Factor Analysis in Research

Home Forums Default Forum Factor Analysis in Research

Viewing 5 posts - 1 through 5 (of 5 total)
  • Author
    Posts
  • #5352

    I lke to have a detailed discussion on “Factor Analysis in Research”. I request this forum to discuss Factor analysis and its applications in research with practical examples and actual cases of your live projects. How to identify factors? How to do Reliability and Validity Test? How to make questionnare for collecting data? How to do factor analysis with the data?

    #5356
    Sunny Bose
    Member

    George,

    Factor Analysis is done to find out the underlying dimensions or factors that actually contributes to a number of observed attributes. It can be used in any research for data reduction. The funda of factor analysis is actually:

    Factor analysis is a statistical approach that can be used to analyze interrelationships among a large number of variables and to explain these variables in terms of their common underlying dimensions (factors).
    The statistical approach involves finding a way of condensing the information contained in a number of original variables into a smaller set of dimensions (factors) with a minimum loss of information.

    By the lloks of your question, I believe you would require the Exploratory Factor Analysis. There is also a Confirmatory Factor Analysis, but I doubt you require that right now. So I am sticking to exploratory one.

    Exploraatory Factor Analysis or simple factor analysis is applied to:
    1.Identification of Underlying Factors:
    clusters variables into homogeneous sets
    creates new variables (i.e. factors)
    allows us to gain insight to categories
    2. Screening of Variables:
    identifies groupings to allow us to select one variable to represent many
    useful in regression (recall collinearity)

    Let say, you are measuring satisfaction from service recovery and you find that to measure it you have some twenty odd items or attributes. Now if you run a statistical test based on those items, it becomes cumbersome. So what you do is you run a factor analysis to see whether some of these items have a common latent factor or in a more common man’s language a coomon root. If it is so, then it can be said that (lets say five items can be clubbed together under one factor) the factor is representing those original items and therefore, now your quationnaie or test items get reduced by four with minimum loss of information.

    When you run a factor analysis you might even find that those twenty original items are actually getting represented by six or seven factors and it is not surprising. Therefore, factor analysis is one of the most popular data reduction tools.

    Now, coming to its technicalities, this analysis uses correlations among the items and items that show the highest correlation among each other is generally grouped together. Note that a single original variable can have more than one factor explaining that variable and it is called communality. The portions of the different original variables that cannot be explained by any factor at all are the idependent terms and have to be considered as they are.

    For a starter, let me suggest that on runnig an analysis, group together all those items whose values you consider to be falling in one group. In that way, you would get factors. As you go on further churning the items, the factors would lesson further. However, ensure that in the effort of reducing data you donot loose infirmation ie. you go on reducing the items and reducing factors,then the correlations and compactness of items grouping under the latent factor also reduces.

    Another thing is that before running a factor analysis, consult the literature for the items you choosing for your research. Literatuee support is necessary. In case, you donot have extant literature for the research, its always better you conduct a focus group and/or take expert opinion for the initial items. To run the analysis, its better to use 5 point or 7 point Likert scale, though Semantic- Diefferential can also be used.

    Finally, consult the book Multivariate Data Analysis by Hair, Black, Anderson, Babin and tatham and Marketing Research by Naresh K Malhotra. They would help you immensly to understand this technique.

    #5355

    Hi Sunny Boss, Thank you very much. Your comment is very useful to me and may be useful for many researchers

    #5354
    mita dixit
    Member

    Hi George,

    I have used factor analysis for my doctorate research recently. FA gave me good results on groping several variables into factors in line with the theory and the conceptual framework. But the critical aspect is of calculating factor scores and forming new factors if they have to be used for further analysis.

    #5353
    Amir Foroughi
    Participant

    dear George what kind of factor analysis would u like discuss CFA  or EFA?

Viewing 5 posts - 1 through 5 (of 5 total)
  • The forum ‘Default Forum’ is closed to new topics and replies.