What is the difference between exploratory and confirmatory factor analysis

The aim of the present study was to use exploratory and CFA to investigate the factorial structure of the UWES in a multi-occupational sample of Swedish women. Confirmatory factor analysis was then performed using three different models: one-factor, two-factor, and three-factor. Goodness-of-fit statistics were obtained for all models and showed that none of them showed overall good fit, with RMSEA never going below 0.

As previously mentioned, a recent review of the factorial structure of the UWES showed inconclusive results, with some included studies showing best fit for a one-factor structure, some showing best fit for a three-factor structure, and some showing an equally good or poor fit for both Kulikowski, This indicates a need for further research into the underlying factors impacting the factor structures in various samples. One of the studies included in the Kulikowski review found that neither the one-factor nor the three-factor structure of the UWES-9 was a good fit for their data Wefald et al.

This used a sample similar to ours, both in terms of size vs. A previous study by Kulikowski has also attempted a two-factor structure, merging Dedication and Vigor into a single factor, letting Absorption constitute the second factor Kulikowski, Although this sample was Swedish, it was different from that of the present study in other significant ways, such as gender a majority were male and occupation all the participants were IT consultants, whilst ours was a multi-occupational sample , which may explain the differences in the results.

If our results are compared with those of other studies also using multi-occupational samples, several of them have, in agreement the Swedish study by Hallberg and Schaufeli , found that both the one-factor and three-factor structures may be used. For example, this was the case for Schaufeli et al. These differing results support the recommendation made by Kulikowski , namely that each study using the UWES-9 should undertake their own factor analysis based on their own sample, and make a decision on which structure to use based on their own results Kulikowski, In addition to this, and in agreement with the current study, several previous studies have found that none of the factor structures tested have shown an acceptable fit Hallberg and Schaufeli, ; Wefald et al.

Subsequently, researchers looking to use a measure of work engagement may wish to use another instrument in parallel with the UWES.

The present study has strengths, as well as weaknesses. The relatively large sample size of approximately women made it possible to randomly divide the group into half so that both an exploratory and a CFA could be undertaken.

The fact that the sample consisted exclusively of women may be seen both as a strength and as a weakness. On the one hand, it ensures that the results are not skewed by an uneven gender balance, but on the other hand our results should not be assumed to be generalizable to males.

An Iranian study investigating determinants of work engagement in hospital staff found no significant effect of gender Mahboubi et al. However, a Dutch study exploring work engagement and burnout in veterinarians found that women rated their work engagement lower than men, indicating that gender differences may vary with different occupational groups, nationalities, or other, hitherto unknown factors Mastenbroek et al. In addition to this, in terms of generalizability, it should be acknowledged that the sample used in the present study should be considered to represent the white-collar population, based on the higher-than-average level of education.

In addition to this, only Swedish-speaking girls participated. However, The present study used a large, multi-occupational female sample to explore the factorial structure of the UWES Despite indication from EFA that a one-factor structure best fit the data, we were unable to find good model fit for a one-, two-, or three-factor model using CFA.

Until such data exists, researchers would be wise to conduct their own factor analysis in order to determine whether the total score, the three dimensions representing Vigor, Dedication and Absorption, or even a two-factor structure is applicable for their sample. The datasets generated for this study are available on request to the corresponding author. At the time of the data collection for the present study, the participants were again asked to give their consent and reminded that their participation was voluntary, could be withdrawn any time without giving a reason, and that all information would be treated confidentially.

MW contributed to the conception and design of the work, performed the analyses, and drafted the manuscript. JW and ML contributed to the conception and design of the work, took part in the data collection and analyses, and revised the work critically.

All authors approved the final version to be published, and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Bakker, A. Weekly work engagement and performance: a study among starting teachers.

Bentler, P. Comparative fit indexes in structural models. Bollen, K. Structural Equations with Latent Variables. New York, NY: Wiley. Google Scholar. Christian, M. Work engagement: a quantitative review and test of its relation with task and contextual performance. IBM Corp SPSS for Windows.

Dziuban, C. When is a correlation matrix appropriate for factor analysis? Some decision rules. Gable, S. What and why is positive psychology? Burnout and work engagement: independent factors or opposite poles?

Hallberg, U. Can work engagement be discriminated from job involvement and organizational commitment? Ho Kim, W. Work engagement in South Korea.

Hooper, D. Structural equation modelling: guidelines for determining model fit. Methods 6, 53— Kulikowski, K. Do we all agree on how to measure work engagement? Factorial validity of Utrecht work engagement scale as a standard measurement tool — A literature review. SEM is used to find if relationships exist between these items and constructs structural model. Asked by: Saulius Ferragut asked in category: General Last Updated: 16th May, What is the difference between exploratory and confirmatory factor analysis?

Exploratory factor analysis is a method for finding latent variables in data, usually data sets with a lot of variables. Confirmatory factor analysis is a method of confirming that certain structures in the data are correct; often, there is an hypothesized model due to theory and you want to confirm it. Can SPSS do confirmatory factor analysis? How do you do a confirmatory factor analysis?

In order to identify each factor in a CFA model with at least three indicators, there are two options: Set the variance of each factor to 1 variance standardization method Set the first loading of each factor to 1 marker method. What does a factor analysis tell you? Factor analysis is a statistical method used to describe variability among observed, correlated variables in terms of a potentially lower number of unobserved variables called factors.

Factor analysis searches for such joint variations in response to unobserved latent variables. How many types of factors are there? How do you do exploratory factor analysis in SPSS? Assumptions: Variables used should be metric. Dummy variables can also be considered, but only in special cases. What is SEM data analysis?

Structural equation modeling is a multivariate statistical analysis technique that is used to analyze structural relationships. This technique is the combination of factor analysis and multiple regression analysis, and it is used to analyze the structural relationship between measured variables and latent constructs. Or you may have formulated a research question based on your theoretical understanding, and are now testing it.

Of course, in an exploratory factor analysis, the final number of factors is determined by your data and your interpretation of the factors. Cut-offs of factor loadings can be much lower for exploratory factor analyses. When you are developing scales, you can use an exploratory factor analysis to test a new scale, and then move on to confirmatory factor analysis to validate the factor structure in a new sample.

For example, a depression scale with the underlying concepts of depressed mood, fatigue and exhaustion, and social dysfunction can first be developed with a sample of rural US women using an exploratory factor analysis.

If you would like to next use that scale in a sample of urban US women, you would use a confirmatory factor analysis to validate the depression scale in your new sample. About the Author: Maike Rahn is a health scientist with a strong background in data analysis. Maike has a Ph. This is just a simple, yet a perfect explanation of Factor analysis.

Thank you for sharing the valuable information. Thank you very much for describing this in a clear and easily understood manner. Because of this, I may now actually finish my PhD! Respected Professor, Thank you very much for your kind clarification. The inputs given by you are simple and comprehensive. With warm regards Dr.

Your email address will not be published. Skip to primary navigation Skip to main content Skip to primary sidebar by Maike Rahn, PhD An important question that the consultants at The Analysis Factor are frequently asked is: What is the difference between a confirmatory and an exploratory factor analysis? Principal Component Analysis.