The essential guide to effect sizes : statistical power, meta-analysis, and the interpretation of research results
Citations
5,374 citations
Cites methods from "The essential guide to effect sizes..."
...To report this analysis, researchers could write in the procedure section that: “Twenty participants evaluated either Movie 1 (n = 10) or Movie 2 (n = 10). Participants reported higher evaluations of Movie 1 (M = 8.7, SD = 0.82) than Movie 2 (M = 7.7, SD = 0.95), F(1, 18) = 6.34, p = 0.022, η(2)p = 0.26, 90% CI [0.02, 0.48].” Whereas in a t-test, we compare two groups, and can therefore calculate a confidence interval for the mean difference, we can perform F-tests for comparisons between more than two groups. To be able to communicate the uncertainty in the data, we should still report a confidence interval, but now we report the confidence interval around the effect size. An excellent explanation of confidence intervals around effect size estimates for F-tests, which is accompanied by easy to use syntax files for a range of statistical software packages (including SPSS) is provided by Smithson (2001) 2....
[...]
...Although η2 is an efficient way to compare the sizes of effects within a study (given that every effect is interpreted in relation to the total variance, all η2 from a single study sum to 100%), eta squared cannot easily be compared between studies, because the total variability in a study (SStotal) depends on the design of a study, and increases when additional variables are manipulated. Keppel (1991) has recommended partial eta squared (η(2)p) to improve the comparability of effect sizes between studies, which expresses the sum of squares of the effect in relation to the sum of squares of the effect and the sum of squares of the error associated with the effect....
[...]
3,117 citations
Cites background from "The essential guide to effect sizes..."
...Readers may prefer to consult specialized statistics books addressing effect sizes (e.g., Cumming, 2012; Ellis, 2010; Grissom & Kim, 2005; Rosenthal et al., 2000)....
[...]
...…columns of a contingency table represent a predictor and a predicted variable, Goodman–Kruskal’s lambda (L) describes how much the prediction is improved by knowing the category for the predictor, a potentially useful description of the size of the effect (Ellis, 2010; Siegel & Castellan, 1988)....
[...]
...Rosnow and Rosenthal (1989), for example, illustrated how a very small effect relating to life-threatening situations, such as the reduction of heart attacks, is important in the context of saving lives on a worldwide basis (see Table 12 and Ellis, 2010)....
[...]
...They provide a description of the size of observed effects that is independent of the possibly misleading influences of sample size....
[...]
...A discussion of the rather confusing history of the chosen symbols for these statistics can be found in Ellis (2010)....
[...]
2,339 citations
Cites background from "The essential guide to effect sizes..."
...Ellis (2010) provided an accessible introduction to a range of ES measures....
[...]
1,681 citations
1,554 citations
References
40,195 citations
[...]
38,291 citations
37,989 citations
16,435 citations
[...]
16,383 citations