When to Choose Factor Analysis
Factor analysis is tool for understanding the structure or order of a certain phenomenon or phenomena. Since phenomena take place across space and time, they can have a pattern, even be independent of each other, or exhibit multiple patterns. A lot of PhD Thesis and Dissertations involve use of factor analysis in data analysis chapter.
For instance, we tend to associate a set of behaviours or a certain pattern of attitudes with business people and another set wth the farming community. Different economic systems exhibit different patterns of behaviour and even the idea of conflict manifests a pattern of elements, such as two or more parties and a clash or contradiction in goals. Even the weather exhibits distinct patterns.
Factor analysis can be used to deal with a great volume of measurements as well as subjective observations and organize them into specific patterns of occurring. The patterns become clear and it is then easier to establish relationships and equations. Factor analysis is a useful tool which can be made use of to understand a certain area, provide structure, map out unknown ideas, categorize or condense information, shed light on causal linkages, filter or reshape data, pinpoint equations, test and form theories, limit variables and arrive at conclusions. In case you wish to de-tangle linear relationships into distinct patterns, the factor analysis is a useful method since each pattern is made to appear as a factor which delineates a separate group of connected information.
Further, factor analysis is handy if you wish to condense a great volume of data or information and provide a more parsimonious summary. For instance, it can be difficult to deal either descriptively or objectively with 50 aspects each of 300 countries. Here data analysis can help in the organization, structuring and making lucid of such data by merely reducing them to their commonalities and mutual patterns. Even a glance at the condensed information, provided by factor analysis can give us an idea of the greater volume of data. Countries can be far more conveniently discussed and contrasted on the basis of financial growth, extent and dimensions of politics.
In case you have to deal with unorganized data, unfamiliar interdependencies, great volumes of both quantitative and qualitative variables or just bad data, factor analysis can be an extremely useful tool especially in the social sciences. Use it to manage more than a hundred variables at the same time, reduce the possibility of random errors or invalidity and simplify complicated equations into their basic premises.
However, factor analysis is not without its disadvantages. It is mathematically complex and includes very different and aspects, which need to be understood and applied. Its technological jargon is complex and includes terms such as simple structure, loadings, communality, as eigenvalues orthogonal, etc.
The findings or results of a factor analysis technique can take up many pages of a report, and consequently leave little space for an introduction to methodology or an explanation of terms. Further, factor analysis as a technique is usually not learnt by students in their formal training. Besides, even social scientists and policy makers have complained that the nature and the importance of the results of factor analysis are hard to understand.