Reference no: EM132374024

Factor Analysis Assignment -

Introduction - Many researchers feel frustration with factor analysis (FA) at one time or another. Unlike most other inferential tests that have an unambiguous solution (i.e., a clear statistical decision is indicated based on interpretation of the calculated p-value), factor analysis does not most of the time. The ideal solution of having unambiguous and salient loading of all variables (i.e., test items or questions) on only one factor is often not attainable. It is possible to reject the null hypothesis of unidimensionality but not have clean loadings for all factors. You would note this situation in your results. Consequently, there is a measure of subjectivity in interpretation of factor analysis results and answering questions such as:

What are the optimum number of factors given the statistics and related theory?

What name should I give to the factors?

What questions should I retain?

What questions should I delete?

Don't assume that this assignment has one and only one correct solution - FA usually doesn't operate that way.

In the present assignment you will use factor analysis procedures on a test instrument that measures help self-efficacy, that is, the belief that one has the ability to help effectively. Help self-efficacy is measured by 15 questions created by the authors of the instrument used in this assignment. Three categories of help are represented in these 15 questions: six questions measure empathetic types of helping, four questions measure informational types of helping, four questions measure instrumental (i.e., doing things) types of helping, and the fifteenth question is open ended to allow any additional type of help given to be inserted. You will conduct a factor analysis to determine if the results of testing a sample of 517 subjects yield the three categories of self-efficacy that were originally hypothesized by the instrument's creators: empathetic types of helping, informational types of helping, and instrumental types of helping.

As far as the instrument's questions go, and in particular the fifteenth question, you need not consider the meanings of the options respondents are asked to consider in their evaluation of each question (otherwise we would need to conduct a literature review on the efficacy construct in order to understand the items and choices as they relate to theory, which I would prefer to avoid in this course). Suffice it to say that although question number 15 is an open-ended question according to the instrument's authors, the authors also provided choices. One issue that you must wrestle with is the role of question number 15 in any solution you recommend (assuming you provide evidence to support multidimensionality).

Specific course competencies assessed by this assignment -

• Apply structural procedures using SPSS to analyze research data.
• Apply the APA style in writing the results section of a research report.

Task - Conduct a factor analysis by completing the tasks listed in the evaluation checklist using the efficacy dataset.

Tasks -

Provide informationally adequate descriptive statistics.

Identify the factor analysis hypothesis(es) in null form.

Provide the results of your evaluations of multivariate normality. Identify how you evaluated this assumption.

Create and include a scree plot. Interpret the results.

Provide the results of your factor analysis. Use Maximum Likelihood Factoring with the Direct Oblimin method of rotation. Provide a rationale for use of these methods in this situation.

Conclusion - In this assignment you conducted factor analysis. You will sometimes encounter these procedures when you conduct or read research, particularly research involving the construct validity of measurement instruments.

Note - "For the KMO and Bartlett's Test table, the Kaiser-Meyer-Olkin Measure of Sampling Adequacy should have a value of .871, and for the Goodness-of-fit Test table, the Chi-Square should have a value of 80.359 and a Sig. of .005. These values may vary depending on your selection for treating missing values - mine are based upon what should be the default of Missing Listwise."

Attachment:- Factor Analysis Assignment Files.rar