Distributions of Goodman and Kruskal's Gamma and Spearman's Rho in 2 × 2 Tables for Small and Moderate Sample Sizes

Gans, Lydia P.; Robertson, Craig

doi:10.1080/01621459.1981.10477745

Cited by 8 publications

(7 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For small to moderate sample sizes and the 2 ϫ 2 and 2 ϫ 3 cases, however, the distribution of sample G values has rather high variability and is irregular and skewed, resulting in a biased estimate of ␥. Moreover, the distribution of G converges very slowly to the ideal form reported by Goodman and Kruskal for the asymptotic case (Gans & Robertson, 1981a, 1981b; see also Lui & Cumberland, 2004;Rosenthal, 1966;Woods, 2007). Method.…”

Section: Simulation 3: Population ␥ D and Estimation Accuracymentioning

confidence: 62%

Sources of bias in the Goodman–Kruskal gamma coefficient measure of association: Implications for studies of metacognitive processes.

Masson¹,

Rotello²

2009

Journal of Experimental Psychology: Learning, Memory, and Cogni

140

143

View full text Add to dashboard Cite

In many cognitive, metacognitive, and perceptual tasks, measurement of performance or prediction accuracy may be influenced by response bias. Signal detection theory provides a means of assessing discrimination accuracy independent of such bias, but its application crucially depends on distributional assumptions. The Goodman-Kruskal gamma coefficient, G, has been proposed as an alternative means of measuring accuracy that is free of distributional assumptions. This measure is widely used with tasks that assess metamemory or metacognition performance. The authors demonstrate that the empirically determined value of G systematically deviates from its actual value under realistic conditions. A distribution-specific variant of G, called G-sub(c), is introduced to show why this bias arises. The findings imply that caution is needed when using G as a measure of accuracy, and alternative measures are recommended.

show abstract

Section: Simulation 3: Population ␥ D and Estimation Accuracymentioning

confidence: 62%

Sources of bias in the Goodman–Kruskal gamma coefficient measure of association: Implications for studies of metacognitive processes.

Masson¹,

Rotello²

2009

Journal of Experimental Psychology: Learning, Memory, and Cogni

140

143

View full text Add to dashboard Cite

show abstract

“…We then used these network-derived measures to conduct individual-based (ego) statistical analyses using SPSS VI in order to test the level and significance of association between a respondent's self-reported innovation and their Bsizes^and Bties^. We selected Goodman and Kruskal's (1954) Gamma to measure this association because of our small sample size, and use of ordinal variables (Gans and Robertson 1981).…”

Section: Discussionmentioning

confidence: 99%

Exploring the role of social capital in influencing knowledge flows and innovation in smallholder farming communities in the Caribbean

et al. 2016

View full text Add to dashboard Cite

This paper presents the results of an exploratory study into how different forms of social capital embedded within community-based social networks may affect innovation in smallholder farming systems to better support food security in the Caribbean. Focusing on two rural communities in the small island developing nation of Saint Lucia, our results indicate the strong presence of interpersonal agricultural knowledge networks operating to: 1) facilitate farmer-tofarmer knowledge exchange; 2) increase farmer access to information; and 3) connect farmers to sources of support. In both communities, 'peer farmers' were reported as being the primary source of new agricultural knowledge for farmers, with government 'extension officers' the secondary source. Comparative social network analysis reveals how different forms of social capital within the two agricultural knowledge networks can affect self-reported farmer innovation in different contexts. Based on these findings we identify a number of opportunities for policy initiatives to better support, coordinate and enhance innovation opportunities among smallholder farmers in the Caribbean with a view to building their adaptive capacity in the face of environmental change. The findings provide important evidence and insights relevant to the governance of domestic agricultural systems and regional food security programming in the Caribbean.

show abstract

“…In addition, we compute simple rank order correlations (Spearman's ρ). Also, we compute Goodman and Kruskal's gamma (Goodman & Kruskal, 1954) but note that it performs less well in small samples than Spearman's rank order correlation (Gans & Robertson, 1981). Finally, we make use of the concept of mutual information (MI; Brillinger, 2004).…”

Section: Discussionmentioning

confidence: 99%

Agreement of Teachers on Evaluating Assessments of Learning Progressions in English Language Arts and Mathematics

Rijn¹,

Graf

Arieli‐Attali

et al. 2018

ETS Research Report Series

View full text Add to dashboard Cite

In this study, we explored the extent to which teachers agree on the ordering and separation of levels of two different learning progressions (LPs) in English language arts (ELA) and mathematics. In a panel meeting akin to a standard‐setting procedure, we asked teachers to link the items and responses of summative educational assessments to LP levels. We evaluated four types of agreement among teachers, with test developers, and with trained raters. Although the results were quite mixed, agreement on the LP levels of mathematics items was generally better than that for ELA tasks. The agreement on the LP levels of ELA sample responses was generally better than that for mathematics. Implications of the results are discussed.

show abstract

Distributions of Goodman and Kruskal's Gamma and Spearman's Rho in 2 × 2 Tables for Small and Moderate Sample Sizes

Cited by 8 publications

References 7 publications

Sources of bias in the Goodman–Kruskal gamma coefficient measure of association: Implications for studies of metacognitive processes.

Sources of bias in the Goodman–Kruskal gamma coefficient measure of association: Implications for studies of metacognitive processes.

Exploring the role of social capital in influencing knowledge flows and innovation in smallholder farming communities in the Caribbean

Agreement of Teachers on Evaluating Assessments of Learning Progressions in English Language Arts and Mathematics

Contact Info

Product

Resources

About