Intro
Effect Size Determination Program
Last Updated 07/25/2001
by
David B. Wilson, Ph.D.
University of Maryland, College Park
Introduction
The Effect Size Determination Program is for use with the book Practical Meta-Analysis
written by Mark W. Lipsey and David B. Wilson and published by Sage. The creation of this
program was supported by HSRI.
This program computes standardized mean difference effect sizes (d) and the correlation
coefficients (r) from summary statistics, such as means and standard deviations, t-tests,
frequencies, etc. It is useful during the coding phase of a meta-analysis for converting
reported results into an effect size index.
To use, "click" on the "Main Menu" button with the mouse. From the Main Menu, select the
desired transformation, again by "clicking" on the approrpiate button. Enter the requested
data in all cells that are shaded in yellow. The effect size is computed automatically when
all necessary data cells are filled in.
Page 1
Intro
Page 2
Intro
Page 3
Intro
Page 4
Menu
Main Menu
Page 5
Menu
Page 6
Menu
Page 7
Means & SDs
Means and Standard Deviation, Post-Treatment Scores
Treatment Group
Mean =
SD =
n=
Comparison Group
Mean =
SD =
n=
d =
r =
Page 8
Means & SDs
Page 9
Means & SDs
Page 10
Means & t-test
Means and t-test, Post-Treatment Scores
Treatment Group
Mean =
n=
t-value =
Comparison Group
Mean =
n=
d =
r =
Page 11
Means & t-test
Page 12
Means & t-test
Page 13
Mean Gain Scores
Mean Gain Scores and Gain Score Standard Deviation
Treatment Group
Mean =
SD =
n=
Comparison Group
Mean =
SD =
n=
Pre & Post-Test Scores Correlation =
d =
r =
Page 14
Mean Gain Scores
Page 15
Mean Gain Scores
Page 16
t-test (Independent)
Independent t-test, No Means Reported
t-value =
Treatment n =
Comparsion n =
d =
r =
Page 17
t-test (Independent)
Page 18
t-test (Independent)
Page 19
t-test (p-value only)
Independent t-test, p-value only
p-value =
df =
d=
r =
Note: This will always return a positive value
Page 20
t-test (p-value only)
Page 21
t-test (p-value only)
Page 22
t-test (Dependent)
Dependent t-test, No Means Reported
t-value =
n (pairs) =
r for paired values =
d =
r =
Page 23
t-test (Dependent)
Page 24
t-test (Dependent)
Page 25
Proportions1
Proportion or Frequency (Dichotomous)
Frequency with Positive Outcome
Treatment Group Successes =
Treatment Group n =
Control Group Successes =
Control Group n =
Proportion with Positive Outcome
Treatment Group =
Control Group =
Logit Method
d =
r =
Probit Method
d =
r =
Logit Method
d =
r =
Probit Method
d =
r =
Logit = logged odds-ratio
Page 26
Proportions1
Page 27
Proportions1
Page 28
Proportions2
Proportions (Ordinal Polychotomous)
Treatment Group
Values
Percent
0
1
2
3
4
5
6
7
8
9
Comparison Group
Values
Percent
0
1
2
3
4
5
6
7
8
9
Page 29
Tx Group n =
Comp Grp n =
Tx Mean =
Tx SD =
Comp Mean =
Comp SD =
d=
r=
Proportions2
Page 30
Proportions2
Page 31
Frequencies
Frequencies (Ordinal Polychotomous)
Treatment Group
Values
Frequency
0
1
2
3
4
5
6
7
8
9
Comparison Group
Values
Frequency
0
1
2
3
4
5
6
7
8
9
Page 32
Tx Mean =
Tx SD =
Comp Mean =
Comp SD =
d=
r =
Frequencies
Page 33
Frequencies
Page 34
F-test (k=2)
Oneway ANOVA with Two Groups (k=2)
F-value =
Treatment n =
Comparison n =
d =
r =
Note: This will always return a positive value
Page 35
F-test (k=2)
Page 36
F-test (k=2)
Page 37
F-test (k>2)
Oneway ANOVA with Three or More Groups (k>2)
Means
Indicate
Group Type*
MSw =
Method 1:
d=
r=
F-value =
Method 2:
d=
r=
Group Type: 1 = treatment group, -1 = comparison group, 0 = other
Page 38
F-test (k>2)
Page 39
F-test (k>2)
Page 40
Oneway ANCOVA
Oneway Analysis of Covariance
Treatment Group
Mean =
MS error =
df error =
r* =
Comparison Group
Mean =
d=
r =
r* = correlation between the covariate and the dependent measure
Page 41
Oneway ANCOVA
Page 42
Oneway ANCOVA
Page 43
Chi-Square
Chi-Square (df=1)
Chi-Square =
Total N =
d=
r =
Note: This will always return a positive value
Page 44
Chi-Square
Page 45
Chi-Square
Page 46
Chi-Square (p-value only)
Chi-Square, p-value only (df=1)
p-value =
Total N =
d=
r =
Note: This will always return a positive value
Page 47
Chi-Square (p-value only)
Page 48
Chi-Square (p-value only)
Page 49
Correlation
Correlation and Mean Difference Effect Size Conversion
r = 0.23
d = 0.4727
d = 0.47
r = 0.2288
Page 50
Correlation
Page 51
Correlation
Page 52
Weighted Mean
Weighted Mean
Means
Weighted Mean =
Page 53
Weighted Mean
Page 54
Weighted Mean
Page 55
Pooled SD
Pooled SD
SDs
Pooled SD =
Page 56
Pooled SD
Page 57
Pooled SD
Page 58
