Structural Equation Modeling: Strengths, Limitations, and Misconceptions

Department of Psychology, Vanderbilt University, Nashville, Tennessee 37203, USA.
Annual Review of Clinical Psychology (Impact Factor: 12.67). 02/2005; 1(1):31-65. DOI: 10.1146/annurev.clinpsy.1.102803.144239
Source: PubMed


Because structural equation modeling (SEM) has become a very popular data-analytic technique, it is important for clinical scientists to have a balanced perception of its strengths and limitations. We review several strengths of SEM, with a particular focus on recent innovations (e.g., latent growth modeling, multilevel SEM models, and approaches for dealing with missing data and with violations of normality assumptions) that underscore how SEM has become a broad data-analytic framework with flexible and unique capabilities. We also consider several limitations of SEM and some misconceptions that it tends to elicit. Major themes emphasized are the problem of omitted variables, the importance of lower-order model components, potential limitations of models judged to be well fitting, the inaccuracy of some commonly used rules of thumb, and the importance of study design. Throughout, we offer recommendations for the conduct of SEM analyses and the reporting of results.

111 Reads
  • Source
    • "Secondly, unlike traditional multivariate analyses that ignore errors, SEM estimates errors of variance parameters. Thirdly, Structural Equation Modelling (SEM) makes it possible for researchers to estimate relations among constructs that are corrected for bias attributable to random error and construct-irrelevant variance by providing separate estimates of relations among latent constructs and their manifest indicators (Tomarken & Waller, 2005). "
    [Show abstract] [Hide abstract]
    ABSTRACT: Understanding the psychological factors related to waste-prevention behaviours of university students could enable local governments and policy makers to craft effective policies to reduce waste. This study utilised a questionnaire-based survey to assess the associations between spiritual intelligence and Machiavellianism as factors that influence waste-prevention behaviours. A total of 210 participants from Universiti Putra Malaysia completed the questionnaires, including the demographic questions, spiritual intelligence inventory, Mach IV, and the waste-prevention behaviours. The data were analysed using Structural Equation Modelling (SEM). The results showed that individuals with higher spiritual intelligence and lower Machiavellianism were more likely to report a positive attitude towards waste-prevention behaviours, and that an inverse association existed between spiritual intelligence and Machiavellianism. Overall, these variables explained 12.0% of the variance in waste-prevention behaviours. Therefore, these findings reinforce the importance of personality traits and cognitive abilities in waste-prevention behaviours.
    Pertanika Journal of Social Science and Humanities 09/2015; 23(3):623-638.
  • Source
    • "As college graduation varied systematically by mothers' education, gender, and race/ethnicity, the characteristics of the missing data in this study were problematic (Cooper et al. 2005). Multiple imputation (MI), which yields less biased results than listwise deletion (Tomarken and Waller 2005), was used to mitigate these missing data concerns. MI involves creating several samples (for the current analyses, five) that estimate all missing values based on the existing patterns in the data, resulting in several ''complete'' data sets. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Critics of emerging adulthood theory have suggested that it only applies to college students, but this assertion has largely gone untested. The purpose of the present study was to compare developmental trajectories of non-students versus college-educated youth in theoretically relevant domains of work, love, and financial independence. Using data from the Youth Development Study (N = 1139, 49.6 % female, 63.3 % White, 10.9 % Southeast Asian, 1.5 % Other Asian, 8.6 % Black, 5.3 % Mixed Race, 4.0 % Latino, 0.8 % Native American), latent growth curve models were fitted to chart each group's development, from ages 14 to 30. Different trajectories were revealed for hours worked, children, and financial dependence on parents, spouses, and government aid. No differences were found in employment rates, marriage rates, or financial dependence on own income. These results provide a clearer picture of emerging adulthood for non-students, and highlight problems with generalizing college student research to all emerging adults.
    Journal of Youth and Adolescence 07/2015; 44(11). DOI:10.1007/s10964-015-0330-0 · 2.72 Impact Factor
    • "were not significant. SEM is an advantageous analytical approach because it allows the simultaneous test of multiple mediators and more complex mediating chains (as we intended to test in our study) and also provides global model fit indices useful to estimate the general accuracy of comprehensive models (Tomarken & Waller, 2005). Nevertheless, to further check the robustness of our analyses, we used a resampling procedure and individually tested hypotheses 2 and 3 using bootstrapping as recommended by Preacher and Hayes (2004). "
    [Show abstract] [Hide abstract]
    ABSTRACT: Using a cross-lagged design, the present study tests an integrative model of emergent collective emotions in learning groups. Our results indicate that the percentage of women in the group fosters the emergence of collective emotional intelligence, which in turn stimulates social integration within groups (increases group cohesion and reduces relationship conflict) and the associated affective similarity, with beneficial effects for group effectiveness.
    British Journal of Psychology 04/2015; 106(2):217–234. DOI:10.1111/bjop.12075 · 2.37 Impact Factor
Show more


111 Reads
Available from