Limits...
Meta-analytic evidence of low convergence between implicit and explicit measures of the needs for achievement, affiliation, and power.

Köllner MG, Schultheiss OC - Front Psychol (2014)

Bottom Line: Studies from a comprehensive search in PsycINFO, data sets of our research group, a literature list compiled by an expert, and the results of a request for gray literature were examined for relevance and coded.Participant age did not moderate the size of these relationships.However, a greater proportion of males in the samples and an earlier publication year were associated with larger effect sizes.

View Article: PubMed Central - PubMed

Affiliation: Human Motivation and Affective Neuroscience Lab, Department of Psychology and Sport Sciences, Institute of Psychology, Friedrich-Alexander University Erlangen-Nürnberg (FAU) Erlangen, Germany.

ABSTRACT
The correlation between implicit and explicit motive measures and potential moderators of this relationship were examined meta-analytically, using Hunter and Schmidt's (2004) approach. Studies from a comprehensive search in PsycINFO, data sets of our research group, a literature list compiled by an expert, and the results of a request for gray literature were examined for relevance and coded. Analyses were based on 49 papers, 56 independent samples, 6151 subjects, and 167 correlations. The correlations (ρ) between implicit and explicit measures were 0.130 (CI: 0.077-0.183) for the overall relationship, 0.116 (CI: 0.050-0.182) for affiliation, 0.139 (CI: 0.080-0.198) for achievement, and 0.038 (CI: -0.055-0.131) for power. Participant age did not moderate the size of these relationships. However, a greater proportion of males in the samples and an earlier publication year were associated with larger effect sizes.

No MeSH data available.


“Funnel-Plot” of single correlations (A) and of study correlations corrected for sampling error (B) for identification of a publication bias.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4126572&req=5

Figure 4: “Funnel-Plot” of single correlations (A) and of study correlations corrected for sampling error (B) for identification of a publication bias.

Mentions: A publication bias can result in overestimation of effect sizes. Like Hofmann et al. (2005), we used a graphical method. A funnel plot detects a publication bias by showing the absence of small effects for studies with small Ns, because such effects would not become significant. Regardless of N, the plots in Figure 4 feature no obvious absence of correlations close to zero. Usually the variance of the correlations in small studies was substantially larger than in studies with large samples. In large samples with more than 300 subjects, values are centered closely around 0.00. The resulting impression of an inverted funnel is desired for a funnel-plot which shows no publication bias (Hunter and Schmidt, 2004; Roberts et al., 2007). But while our funnel plot suggests that a publication bias was unlikely across the entire publication time span of the included studies, in light of the negative association between publication year and the relationship we cannot rule out that a publication bias has existed at some point in implicit motive research.


Meta-analytic evidence of low convergence between implicit and explicit measures of the needs for achievement, affiliation, and power.

Köllner MG, Schultheiss OC - Front Psychol (2014)

“Funnel-Plot” of single correlations (A) and of study correlations corrected for sampling error (B) for identification of a publication bias.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4126572&req=5

Figure 4: “Funnel-Plot” of single correlations (A) and of study correlations corrected for sampling error (B) for identification of a publication bias.
Mentions: A publication bias can result in overestimation of effect sizes. Like Hofmann et al. (2005), we used a graphical method. A funnel plot detects a publication bias by showing the absence of small effects for studies with small Ns, because such effects would not become significant. Regardless of N, the plots in Figure 4 feature no obvious absence of correlations close to zero. Usually the variance of the correlations in small studies was substantially larger than in studies with large samples. In large samples with more than 300 subjects, values are centered closely around 0.00. The resulting impression of an inverted funnel is desired for a funnel-plot which shows no publication bias (Hunter and Schmidt, 2004; Roberts et al., 2007). But while our funnel plot suggests that a publication bias was unlikely across the entire publication time span of the included studies, in light of the negative association between publication year and the relationship we cannot rule out that a publication bias has existed at some point in implicit motive research.

Bottom Line: Studies from a comprehensive search in PsycINFO, data sets of our research group, a literature list compiled by an expert, and the results of a request for gray literature were examined for relevance and coded.Participant age did not moderate the size of these relationships.However, a greater proportion of males in the samples and an earlier publication year were associated with larger effect sizes.

View Article: PubMed Central - PubMed

Affiliation: Human Motivation and Affective Neuroscience Lab, Department of Psychology and Sport Sciences, Institute of Psychology, Friedrich-Alexander University Erlangen-Nürnberg (FAU) Erlangen, Germany.

ABSTRACT
The correlation between implicit and explicit motive measures and potential moderators of this relationship were examined meta-analytically, using Hunter and Schmidt's (2004) approach. Studies from a comprehensive search in PsycINFO, data sets of our research group, a literature list compiled by an expert, and the results of a request for gray literature were examined for relevance and coded. Analyses were based on 49 papers, 56 independent samples, 6151 subjects, and 167 correlations. The correlations (ρ) between implicit and explicit measures were 0.130 (CI: 0.077-0.183) for the overall relationship, 0.116 (CI: 0.050-0.182) for affiliation, 0.139 (CI: 0.080-0.198) for achievement, and 0.038 (CI: -0.055-0.131) for power. Participant age did not moderate the size of these relationships. However, a greater proportion of males in the samples and an earlier publication year were associated with larger effect sizes.

No MeSH data available.