Factor Analysis Flashcards
What can exploratory factor analysis be used for (2)?
Effectively extracting information from large bodies of interrelated data
Defining the underlying structure among the variables in the analysis
When to use exploratory technique over confirmatory factor analysis?
When researcher has little control over the specifics of the structure
What are the (7) stages of applying factor analysis?
Clarifying the objectives of factor analysis
Designing a factor analysis
Assumptions of exploratory factor analysis
Deriving factors and assessing overall fit
Rotating and interpreting the factors
Validation of exploratory factor analysis solutions
Additional uses of exploratory factor analysis results
What is the difference between R and Q in factor analysis?
R: factors are calculated from the correlation
Q: factors are calculated from the individual respondent
What is the main difference between principal component analysis and common factor analysis?
PCA considers total variance, whereas for CFA unique and error variance are not of interest in defining the structure of the variables.
How to determine the numbers of factors to extract (2)?
Combine a conceptual foundation with empirical evidence
Number of empirical stopping criteria for deciding the factors to extract
What is meant by rotation of factors?
Redistribute the variance from earlier factors to later ones to achieve a simpler, theoretically more meaningful factor pattern.
Which variables are considered more important in factor analysis?
Variables with higher loadings
What might the presence of cross-loading indicate?
Deletion of that variable since it does not represent simple structure and complicates the naming process
What are the (3) major limitations of exploratory factor analysis?
Controversy over which technique is the best
The subjective nature of EFA are all subject to many difference in opinion
The problem of reliability is real
What is the purpose of EFA?
Estimate a model which explains variance/covariance between a set of observed variables (in a population) by a set of (fewer) unobserved factors and weightings.
How do you decide which variables to include (3)?
Past research and theory
Measurement properties
Sample size
When to use CFA rather than EFA?
When testing a hypothesis
How to pass KMO measure of sampling frequency?
At least above 0.5, the closer to 1, the better.
How to pass Bartletts test of sphericity?
Smaller than 0.05
How to assess factor loadings (3)?
Minimal between 0.4 and 0.5
Significant > 0.5
Desirable > 0.7
How can the number of factors be determined (5)?
A priori determination
Based on eigenvalues
Based on screeplot
Based on 0.6>% of variance
Based on split-half reliability
Why rotate the axes?
To make the location of the axes fit the actual data points better.
What is the difference between varimax and oblimin?
Axes are maintained at right angles vs axes are NOT maintained at right angles.
What are factor scores?
Composite measures of each factor for each respondent.
How to create surrogate variables?
Select for each factor the variable with the highest loading.
What are the advantages of summated scores (3)?
Represents multiple aspects in one variable
Easier to use for prediction-oriented research
Reducing measurement error
How to determine model fit (2)?
Via residuals
1. Reproduce correlations between variables from estimated correlations
2. Compare difference between observed and reproduced correlations
Which value of Cronbachs alpha is accepted in reliability analysis?
0.7 or more
How to assess the construct validity of a measurement model (4)?
Convergent validity
Discriminant validity
Face validity
Nomological validity
How to select a candidate for deletion in CFA? (2)
Modification indices
Relatively high residual indicates a covariance for which the proposed model underperforms
What are the (5) generic steps in CFA?
Model specification
Identification
Estimation
Testing fit
Respectification
How to estimate parameters in CFA (4)?
Maximum likelihood
Unweighted least squares
Generalized least squares
Distribution free
What is meant by comparative fit measures?
Comparison between the current model and a reference model.
How does statistics inference work?
A significant test result indicates that the implied covariance matrix differs significantly from the imperial covariance matrix. Thus, model is wrong.
What is the danger of model respectification?
Overfitting
