Response to Another Student Discussion-Factor Analysis
DIRECTIONS: PLEASE READ THE FOLLOWING FROM ANOTHER STUDENT AND:
Identify an element of the learner's posting that applies to your own understanding of factor analysis or your own research plans and elaborate on the point with specific concerns or examples. Please cite references.
Robert Laukaitis
U04D1 - Factor Analysis - R. Laukaitis
Top of Form
U04D1 - Factor Analysis
When considering learning about factor analysis, it was particularly helpful to distinguish where the underlying assumptions of factor analysis differ from other statistical methods (e.g., ANOVA, t tests, correlation, etc.) studied thus far. In a way, factor analysis has played a part in the formulation of variables being considered for my dissertation.
In my DRP, I describe variables to be considered. The variables represent latent variables for unobservable phenomena that the organization captures as proxy indicators for organization effectiveness. For example, where Warner (2013) describes latent variables as those not directly observable, the organization for which this research will focus uses similar indicators that might exist in complex sets of data. The complexity of my organization drives the organization to summarize multitudes of data into few indicators of organization effectiveness. Where factor analysis could contribute to the process is in the fundamental ability of the approach to consider large volumes of data and propose groupings based on statistical approaches. Yet, the process begins with correlations being generated for each possible pair for each variable. The logic here being that factors evolve from the loading within the construct being studied (Field, 2009).
Warner (2013) suggested when considering using factor analysis, the selection and measurement of variables representing a given domain must occur. Without this step, it would be mute to attempt to define a construct. Next, data should be screened in order to validate that there is no missing or miscoded data. This step helps prepare the data for the correlation matrix. Once the correlation matrix is formed, factor extraction begins. This process uses sophisticated mathematical approaches to determine factors for extraction. Next, factor rotation is computed and interpreted in order to increase the visibiliy of factors within the construct. Once the factors have been extracted, interpretation of factors based on their contributing components can be conducted (Field, 2009; Warner, 2013).
Most interesting is the ability of exploratory factor analysis to help understand primary constructs (Warner, 2013). For example: while my organization has developed or is using unobservable latent variables to determine organization effectiveness, it could be incredibly useful to explore the underlying data in order to see if any new or different factors emerge from unknown constructs. Since primary component analysis uses a correlation matrix as the baseline indicator of potential factor groups, this approach could be useful in determining common threads within the organization’s effectiveness indicator array (George & Mallery, 2013).
Factor analysis could provide useful insight into the complex data present in my organization. Indicators could also be added to or improved in order to help organization leaders make decisions based on better constructed indicators of unobservable organization information.Â
References
Field, A. (2009). Discovering statistics using SPSS (3rd ed.). Thousand Oaks, CA: Sage.
George, D., & Mallery, P. (2013). IBM statistics 21 step by step: A simple guide and reference (13th ed.). Pearson. Retrieved from http://online.vitalsource.com/books/9781269627795
Warner, R. M. (2013). Applied statistics: From bivariate through multivariate techniques (2nd ed.). Thousand Oaks, CA: Sage. Retrieved from http://online.vitalsource.com/books/9781452268705
Bottom of Form
