Business Research Method Factor Analysis

Factor analysis is a general name denoting a class of procedures primarily used for data reduction and summarization. It helps in reducing the number of variables being studied to a smaller number by combining related ones into factors. Factor is an underlying dimension that accounts for several observed variables Factor analysis is an interdependence technique in that an entire set of interdependent relationships is examined without making the distinction between dependent and independent variables.

Factor Analysis Factor analysis is used in the following circumstances: –To identify latent or underlying factors, from an array of seemingly important variables by analysing correlations between variables –To identify a new, smaller, set of uncorrelated factors to replace the original set of correlated variables in subsequent multivariate analysis (regression or discriminant analysis). –To identify a smaller set of salient ( Surrogate) variables from a larger set for use in subsequent multivariate analysis.

Factor Analysis Evaluate credit card usage & behavior of customers. Initial set of variables is large: Age, Gender, Marital status,Income, education,employment status, credit history& family background Reduction of 9 variables in 3 factors : --Demographic Characteristic (Age, Gender, Marital status) --Socio-economic Status (Income, education, employment status ) --Background status (credit history& family background)

Height Weight Occupation Education Source of Income Size Social Status Factor Analysis Copyright © 2000 Harcourt, Inc. All rights reserved.

Factor Analysis Evaluate buying behavior of customers for a two wheeler Initial set of variables is large: Affordable, Sense of freedom, Economical, Man’s Vehicle, Feel Powerful, Friends Jealous, Feel good to see Ad of My Brand, Comfortable ride, Safe travel, Ride for three to be allowed Reduction of 10 variables to 3 factors through factor analysis --Pride(Man’s Vehicle,Feel Powerful, Friends Jealous, Feel good to see Ad of My Brand,) --Utility( Economical, Comfortable ride, Safe travel) --Economy (Affordable, Ride for three to be allowed)

Factor Analysis Factors determining buying behaviour of small cars Factors determining choice of an airlines Factors leading to cigarette smoking Underlying dimensions for willingness to donate regenerative & non regenerative body parts Factors determining choice of a bank

Factor Analysis Two Stages in Factor Analysis ----Factor Extraction ----Factor rotation

Factor Extraction Determines Number of factors to be extracted Factors are linear combinations of original variables Maximum number of factors equals no. of variables Purpose is to reduce variables to fewer no. of factors Popular method is Principal Component Analysis. Based on the Concept of Eigen Value Higher the eigen value of the factor, higher is the amount of variance explained by the factor Extract least number of factors to explain maximum variance

Factor Analysis-Extraction Each original variable has Eigen value =1due to standardization Only factors with eigen value >= 1 are retained Factors with eigen value < 1 are no better than a single variable The number of factors extracted is determined so that cumulative % of variance extracted reaches a satisfactory level ( at least 60% )

Factor Analysis-Extraction Scree plot. A scree plot is a plot of the Eigen values against the number of factors in order of extraction. The Shape of the plot is used to determine the number of factors The plot has a distinct break between steep slope of factors with large eigen values & a gradual trailing off associated with rest of the factors The gradual trailing off is referred to as Scree The point at which scree begins denotes the No. of factors Generally number of factors determined by scree plot is 1or 2 more than determined by eigen values

Scree Plot Component Number Eigenvalue

Factor rotation After extraction the next task is to interpret & name the factors This is done by identifying which factors are associated with which original variables The factor matrix is used for this purpose The original factor matrix is unrotated & comes as output of stage I The rotated factor matrix comes as output of stage II when we request the computer package to perform rotation & give us a rotated factor matrix The popular method of rotation is Orthogonal(varimax )

Although the initial or unrotated factor matrix indicates the relationship between the factors and individual variables, it seldom results in factors that can be interpreted, because the factors are correlated with many variables. Therefore, through rotation the factor matrix is transformed into a simpler one that is easier to interpret. In rotating the factors, we would like each factor to have nonzero, or significant, loadings or coefficients for only some of the variables. Rotation does not affect communalities& % of total variance explained. However % of variance accounted for by each factor does change In factor rotation smallest loadings tend towards 0 & largest loadings tend towards 1. The rotation is called orthogonal rotation if the axes are maintained at right angles. Factor rotation

