Here, you are going to work with the data from the Kraut/Kiesler Internet study. The data in SPSS format are found here. Here is a brief codebook; most of the description of the data can be found in...

Here, you are going to work with the data from the Kraut/Kiesler Internet study. The data in SPSS format are found here. Here is a brief codebook; most of the description of the data can be found in the article itself.
For the actual analyses you will use both SPSS and a program called AMOS. This program is best known as a tool for structural equation modeling, and when we get to it.

Assignment:

1. Prepare and present appropriate descriptive statistics for the variables in the study. I trust that you remember how to do this. The key word is “appropriate” – that does NOT entail giving me the average gender of the participants, nor just cutting and pasting endless quantities of SPSS printout into your paper. SELECT what you want to present carefully, and EDIT it using Word or Excel or whatever to create a good-looking summary of the data. Use the article for a guide if you’re unclear on what you’re trying to make it look like.


2. Using SPSS REGRESSION, try to replicate the K/K analysis of the model predicting Depression at T3 from Internet use and a set of T1 variables. You can refer to this Guidance on Path Analysis in SPSS for assistance. Basically, you’re running a series of regressions dictated by the model, and selecting your coefficients from the output. Again, I don’t want to see all your output, just the usable numbers and a brief text description of how you got what you did. You may not get exactly the coefficients reported in the article, but your results should be pretty close (there are some minor data differences).


3. Now using AMOS, read in the K/K data and using the graphics tools, draw the full path model and estimate it. Refer to this Guidance on Path Analysis in AMOS for assistance if necessary. Interpret the results, remove from the model any non-significant paths, and re-estimate it. Repeat again if necessary. When you’re done, you should have only significantly non-zero paths in the model. Include your final model with its coefficients in your brief write-up of this process.