Reference no: EM132262213

Assignment -

Instructions: On this assignment, you will be asked to do two things:

1) Conduct analyses using a database provided for you, and

2) Answer questions regarding these analyses as well as conceptual principles regarding these analyses. Consequently, you will need to turn in 1) the output of the analyses that you conducted and 2) a document including your typed answers to the questions.

For the output of your analyses: Please do not send me your SPSS output files. There are too many different versions of SPSS out there, and the output files do not always open correctly. Please "Copy and Paste" (crtl-c; ctrl-v) the relevant sections of your output into a word document. This way there won't be any compatibility issues. Alternatively, you can export the SPSS output into a PDF d ocument.

For example: if asked what the limits of a correlation coefficient are, if you answer, "It's limited between -1 and +1, and tells you whether variable 1 causes variable 2".... Then you lose credit. Yes +1 and -1 are the limits, but to add that it implies causation is definitely wrong. If you had just said "It's limited between -1 and +1" then that's a full credit answer. I do this so that people cannot write 20 sentences of junk, and then try to argue that they were "correct" in one of these sentences. Unnecessarily long answers thus can be problematic - do not confuse my 'maximum length' restrictions for 'suggested length.'

What does this mean? You are responsible for providing a complete answer to the questions; but be careful to not just babble. Stick to the question, and make sure that your wording is correct. Balance being brief and being complete.

Also, there is no need for you to cite any of your answers. You should be able to answer all of these questions from the material presented in lecture.

Part I: Mediation

1) Explain conceptually (don't tell me how to test it statistically...yet) what mediation is. What does it mean if a variable mediates the effect of X and Y? Feel free to speak about causation because we are discussing the topic here at a conceptual level. Please give an example. (no more than 6 sentences)

2) Explain statistically how mediation is tested.

A) How is this done with the Baron & Kenny method?

B) How is this done with this Sobel method?

C) How is this done with the bootstrapping method?

Your answers to each of these should communicate that you understand not only how to create the output (simple) but also how this method accomplishes the goal of testing mediation. Note that to provide a complete understanding of part C, you will need to briefly describe what bootstrapping is and not just how it can be used to test for mediation. (no more than 1 page total)

3) Get dataset ExamMediation.sav. Test for mediation using all three of these approaches (variable names are obvious; feel free to label them something of your choosing). A) Provide the output of these results, B) conceptually interpret the test that you think is most trustworthy for testing mediation (like you would write in a journal article. Here you are not graded based on style or formatting; I'm just looking for clear and accurate language). (1-2 paragraphs for interpretation)

Part II: Moderation

4) Explain conceptually (don't tell me how to test it statistically...yet) what moderation is. What does it mean if a variable moderates the effect of X and Y? Feel free to speak about causation because we are discussing the topic here at a conceptual level. Please give an example. (no more than 6 sentences)

5) Explain statistically how moderation is tested.

A) Why is a multiplicative term added; what does it do?

B) If the interaction is significant, what does that mean?

C) When you have a significant interaction, what should you do next?

D) When you have a non-significant interaction, what should you do?

E) What is the difference between a main effect and a simple effect (aka 'simple slope')? (no more than 1 page total)

6) Get dataset ExamModeration.sav. Test for moderation using regression (traditional or PROCESS) (variable names are obvious; feel free to label them something of your choosing).

A) Provide the output of these results,

B) Conceptually interpret the test, including simple slope tests if the interaction is significant. You do not need to plot it, if there is an interaction. (1-2 paragraphs for interpretation)

Part III: Logistic Regression

1. What problems occur when we use standard Ordinary Least Squares regression to predict a dichotomous outcome?

2. How does Logistic Regression address the problems in question 1?

3. Describe the relationship between probability, odds, and odds ratios. Make sure that you give an example (make one up) that illustrates how you would interpret each of these three things.

4. Using logistic regression and the "Exam1 Logistic.sav" dataset, do the following: Predict the variable "outcome" from the variables "variable1", "variable2", and "variable3." Which of these variables are predictive of the outcome? Interpret any significant effects. Make sure to be clear in this interpretation.

Attachment:- Assignment Files.rar