Reading...
Play button
Play button
Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
7 Cards in this Set
- Front
- Back
- 3rd side (hint)
|
Unidimensionality at 3 Levels
|
* Construct
* Items * Set of Items |
* Construct
- Unidimensional if multiple sets of n items taken from the domain of the construct and the partial correlations among each set are equal to zero * Items - Unidimensional if it is a measure of only one construct or latent factor * Set of Items - Unidimensional if the correlations among items can be accounted for by a single factor |
|
|
Significance of parameter estimates and related diagnostics
|
* Statistically significant loadings (t/zvalue>2.57) of .70 and above
* Composite reliability >.7 * AVE >.5 |
|
|
|
Standardized residuals and modification indices
|
* Examine standardized residuals > 2.57 (especially those > 4)
* Modification Indices (MI) > 3.84 * If MI indicate correlated error terms, there may be issues of multi- dimensionality -Construct has multiple dimensions -Presence of method bias * If MI indicate significant cross-loadings, issues with unidimensionalityof items (i.e. items load on more than one construct) |
|
|
|
Undimensionality assessment
|
By model fit and by examining correlated error terms and modification indices
|
|
|
|
CFA Guidelines
|
* Model convergence and acceptable range of parameters
-No negative error variances or completely standardized parameter estimates >1 * Fit indices - Absolute or Stand-Alone Fit Indices - GFI > .90, - AGFI > .80 (or AGFI > .90, in some citations) - RMSEA < .08 - Comparative or Incremental Fit Indices (compare to Null model) - CFI > .90 (> .95 according to Huand Bentler) - NNFI > .90 (> .95 according to Huand Bentler) - Parsimony Fit Indices (useful for model comparison) - AIC --close to zero better fit - PFI –higher the better (range 0-1) - and insignificant χ2, combined with significant t-values for item loadings. |
|
|
|
Discriminant Validity
|
INTERCONSTRUCT CORRELATIONS AND LOADINGS
* Items load more highly on factor than on other factors (loadings > cross-loadings) * Loadings above .70 desirable * Square root of AVE > inter-construct correlation (items share more variance with construct than with other constructs) |
|
|
|
Multitrait Multimethod
|
* The Reliability Diagonal (monotrait-monomethod)
- Estimates of the reliability of each measure in the matrix. * The Validity Diagonals (monotrait-heteromethod) - Correlations between measures of the same trait measured using different methods. * The Heterotrait-monomethod Triangles - These are the correlations among measures that share the same method of measurement. Note that what these correlations share is method, not trait or concept. If these correlations are high, you have common method bias. Heterotrait-HeteromethodTriangles These are correlations that differ in both trait and method. The MonomethodBlocks These consist of all of the correlations that share the same method of measurement. The HeteromethodBlocks These consist of all correlations that do not share the same methods. |
|