The factor matrix ( whether unrotated or rotated ) gives us the loadings of each variable on each of the extracted factors This is similar to correlation matrix with loadings having values between o to 1 Values close to 1 represent high loadings & close to 0 low loadings The objective is to find variables which have a high loading on one factor low loadings on other factors. If Factor 1 is loaded highly by variables,say, 3.6 & 10,then, it is assumed that Factor 1 is a linear combination of variables 3,6 & 10 It is given a suitable name representing essence of original variables (3,6 & 10)

Example To determine benefits consumer seeks from purchase of a toothpaste A sample of 30 respondents was interviewed Respondents were asked to indicate their degree of agreement with the following statements using a 7 point scale(1=Strongly agree,7= Strongly disagree) V1:Important to buy a toothpaste that prevents cavities V2:Like a toothpaste that gives shiny teeth V3:A toothpaste should strengthen your gums V4:Prefer toothpaste that freshens breath V5:Prevention of tooth decay is not an important benefit V6:The most important consideration is attractive teeth Data obtained are given in the next slide

Conducting Factor Analysis

Formulate the Problem The objectives of factor analysis should be identified. The variables to be included in the factor analysis should be specified based on past research, theory, and judgment of the researcher. It is important that the variables be appropriately measured on an interval or ratio scale. An appropriate sample size should be used. As a rough guideline, there should be at least four or five times as many observations (sample size) as there are variables.

The analytical process is based on a matrix of correlations between the variables. Bartlett's test of sphericity can be used to test the null hypothesis that the variables are uncorrelated in the population: in other words, the population correlation matrix is an identity matrix. Rejection of this hypothesis indicates the appropriateness of factor analysis Another useful statistic is the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy.It is an index used to examine the appropriateness of factor analysis. High values (between 0.5 and 1.0) indicate factor analysis is appropriate. Values below 0.5 imply that factor analysis may not be appropriate Construct the Correlation Matrix

Correlation Matrix

Tests For Appropriateness of FACTOR ANALYSIS Barlett test of sphericity Approx. Chi-Square = df = 15 Significance = Kaiser-Meyer-Olkin measure of sampling adequacy = 0.660

Interpretation of correlation matrix For factor analysis to be appropriate variables must be correlated Barlett test of sphericity H0: Correlation matrix is unit matrix Approx. Chi-Square = , df = 15, Significant at 5% level, p-value.0000<.05: Reject Ho: Variables in the matrix are correlated Kaiser-Meyer-Olkin measure of sampling adequacy = >.5: Sample is adequate Both tests indicate appropriateness of factor analysis

In principal components analysis, the total variance in the data is considered. The diagonal of the correlation matrix consists of unities, and full variance is brought into the factor matrix. Principal components analysis is recommended when the primary concern is to determine the minimum number of factors that will account for maximum variance in the data for use in subsequent multivariate analysis. The factors are called principal components. Determine the Method of Factor Analysis

Results of Principal Components Analysis

.

A Priori Determination. Sometimes, because of prior knowledge, the researcher knows how many factors to expect and thus can specify the number of factors to be extracted beforehand. Determine the Number of Factors

Determination Based on Eigenvalues.In this approach, only factors with Eigenvalues greater than 1.0 are retained. An Eigenvalue represents the amount of variance associated with the factor. Hence, only factors with a variance greater than 1.0 are included. Factors with variance less than 1.0 are no better than a single variable, since, due to standardization, each variable has a variance of 1.0

Determination Based on Percentage of Variance. In this approach the number of factors extracted is determined so that the cumulative percentage of variance extracted by the factors reaches a satisfactory level. It is recommended that the factors extracted should account for at least 60% of the variance. Determine the Number of Factors

