Scribd Logo
Upload

Welcome to Scribd!

  • 76 views

    Converting Among Effect Sizes

    Uploaded by

    Leonidas Ioannou
    Effect Size

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Converting Among Effect Sizes

    Uploaded by

    Leonidas Ioannou
    76 views5 pages
    Effect Size

    Original Title

    Effect Size

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Effect Size

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    76 views5 pages

    Converting Among Effect Sizes

    Uploaded by

    Leonidas Ioannou
    Effect Size

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    You are on page 1of 5

    CHAPTER 7

    Converting Among Effect Sizes

    Introduction
    Converting from the log odds ratio to d
    Converting from d to the log odds ratio
    Converting from r to d
    Converting from d to r

    INTRODUCTION
    Earlier in this Part we discussed the case where different study designs were used
    to compute the same effect size. For example, studies that used independent groups
    and studies that used matched groups were both used to yield estimates of the
    standardized mean difference, g. There is no problem in combining these estimates
    in a meta-analysis since the effect size has the same meaning in all studies.
    Consider, however, the case where some studies report a difference in means,
    which is used to compute a standardized mean difference. Others report a difference
    in proportions which is used to compute an odds ratio. And others report a correla-
    tion. All the studies address the same broad question, and we want to include them
    in one meta-analysis. Unlike the earlier case, we are now dealing with different
    indices, and we need to convert them to a common index before we can proceed.
    The question of whether or not it is appropriate to combine effect sizes from
    studies that used different metrics must be considered on a case by case basis. The
    key issue is that it only makes sense to compute a summary effect from studies that
    we judge to be comparable in relevant ways. If we would be comfortable combining
    these studies if they had used the same metric, then the fact that they used different
    metrics should not be an impediment.
    For example, suppose that several randomized controlled trials start with the
    same measure, on a continuous scale, but some report the outcome as a mean
    and others dichotomize the outcome and report it as success or failure. In this
    case, it may be highly appropriate to transform the standardized mean differences

    Introduction to Meta-Analysis. Michael Borenstein, L. V. Hedges, J. P. T. Higgins and H. R. Rothstein


    © 2009 John Wiley & Sons, Ltd. ISBN: 978-0-470-05724-7
    46 Effect Size and Precision

    Binary data Continuous data Correlational data

    Standardized
    Log odds ratio Mean Difference Fisher’s z
    (Cohen’s d )

    Bias-corrected
    Standardized
    Mean Difference
    (Hedges’ g)

    Figure 7.1 Converting among effect sizes.

    and the odds ratios to a common metric and then combine them across studies.
    By contrast, observational studies that report correlations may be substantially
    different from observational studies that report odds ratios. In this case, even if
    there is no technical barrier to converting the effects to a common metric, it may be
    a bad idea from a substantive perspective.
    In this chapter we present formulas for converting between an odds ratio and d, or
    between d and r. By combining formulas it is also possible to convert from an odds
    ratio, via d, to r (see Figure 7.1). In every case the formula for converting the effect
    size is accompanied by a formula to convert the variance.
    When we convert between different measures we make certain assumptions
    about the nature of the underlying traits or effects. Even if these assumptions do
    not hold exactly, the decision to use these conversions is often better than the
    alternative, which is to simply omit the studies that happened to use an alternate
    metric. This would involve loss of information, and possibly the systematic loss of
    information, resulting in a biased sample of studies. A sensitivity analysis to
    compare the meta-analysis results with and without the converted studies would
    be important.
    Figure 7.1 outlines the mechanism for incorporating multiple kinds of data in the
    same meta-analysis. First, each study is used to compute an effect size and variance
    of its native index, the log odds ratio for binary data, d for continuous data, and r for
    correlational data. Then, we convert all of these indices to a common index, which
    would be either the log odds ratio, d, or r. If the final index is d, we can move from
    there to Hedges’ g. This common index and its variance are then used in the
    analysis.
    Chapter 7: Converting Among Effect Sizes

