Exploratory factor analysis Flashcards
What is a concept?
an abstraction of some sort of idea or phenomenon
What is a variable?
Something that can take on different values
What are different forms of validity?
- content validity
- construct validity
- face validity
- statistical conclusion validity
What is reliability?
the extent to which a measurement produces consistent results when repeated. All measurements are subject to random error. A measurement with low random error has high reliability.
What is exploratory factor analysis?
a class of procedures used for data reduction and summarization. is very important when questionnaire surveys are used
When is factor analysis used?
- to identify underlying dimensions, or factors that explain the correlations among a set of indicators
- to identify the factors that make up the variables for hypothesis testing
Designing a factor analysis requires a minimum of sample size. What is the minimum?
100 samples (5 observations per variable)
What is a correlation matrix?
a lower triangle matrix showing the simple correlations, r, between all possible pairs of variables included in the analysis
What is communalities?
amount of variance a variable shares with all other variables
What is Kaiser-Meyer-Olkin measure of sampling adequacy?
it explains how much one item’s variance overlaps with another item’s variance
What is Eigenvalue?
it represents the total variance explained by each factor
What is factor loadings?
correlations between the variables/the factors
When do we have construct validity?
When we have Convergent validity and Discriminant validity
What does it mean if we say that correlation = good?
variables correlate with what they should (convergent) and what they should not (discriminant)
How should the bivariate correlation matrix look like in an exploratory factor analysis?
we want HIGH correlations with a construct of LOW between independent constructs/variables
What is commonalities?
they represent the extent to which an item correlates with all other items. Should be above 0.50. if they are not, remove them and re-run the analysis. but above 0.7 is what we strive for
What is the Kaiser-Meyer-Olkin Measure of Sampling adequacy? (KMO)
KMO explains how much one item/variables variance overlaps with another item/variables variance. Value should be above 0.50
Why is KMO used?
It is used to test how suited data is for factor analysis
What is rotation? (in EFA)
any of several methods in factor analysis by which the researcher attempts to relate the calculated factors to the theoretical entities
What are the two basic types of rotation to choose from in EFA?
- orthogonal (assume uncorrelated factors)
- oblique (assume correlated factors)
Factor loadings should be above what number?
above 0.50.
Correlation should be above what number?
above 0.30
What is the anti-image correlation?
the measure of KMO for each indicator/variable is given in the diagonal. in the diagonal, items with value under 0.5 should perhaps be deleted
What is one of the most accepted methods to test Reliability?
Coefficient Alpha (Chronbach’s Alpha test). it tests how well the items in each variable relate to each other.
in general, the cutoff value is above 0.7 and accepted for the Chronbachs Alpha. (shows reliable variable)
