Dq week four two
Factor analysis is a quantitative technique that is used by researchers to identify basic patterns in the relationship patterns among variables. In other words, factor analysis is used to find out if the variables observed could be explained largely in terms of a smaller number of variables which are referred to as factors. Therefore, factor analysis is appropriate if the researcher is of the opinion that the dependent variable can be largely explained by fewer variables than the ones obtained during the research. Factor analysis determines the number of factors that are needed to explain the relationship pattern among variables, the nature of the identified factors, the extent to which the hypothesized factors explain the data that was observed and the value of the random variable for each of the observed variables.
An example of where factor analysis could be used is a research that is intended to construct a scale of employee job satisfaction. The researcher will first design a questionnaire with several questions related to employee job satisfaction in both verbal and numeric form. The interest of the researcher is the views of employees concerning underlying dimensions of employee job satisfaction. Normally, there would be few such dimensions that are of psychological state and cannot be measured directly. These underlying dimensions are what are referred to as factors. The factors are assumed to be linearly related to the score.
Assuming that the answers obtained using the questionnaires are condensed to only two underlying factors; pay represented by F1 and management represented by F2 plus a random element. Therefore, the employee satisfaction score will be linearly related to the two factors as follows;
Si = β0i + β1iF1 + β2iF2 + e
The error term, indicates that the hypothesized relationship is not exact. Separate equations could be written for all the questionnaires items. The coefficients β1i and β2i will probably not be the same for each of the items. These coefficients are referred to as factor loadings. An estimate of each loading is derived from a matrix that contains the intercorrelations of the observed scores obtained from a large number of respondents. Each respondent answers all the questions in the questionnaire and the matrix would contain all possible correlations between the pair of the questionnaire items. Factor analysis can be performed by most of the statistical programs including Statistical Package for the Social Services (SPSS). The pattern observed from the factor loadings is used to interpret the underlying factors. Factor loading of more than 0.7 is seen as an indicator of a close association between an observed item and a factor. Assuming in this example there are strong factor loading, then the researcher can conclude that there are two main dimensions to the level of employee job satisfaction that is; pay and management.
Christensen, L. B., Johnson, R. B., & Turner, L. (2011). Research Methods, Design, and Analysis. New Jersey: Prentice Hall .
Jackson, S. L. (2011). Research Methods and Statistics: A Critical Thinking Approach (4 ed.). London: Cengage Learning.
