Tag | Value |
---|---|
file | Factor-analysis_uu-factor-analysis-009-en_uu-factor-analysis-009-en |
name | uu-factor-analysis-009-en |
section | Factor analysis |
type | schoice |
solution | FALSE, TRUE, FALSE, FALSE |
Type | Conceptual |
Language | English |
Level | Statistical Literacy |
A factor analysis is performed on fifteen items designed to measure fear of St. Nicholas in children. Principal component analysis with oblimin rotation results in two statistically and substantively interesting dimensions. Factor A concerns fear of Sinterklaas and factor B concerns fear of Black Pete. Evaluate the following two statements on factor analysis.
I. In general, factors with an Eigenvalue greater than 1 explain more test variance than an individual item. II. If an item (after rotation) has a high factor loading (|a| > .40) on factor A, factor A is strongly related to factor B.
Theorem 1: Eigenvalue is calculated by:
For each item, how much variance of that item is explained by the factor
This is converted into a proportion (for example: .60 (=60%) of the item is explained by the factor)
The proportion of all items are added together to determine the Eigenvalue of a factor.
Thus, each item has a variance of 1.00. When a factor has an Eigenvalue >1.00, the factor explains more variance than an individual item.
Theorem 2: The factor loading of one item on one of the factors says nothing about the correlation between the factors.