Effect size

From Wikiversity

Home

Survey design

Descr/ Graphs

Correl- ation

EFA

Psycho- metrics

Qual.

MLR

ANOVA

Power

Effect size

Review

In statistics, an effect size is a measure of the strength of the relationship between two variables. When reporting statistical significance, it is generally also recommend to also report measure(s) of effect size.

Contents 1 Types 2 Effect sizes in SPSS 3 Data analysis exercises 4 See also 5 External links

[edit] Types

Some common types of effect size are:

• r or r² (bivariate linear correlation or correlation squared)
• R or R² (multiple correlation or coefficient of determination)
• Cohen's d, Hedges g, or other forms of standard deviation unit effect size
• η² and η (total variance explained and partial variance explained in ANOVA; each IV will have a partial eta-squared; total eta-squared is equivalent to R-squared)

[edit] Effect sizes in SPSS

For SPSS users, note that:

• Cohen's d, etc. are not available in SPSS (use a spreadsheet calculator such as Cohensd.xls) instead. The basic forumula for Cohen's d is:

• η² is not available in SPSS (this can be calculated as shown in the Francis SPSS lab manual in the independent t-test section and in the appendices of the sample lab report). To do: Add details here.

For more information, see the effect size article at Wikipedia.

[edit] Data analysis exercises

• Effect size data analysis tutorial

[edit] See also

• Effect size (Wikipedia)
• Meta-analysis
• Statistical power

[edit] External links

• Neill, J. T. (2008). Why use Effect Sizes instead of Significance Testing in Program Evaluation? wilderdom.com.