Soonmook Lee

Exploratory factor analysis for small samples

Traditionally, two distinct approaches have been employed for exploratory factor analysis: maximum likelihood factor analysis and principal component analysis. A third alternative, called regularized exploratory factor analysis, was introduced recently in the psychometric literature. Small sample size is an important issue that has received considerable discussion in the factor analysis literature. However, little is known about the differential performance of these three approaches to exploratory factor analysis in a small sample size scenario. A simulation study and an empirical example demonstrate that regularized exploratory factor analysis may be recommended over the two traditional approaches, particularly when sample sizes are small (below 50) and the sample covariance matrix is near singular.

A Review of CEFA Software: Comprehensive Exploratory Factor Analysis Program

CEFA 3.02(Browne, Cudeck, Tateneni, & Mels, 2008) is a factor analysis computer program designed to perform exploratory factor analysis. It provides the main properties that are needed for exploratory factor analysis, namely a variety of factoring methods employing eight different discrepancy functions to be minimized to yield initial solutions, a variety of tests or fit indices for the factor solutions, a reasonable choice of rotation methods to yield derived solutions, and asymptotic standard errors for the loadings of rotated factor pattern matrices. It is distributed free in an English version. The functionality of CEFA was reviewed and compared with the factor analysis procedures provided in widely used statistical packages, SAS (SAS Institute Inc., 1999) and SPSS (SPSS Inc., 2006).

Confirmatory factor analysis of the Child Oral Health Impact Profile (Korean version)

Empirical support for the factor structure of the Child Oral Health Impact Profile (COHIP) has not been fully established. The purposes of this study were to evaluate the factor structure of the Korean version of the COHIP (COHIP-K) empirically using confirmatory factor analysis (CFA) based on the theoretical framework and then to assess whether any of the factors in the structure could be grouped into a simpler single second-order factor. Data were collected through self-reported COHIP-K responses from a representative community sample of 2,236 Korean children, 8-15 yr of age. Because a large inter-factor correlation of 0.92 was estimated in the original five-factor structure, the two strongly correlated factors were combined into one factor, resulting in a four-factor structure. The revised four-factor model showed a reasonable fit with appropriate inter-factor correlations. Additionally, the second-order model with four sub-factors was reasonable with sufficient fit and showed equal fit to the revised four-factor model. A cross-validation procedure confirmed the appropriateness of the findings. Our analysis empirically supported a four-factor structure of COHIP-K, a summarized second-order model, and the use of an integrated summary COHIP score.

Developing a Domain-Referenced Test given a Selection Ratio and a Cut-Off Score: An application of Item Response Theory

In the present study we developed a domain-referenced test consisting of four subtests with the application of one parameter logistic model of item response theory. Some conditions are required: 10%–15% of the test takers pass the test given a cut-off score of 80% correct. Two of the four subtests satisfied the above requirements. One of the other two showed a passing ratio higher than specified. The other showed a low degree of classification consistency. In regards to validity, we attempted to secure content-related validity. Also there is an evidence of supporting construct-related validity. The correlations among the four subtests were small enough to show adequate discrimination of constructs.