Determination Based on Scree Plot. A scree plot is a plot of the Eigenvalues against the number of factors in order of extraction. Experimental evidence indicates that the point at which the scree begins denotes the true number of factors. Generally, the number of factors determined by a scree plot will be one or a few more than that determined by the Eigenvalue criterion. Determine the Number of Factors

Scree Plot Component Number Eigenvalue Fig 19.2

A factor can then be interpreted in terms of the variables that load high on it from rotated factor matrix FACTOER I has high coefficients for V1:Important to buy a toothpaste that prevents cavities V3:A toothpaste should strengthen your gums V5:Prevention of tooth decay is not an important benefit FACTOR I may be labelled as health benefit FACTOR II has high coefficients on V2:Like a toothpaste that gives shiny teeth V4:Prefer toothpaste that freshens breath V6:The most important consideration is attractive teeth FACTOR II may be labelled as aesthetic factor Interpret Factors

Factor Loading Plot Another useful aid in interpretation is to plot the variables, using the factor loadings as coordinates. Variables at the end of an axis are those that have high loadings on only that factor, and hence describe the factor.

Factor Loading Plot Component 2  Component 1 Component Variable 1 2 V V V V V V Component Plot in Rotated Space      V1 V3 V6 V2 V5 V4 Rotated Component Matrix

By examining the factor matrix, one could select for each factor the variable with the highest loading on that factor. That variable could then be used as a surrogate variable for the associated factor. However, the choice is not as easy if two or more variables have similarly high loadings. In such a case, the choice between these variables should be based on theoretical and measurement considerations. In our example all 3 variables V1,V3 and V5 have high loadings on F1,the highest being V1. But if prior knowledge suggests that V5 is important,it could be selected as surrogate variable. Similarly V6 could be selected as surrogate for F2 Thus future analysis could be done with only 2 variables V5 (Tooth Decay) & V6(Attracive teeth) Select Surrogate Variables

The lower left triangle contains the reproduced correlation matrix; the diagonal, the communalities; the upper right triangle, the residuals between the observed correlations and the reproduced correlations. Results of Principal Components Analysis.

Conducting Factor Analysis Construction of the Correlation MatrixMethod of Factor AnalysisDetermination of Number of Factors Determination of Model Fit Problem formulation Calculation of Factor Scores Interpretation of Factors Rotation of Factors Selection of Surrogate Variables

The factors can be expressed as linear combinations of the observed variables. F i = W i1 X 1 + W i2 X 2 + W i3 X W ik X k where F i =estimate of i th factor W i =weight or factor score coefficient k =number of variables Factor Analysis Model

It is possible to select weights or factor score coefficients so that the first factor explains the largest portion of the total variance. Then a second set of weights can be selected, so that the second factor accounts for most of the residual variance, subject to being uncorrelated with the first factor. This same principle could be applied to selecting additional weights for the additional factors. Factor Analysis Model

Statistics Associated with Factor Analysis Bartlett's test of sphericity. Bartlett's test of sphericity is a test statistic used to examine the null hypothesis that the variables are uncorrelated in the population. In other words, the population correlation matrix is an identity matrix; each variable correlates perfectly with itself (r = 1) but has no correlation with the other variables (r = 0) A large value of test statistics favors rejection of null hypothesis & factor analysis is meaningful. Correlation matrix. A correlation matrix is a lower triangle matrix showing the simple correlations, r, between all possible pairs of variables included in the analysis. The diagonal elements, which are all 1, are usually omitted.

