do is testing the pathway between our
independent variable, class anxiety, and our dependent variable, final score-final exam score. So it can go ahead and click Analyze, Regression. So very
bivariate regression should be familiar to everybody. So in the Dependent
Variable box, I’m going to put final exam score. And in the Independent Variable
box, I’m going to put classic anxiety at time one.
And I can go ahead and run my regression analysis. The output comes up. I can
see the ANOVA. The overall regression is significant. And indeed, classic anxiety
is a significant predictor. So you can see a positive unstandardized coefficient.
As anxiety increases, so does the final exam score.
Now, I need to test for a relationship between my mediator, motivation, and final
exam score. So we just do the same thing– Analyze, Regression, Linear. And so
we’re just going to move our independent variable of anxiety out and move
motivation score, so it’s measured also from an instrument– move that into the
Independent Variable. Click OK. OK.
And we see this, from the ANOVA, the regression is significant. The coefficient is
significant. As motivation increases, so does the final exam score. OK. Then we
need to test for the relationship between our independent variable of class
anxiety and motivation. So back up to Regression, Linear.
So we’re testing each of these direct paths. So motivation moves into the
Dependent Variable box, class anxiety at time one moves into our Independent
Variable. And indeed, we also see, jumping down to our coefficients, that class
© 2017-2022 Walden University, LLC
1
Mediation
anxiety is a significant predictor of motivation. So the final step is we’re going to
conduct a multiple regression analysis. So back to Regression.
And we’re going to put our dependent variable of interest, the final exam score-we already have class anxiety as our independent variable. So we just need to
move motivation back in there. Now that we see that both of our predictive
variables are still statistically significant, and we can see looking at our
coefficients– I’ll look here at our standardized beta– have actually decreased
from when we were just doing that bivariate regression analysis. So that is one of
the variables, motivation, has some sort of mediating effect on class anxiety.
So far we’ve tested all the direct effects in SPSS, but there’s one last important
step to mediation analysis. And that is testing for the indirect effect. There are
many different ways to test for this indirect effect, but one specific test that I’ll
mention is the Sobel test. You can go to an outside Sobel test calculator on the
web– and we’ve listed one of those under your week three resources. But there
are also many others. There are also SPSS macros that can be used to test this
indirect effect as well.
If you’re interested in this, you might want to speak to your instructor about his or
her preference or other resources that he or she may know about.
[MUSIC PLAYING]
Mediation
Additional Content Attribution
FOOTAGE:
GettyLicense_160463144
simonkr/Creatas Video/Getty Images
GettyLicense_626791754
AzmanL/Creatas Video/Getty Images
GettyLicense_114759820
David Baumber/Vetta/Getty Images
© 2017-2022 Walden University, LLC
2
Introduction to Moderation
Introduction to Moderation
Program Transcript
[MUSIC PLAYING]
DR. ANNIE PEZALLA: In the previous week, you refreshed your memory of
some of the more straightforward and useful approaches to multiple regression.
This week, you’re introduced to an expanded approach to multiple regression
that tests for a moderation effect in a model. To illustrate moderation, let’s take
an example, one that is relevant to almost anyone taking a statistics class.
Statistics can scare people, and some people are more likely to shut down and
not ask for help in a statistics class than are others. Now you might wonder, is
there a relationship between gender and fear of asking for help in a statistics
class? You might think that women are more likely than men to report a fear of
asking for help, or maybe you think the opposite.
No matter what you think, the relationship between those two variables– gender
and fear of asking for help– is probably quite different depending on the comfort
one has with math. That is, if you’re studying a group of people who all have a
very low comfort with math, women may, indeed, be more fearful than men to
ask for help in a statistics class. So there may be a strong relationship between
gender and fear of asking for help in such a group.
Yet when math comfortability is high, the relationship between gender and fear of
asking for help may be weak or non-existent. That speculation could be tested
with a moderation model. So how could you frame your research question to
examine the role of math comfortability between those variables?
The research question you’d ask is, does math comfortability moderate the
relationship between gender and fear of asking for help?
Essentially, moderation tells you whether the strength of the relationship between
your predictor variable and your dependent variable changes based on the value
of a third variable. My colleague and friend, Dr. Jones, will walk you through this
test in SPSS.
© 2017-2022 Walden University, LLC
1
Introduction to Moderation
Introduction to Moderation
Additional Content Attribution
FOOTAGE:
GettyLicense_160463144
simonkr/Creatas Video/Getty Images
GettyLicense_626791754
AzmanL/Creatas Video/Getty Images
GettyLicense_114759820
David Baumber/Vetta/Getty Images
© 2017-2022 Walden University, LLC
2
Introduction to Mediation
Introduction to Mediation
Program Transcript
DR. ANNIE PEZALLA: This week you’re exposed to a slightly different approach
to multiple regression. Up to this point, we’ve assumed that, in the conception of
our models, the independent variable either directly affects or does not affect our
dependent variable. But life isn’t always that simple. It’s pretty common for there
to be an intervening variable through which the independent variable passes to
effect the dependent variable.
Those intervening variables are called mediators, and they serve the same sort
of function that a mediator might serve between two people who have some sort
of relationship but the nature of their relationship is unclear and requires
someone to essentially go between them to help clarify the ways in which they
relate. Testing for mediation gives you the ability to test for those sorts of
relationships, to examine both the direct and indirect effects between your
variables of interest.
Let’s return to our example with the statistics professor and his anxious students.
Using a simple linear regression model, this professor might test for a
relationship between student anxiety of statistics and end of course knowledge
attainment of statistical methods. You might think that anxiety might be a pretty
good predictor of end of course knowledge, that high anxiety predicts low end of
course knowledge.
But the relationship between those variables is almost always a little more
complicated than that. Anxiety could be a good thing or a bad thing. Indeed,
sometimes anxiety is just a reflection of motivation to learn the course material.
From that perspective, motivation could mediate the relationship between anxiety
and end of course knowledge
Dr. Jones will be testing those relationships for you this week as he answers the
question– does motivation mediate the relationship between anxiety and final
score in a statistics course? Both mediation and moderation are incredibly
powerful and popular statistical tests and research. Think about how you might
use either or both of these tests with a dataset that you’ve chosen to use.
Introduction to Mediation
Additional Content Attribution
FOOTAGE:
GettyLicense_160463144
simonkr/Creatas Video/Getty Images
© 2017-2022 Walden University, LLC
1
Introduction to Mediation
GettyLicense_626791754
AzmanL/Creatas Video/Getty Images
GettyLicense_114759820
David Baumber/Vetta/Getty Images
© 2017-2022 Walden University, LLC
2
Purchase answer to see full
attachment