    CONVERTING FROM THE LOG ODDS RATIO TO d


    We can convert from a log odds ratio (LogOddsRatio) to the standardized mean
    difference d using
    pffiffiffi
    3
    d 5 LogOddsRatio  ; ð7:1Þ
    p
    where p is the mathematical constant (approximately 3.14159). The variance of d
    would then be
    3
    Vd 5 VLogOddsRatio  ; ð7:2Þ
    p2
    where VLogOddsRatio is the variance of the log odds ratio. This method was originally
    proposed by Hasselblad and Hedges (1995) but variations have been proposed
    (see Sanchez-Meca, Marin-Martinez, & Chacon-Moscoso, 2003; Whitehead,
    2002). It assumes that an underlying continuous trait exists and has a logistic
    distribution (which is similar to a normal distribution) in each group. In practice,
    it will be difficult to test this assumption.
    For example, if the log odds ratio were LogOddsRatio 5 0.9069 with a variance
    of VLogOddsRatio 5 0.0676, then
    pffiffiffi
    3
    d 5 0:9069  5 0:5000
    3:1416
    with variance
    3
    Vd 5 0:0676  5 0:0205:
    3:14162

    CONVERTING FROM d to the log odds ratio


    We can convert from the standardized mean difference d to the log odds ratio
    (LogOddsRatio) using
    p
    LogOddsRatio 5 d pffiffiffi ; ð7:3Þ
    3
    where p is the mathematical constant (approximately 3.14159). The variance of
    LogOddsRatio would then be
    p2
    VLogOddsRatio 5 Vd : ð7:4Þ
    3
    For example, if d 5 0.5000 and Vd 5 0.0205 then
    3:1416
    LogOddsRatio 5 0:5000  pffiffiffi 5 0:9069;
    3
    48 Effect Size and Precision

    and
    3:14162
    VLogOddsRatio 5 0:0205  5 0:0676:
    3
    To employ this transformation we assume that the continuous data have the logistic
    distribution.

    CONVERTING FROM r TO d
    We convert from a correlation (r) to a standardized mean difference (d) using
    2r
    d 5 pffiffiffiffiffiffiffiffiffiffiffiffiffi : ð7:5Þ
    1  r2
    The variance of d computed in this way (converted from r) is
    4Vr
    Vd 5 : ð7:6Þ
    ð1  r 2 Þ 3

    For example, if r 5 0.50 and Vr 5 0.0058, then


    2  0:50
    d 5 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 5 1:1547
    1  0:502
    and the variance of d is
    4  0:0058
    Vd 5  3 5 0:0550:
    1  0:502

    In applying this conversion we assume that the continuous data used to compute r
    has a bivariate normal distribution and that the two groups are created by dichot-
    omizing one of the two variables.

    CONVERTING FROM d TO r
    We can convert from a standardized mean difference (d) to a correlation (r) using
    d
    r 5 pffiffiffiffiffiffiffiffiffiffiffiffiffi ð7:7Þ
    2
    d þa
    where a is a correction factor for cases where n1 ¼
    6 n2 ,
    ðn1 þ n2 Þ 2
    a5 : ð7:8Þ
    n1 n2
    The correction factor (a) depends on the ratio of n1 to n2, rather than the
    absolute values of these numbers. Therefore, if n1 and n2 are not known precisely,
    use n1 5 n2, which will yield a 5 4. The variance of r computed in this way
    (converted from d) is
    Chapter 7: Converting Among Effect Sizes

    a2 V d
    Vr 5 : ð7:9Þ
    ð d 2 þ aÞ 3

    For example, if n15 n2, d 5 1.1547 and vd 5 0.0550, then


    1:1547
    r 5 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 5 0:5000
    1:15472 þ 4
    and the variance of r converted from d will be
    42  0:0550
    Vr 5  3 5 0:0058:
    1:15472 þ 4

    In applying this conversion assume that a continuous variable was dichotomized to


    create the treatment and control groups.
    When we transform between Fisher’s z and d we are making assumptions about
    the independent variable only. When we transform between the log odds ratio and d
    we are making assumptions about the dependent variable only. As such, the two sets
    of assumptions are independent of each other, and one has no implications for the
    validity of the other. Therefore, we can apply both sets of assumptions and trans-
    form from Fisher’s z through d to the log odds ratio, as well as the reverse.

    SUMMARY POINTS
     If all studies in the analysis are based on the same kind of data (means, binary,
    or correlational), the researcher should select an effect size based on that kind
    of data.
     When some studies use means, others use binary data, and others use correla-
    tional data, we can apply formulas to convert among effect sizes.
     Studies that used different measures may differ from each other in substantive
    ways, and we need to consider this possibility when deciding if it makes sense
    to include the various studies in the same analysis.

    You might also like