At what sample size do correlations stabilize

By Felix D. Schönbrodt, Marco Perugini. 2013

2,022 citations recorded by the source corpus.

Subfields: Economics and Econometrics, Experimental and Cognitive Psychology, Statistical and Nonlinear Physics

View Paper: https://www.semanticscholar.org/paper/6815c19823c9af5b2e8b4b4fcb2f67c0056e0f05

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

In Experimental and Cognitive Psychology

In Statistical and Nonlinear Physics

Citations in the corpus

Reproducible brain-wide association studies require thousands of individuals (2022, positive): “In the revised manuscript we have clariﬁed the wording and made a distinction between this method of assessing “stabilization of effect size” as in Schönbrodt and Perugini, 2013 via bootstrapped subsampling and classical “analytic statistical power estimates” through G*Power.”
High Boredom Proneness and Low Trait Self-Control Impair Adherence to Social Distancing Guidelines during the COVID-19 Pandemic (2020, background): “We aimed for a sample size that allows stable and precise estimates of correlation coefficients [21].”
Performing High-Powered Studies Efficiently with Sequential Analyses (2014, positive): “A recent simulation by Schönbrodt and Perugini (2013) provides useful recommendations for the required sample size to yield reasonably stable effect sizes estimates that do not change considerably when additional participants are collected.”
Morality as a Basic Psychological Need (2018, result): “Further, given that the ralerting values tended to be quite large in Study 1 (i.e., rs > .65), we did not need large samples to obtain stable estimates of those coefficients in Study 2 (cf. Schönbrodt & Perugini, 2013).”
Powerful Pharma and Its Marginalized Alternatives?: Effects of Individual Differences in Conspiracy Mentality on Attitudes Toward Medical Approaches (2018, background): “Following recent recommendations regarding the minimum sample size to receive stable correlations (n = 250; Schönbrodt & Perugini, 2013), 401 participants from Germany completed an online survey.”
Building a Science of Individual Differences from fMRI. (2016, methodology): “(A) The ‘cone of confidence’ for correlation analysis, as in [139].”
The Song Is You (2016, methodology): “…et al., 2008; Dunn et al., 2011; George et al., 2007; Langmeyer, et al., 2012; Ren-tfrow & Gosling, 2003; Zweigenhaft, 2008), and considering the comparatively large sample size used, the correlation estimates we observed should be quite stable (Ioannidis, 2008; Scho¨nbrodt & Perugini, 2013).”
The reliability paradox: Why robust cognitive tasks do not produce reliable individual differences (2017, result): “Though our sample sizes are larger than many previous retest reliability studies of these tasks, it has been argued that samples approaching 250 are necessary for a stable estimate of the (Pearson’s) correlation effect size (Schonbrodt & Perugini, 2013).”
Registered Reports A Method to Increase the Credibility of Published Results (2014, result): “At the same time, effects size estimates in small samples – common in original research – can vary considerably and are more likely to elicit an exaggerated effect size than results from larger sample sizes (Schçnbrodt & Perugini, 2013).”
Perfectionism and narcissism: A meta-analytic review (2016, background): “Evidence suggests correlations do not stabilize until N > 250 (Schönbrodt & Perugini, 2013).”
Collective Narcissism Predicts Hypersensitivity to In–group Insult and Direct and Indirect Retaliatory Intergroup Hostility (2016, background): “In our next studies, we sought to increase the sample sizes to over 250 participants in light of new research suggesting that at this sample size, correlations stabilize (Schönbrodt & Perugini, 2013).”
Science Skepticism Across 24 Countries (2021, positive): “Our aim was to select at least 250 participants per country (Schönbrodt & Perugini, 2013).”
At what sample size do latent variable correlations stabilize? (2018, methodology): “Based on the results of their simulation, as well as balancing accuracy and confidence, Schönbrodt and Perugini (2013) suggested that sample sizes approaching 240 to 250 may be minimally appropriate for typical differential psychology research scenarios (i.e., r ≈ .20).(1) Although Schönbrodt and Pe”