Fixed Effects vs. Random Effects Meta‐Analysis Models: Implications for Cumulative Research Knowledge

By J. Hunter, Frank L. Schmidt. 2000

842 citations recorded by the source corpus.

Subfields: Economics and Econometrics, Information Systems and Management, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/9d0da258a66cc36af3eb0fbd68d5192914681308

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Economics and Econometrics

1. Mike W.-L. Cheung
2. Andy Field
3. G. Banks
4. Peter Sedlmeier
5. Raphael Gillett
6. Juliane Eberth
7. In-Sue Oh
8. W. Viechtbauer

In Information Systems and Management

1. Juho Hamari
2. W. Viechtbauer
3. Emma García‐Meca
4. J. P. Sánchez‐Ballesta
5. John K. Kruschke
6. C. Schwalbe
7. N. Denson
8. Hichem Khlif

In Statistics and Probability

1. Mike W.-L. Cheung
2. J. Sánchez-Meca
3. Fulgencio Marín-Martínez
4. G. Banks
5. Raphael Gillett
6. C. Abraham
7. Martin Eisend
8. John G. Grundy

Citations in the corpus

• An other perspective on personality: meta-analytic integration of observers' accuracy and predictive validity. (2010, background): “Such random-effects models more accurately estimate confidence intervals around point estimates and better avoid Type I errors in detecting moderators than do fixed-effects models (Hunter & Schmidt, 2000; Schmidt, Oh, & Hayes, 2009).”
• A Meta-Analysis of Motivational Interviewing: Twenty-Five Years of Empirical Studies (2010, methodology): “A random effects model was used for all analyses, which is more conservative than fixed effects models and assumes that effect sizes are likely to vary across samples and populations (Hunter & Schmidt, 2000).”
• Fear and Loathing in Psychopaths: a Meta-Analytic Investigation of the Facial Affect Recognition Deficit (2011, positive): “In recent years, random effects models have become viewed as generally more valid (Hunter & Schmidt, 2000); however, the relative benefits of using a random versus fixed effects model appear to break down when examining small numbers of studies (≤ 30), and the conservative nature of random effects…”
• Do Curricular and Cocurricular Diversity Activities Influence Racial Bias? A Meta-Analysis (2009, methodology): “…conservative in nature, reducing the chances of committing a Type I error and are thus a more rigorous meta-analytic method (Hedges & Vevea, 1998; Hunter & Schmidt, 2000; National Research Council, 1992).2 The HLM random effects model in meta-analysis specifies two linked equations: a…”
• The Power of Feedback Revisited: A Meta-Analysis of Educational Feedback Research (2020, neutral): “However, by this definition, homogeneity of effect sizes within hypothesized moderator groups will occur rarely in real data, which means that fixed-effect models are rarely appropriate for real data in meta-analyses and random-effects models should be preferred (Hunter and Schmidt, 2000).”
• Weighting by Inverse Variance or by Sample Size in Random-Effects Meta-Analysis (2010, methodology): “…the scientific aims of a meta-analysis (Brockwell & Gordon, 2001; Erez, Bloom, & Wells, 1996; Field, 2001, 2003, 2005; Hedges & Vevea, 1998; Hunter & Schmidt, 2000; Kisamore & Brannick, 2008; National Research Council, 1992; Overton, 1998; Raudenbush, 1994; Sánchez-Meca &…”
• The Association of Board Independence and Ownership Concentration with Voluntary Disclosure: A Meta-analysis (2010, methodology): “The hypothesis of homogeneity will be rejected in many cases, so in order to limit Type I error rates we use a random effects model (Shadish and Haddock, 1994; Overton, 1998; Hunter and Schmidt, 2000) using S(2)r/k √ as the standard error in order to create a 95% confidence interval around the mean ”
• Do Self-Concept Interventions Make a Difference? A Synergistic Blend of Construct Validation and Meta-Analysis (2006, background): “The increased generalizability afforded by the random effects model has led numerous meta-analytic researchers to conclude that random effects models are more appropriate (e.g., Erez, Bloom, & Wells, 1996; Hunter & Schmidt, 2000).”
• Rejecting or Accepting Parameter Values in Bayesian Estimation (2018, positive): “This latter question is usually answered with advice always to use the random-effects model (e.g., Field, 2003; Hunter & Schmidt, 2000), but here we will consider a Bayesian hypothesis test because that approach was advocated by Scheibehenne, Gronau, et al. (2017).”
• A meta-analysis of gender roles in advertising (2010, background): “(homogeneity statistic) do indeed indicate that the total variance of the log odds of each variable is significantly higher (p Hunter and Schmidt 2000; Raudenbush 1994).”
• Meta-Analytic Studies of Identity Status and the Relational Issues of Attachment and Intimacy (2009, background): “Fixed effects models make the assumption that the population effect size is constant and, unless this assumption is met, the analyses will have inflated the Type I error and will report overly narrow confidence intervals (Hunter & Schmidt, 2000).”