Hello Norman,

Ideally, at the time that a study is being planned (and before data are collected), one would identify the specific research questions, develop or adapt the measure(s), and identify suitable statistical methods for handling the data.

If you could elaborate a bit on your query, specifically indicating: (a) your research questions; and (b) the nature of the scores from the survey measure's two parts (are these each akin to a nominal classification, or something else), you'll be more likely to get more focused recommendations for how to proceed with analysis.

As for study generalizability, that pretty much depends on how you selected your sample, and how confident you can be that the sample is representative of the intended target population.

Good luck with your work.

I think the next step should be to assess each of your scales, using coefficient alpha to determine their reliability. In SPSS click Analyze, then Scale, and then Reliability Analysis.

In making this recommendation, I am assuming that your scales are drawn from the existing literature, but if you have created them yourself, then you will also want to evaluate them using Exploratory Factor Analysis.

Dear Norman Snead ,

David L Morgan is right - you need to analyze the reliability of individual scales or of all your questions together. Please have a look at my study guides on scale reliability with small samples and exploratory factor analysis.

Pinpointing the exact statistical test(s) for you depends on the purpose of your analysis, the number of your variables, and the nature of your data. That being said, you may avail yourself of the decision table on the UCLA Website, fully cited below.

UCLA. (2016, August 22). Choosing the correct statistical test in SAS, Stata, SPSS and R. IDRE Stats—Statistical Consulting Web Resources. https://stats.idre.ucla.edu/other/mult-pkg/whatstat/

Good luck,

I fully appreciate each of your responses to my inquiry. Rest assured that prior to conducting the research a Pilot Study was conducted using a large group of very similar participants to the target group. Upon completion of the Pilot Study I conducted a Chronbach's Alpha on the results and received a 0.95 alpha score.

When I mentioned validity in my inquiry I was referring to making sure that all the statistical analyses were conducted and interpreted so as to be valid (correctly completed and correctly analyzed/interpreted) among my peers.

Some of you have asked for some very specific information as to my survey and Likert Scale statements and for reasons that should be obvious I will not be discussing those. I can say however, that the survey was 116 questions/statements (5 demographic questions, 25 5-point Likert Scale statements that determined one of 5 leadership styles as prominent, using a key, and the remaining 86 5-point Likert Scale statements were created to identify which one of the 2 models relating to one target leadership style was most closely followed). The survey required 20 minutes from each participant. I sent out 10,000 email requests to perspective participants nationally and all respondents were chosen in a highly random manner. I achieved a 6% return rate with a 77% completion rate. My sample size for generalizability was 332 which was exceeded.

With this additional information, can those who responded give me more specific information as to how to analyze the demographic information? (Would it simply be mean, median, mode so as to simply show the numbers in each of the 5 areas?)

The 25 statements which determined the dominant leadership style using a key. What statistical analyses are typically used on a 5-point Likert Scale? (Would it be central tendency, or mean, mode, median again, or some other test that frequency?) The highest score on 5 questions among the 5 areas reveals the dominant style.

Then there are the 43 5-point Likert Scale statements that point to one of the models and another 43 5-point Likert Scale statements that point to the other model. (Same question as before on which type of analyses to use?) The highest score between the two models reveals the dominant model.

Now that you have more information, can anyone help narrow my focus on which analyses to conduct using IBM SPSS for each of the areas?

Thank you for your input and time!

Norman