Factor: Factor is an underlying dimension that accounts for several observed variables Eigenvalue. The eigen value represents the total variance explained by each factor. Factor loadings. Factor loadings are simple correlations between the variables and the factors. Factor loading plot. A factor loading plot is a plot of the original variables using the factor loadings as coordinates. Factor matrix. A factor matrix contains the factor loadings of all the variables on all the factors extracted. Communality.. This is the proportion of variance explained by the common factors for each variable.It shows how much of each variable is accounted for by underlying factors taken together. It equals sum of squares of factor loadings for that variable. It ranges from 0 to 1. In factor analysis the sum of the initial communality values of variables will be equal to total number of variables considered for analysis Statistics Associated with Factor Analysis

Factor scores. Factor scores are composite scores estimated for each respondent on the derived factors. Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy. The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy is an index used to examine the appropriateness of factor analysis. High values (between 0.5 and 1.0) indicate factor analysis is appropriate. Values below 0.5 imply that factor analysis may not be appropriate. Percentage of variance. The percentage of the total variance attributed to each factor. Residuals are the differences between the observed correlations, as given in the input correlation matrix, and the reproduced correlations, as estimated from the factor matrix. Scree plot. A scree plot is a plot of the Eigenvalues against the number of factors in order of extraction. Statistics Associated with Factor Analysis

Appendix

A Priori Determination. Sometimes, because of prior knowledge, the researcher knows how many factors to expect and thus can specify the number of factors to be extracted beforehand. Determination Based on Eigenvalues. In this approach, only factors with Eigenvalues greater than 1.0 are retained. An Eigenvalue represents the amount of variance associated with the factor. Hence, only factors with a variance greater than 1.0 are included. Factors with variance less than 1.0 are no better than a single variable, since, due to standardization, each variable has a variance of 1.0. If the number of variables is less than 20, this approach will result in a conservative number of factors. Determine the Number of Factors

Determination Based on Scree Plot. A scree plot is a plot of the Eigenvalues against the number of factors in order of extraction. Experimental evidence indicates that the point at which the scree begins denotes the true number of factors. Generally, the number of factors determined by a scree plot will be one or a few more than that determined by the Eigenvalue criterion. Determination Based on Percentage of Variance. In this approach the number of factors extracted is determined so that the cumulative percentage of variance extracted by the factors reaches a satisfactory level. It is recommended that the factors extracted should account for at least 60% of the variance. Determine the Number of Factors

Scree Plot Component Number Eigenvalue Fig 19.2

Determination Based on Split-Half Reliability. The sample is split in half and factor analysis is performed on each half. Only factors with high correspondence of factor loadings across the two subsamples are retained. Determination Based on Significance Tests. It is possible to determine the statistical significance of the separate Eigenvalues and retain only those factors that are statistically significant. A drawback is that with large samples (size greater than 200), many factors are likely to be statistically significant, although from a practical viewpoint many of these account for only a small proportion of the total variance. Determine the Number of Factors

Although the initial or unrotated factor matrix indicates the relationship between the factors and individual variables, it seldom results in factors that can be interpreted, because the factors are correlated with many variables. Therefore, through rotation the factor matrix is transformed into a simpler one that is easier to interpret. In rotating the factors, we would like each factor to have nonzero, or significant, loadings or coefficients for only some of the variables. The rotation is called orthogonal rotation if the axes are maintained at right angles. Rotate Factors

The most commonly used method for rotation is the varimax procedure. This is an orthogonal method of rotation that minimizes the number of variables with high loadings on a factor, thereby enhancing the interpretability of the factors. Orthogonal rotation results in factors that are uncorrelated. The rotation is called oblique rotation when the axes are not maintained at right angles, and the factors are correlated. Sometimes, allowing for correlations among factors can simplify the factor pattern matrix. Oblique rotation should be used when factors in the population are likely to be strongly correlated. Rotate Factors

A factor can then be interpreted in terms of the variables that load high on it. Another useful aid in interpretation is to plot the variables, using the factor loadings as coordinates. Variables at the end of an axis are those that have high loadings on only that factor, and hence describe the factor. Interpret Factors

By examining the factor matrix, one could select for each factor the variable with the highest loading on that factor. That variable could then be used as a surrogate variable for the associated factor. However, the choice is not as easy if two or more variables have similarly high loadings. In such a case, the choice between these variables should be based on theoretical and measurement considerations. Select Surrogate Variables

SPSS Windows To select this procedures using SPSS for Windows click: Analyze>Data Reduction>Factor …