-------------------------------------------------------------------------------help parmhet_hettest_opts(Roger Newson) -------------------------------------------------------------------------------

Heterogeneity-test options forparmhetandparmiv

optionsDescription -------------------------------------------------------------------------chi2het(newvarname)Heterogeneity chi-squared statistic variabledfhet(newvarname)Heterogeneity degrees of freedom variablei2het(newvarname)HeterogeneityI-squared statistic variabletau2het(newvarname)Heterogeneity tau-squared statistic variablefhet(newvarname)HeterogeneityF-statistic variableresdfhet(newvarname)Heterogeneity residual degrees of freedom variablephet(newvarname)HeterogeneityP-value variable -------------------------------------------------------------------------

DescriptionThese options specify the names of generated variables, containing heterogeneity-test statistics for the input dataset, or for the by-group if the

by()option is specified. If specified forparmhet, these options refer to generated variables in theparmhetresultsset (see help forparmhet_resultsset), and cause the variable with the same name as the option to be renamed to the name specified by the option. If specified forparmiv, these options cause a new variable of the specified name to be added to the existing dataset in memory.

Options

chi2het(newvarname)specifies the name of an output variable containing the heterogeneity chi-squared statistic defined by Cochrane (1954).

dfhet(newvarname)specifies the name of an output variable containing the degrees of freedom for the heterogeneity chi-squared statistic defined by Cochrane (1954).

i2het(newvarname)specifies the name of an output variable containing the heterogeneityI-squared statistic defined by Higgins and Thompson (2002). This statistic is expressed on a percentage scale from 0 to 100, and denotes the percentage excess of the heterogeneity chi-squared statistic, compared to its mean value under the null hypothesis of no heterogeneity, specified by its degrees of freedom. If the chi-squared statistic is less than its degrees of freedom, then theI-squared statistic is zero.

tau2het(newvarname)specifies the name of an output variable containing the heterogeneity tau-squared statistic defined by Higgins and Thompson (2002). The tau-squared statistic is an estimate of the variance of the true population values of the estimated parameters, in the meta-population from which these populations are sampled. It is expressed in squared units of the input parameter estimates, or in squared log units of the input parameter estimates, if theeformoption is specified (see help forparmhet_basic_opts). If the chi-squared statistic is less than its degrees of freedom, then the tau-squared statistic is zero.

fhet(newvarname)specifies the name of an output variable containing the heterogeneityF-statistic defined by Welch (1951) and popularized by Cochrane (1954). This variable is only calculated if the user specifies an input degrees of freedom variable, in addition to the input estimate and standard error variables. If an input degrees of freedom variable is not provided, then thefhet()option is ignored.

resdfhet(newvarname)specifies the name of an output variable containing the residual (or denominator) degrees of freedom for the heterogeneityF-statistic output to thefhet()variable. This denominator degrees of freedom variable may have non-integer values, and is used, together with the numerator degrees of freedom output to thedfhet()variable, to calculate aP-value (output to thephet()variable) for the heterogeneityF-statistic (output to thefhet()variable). Theresdfhet()variable is only calculated if the user specifies an input degrees of freedom variable, in addition to the input estimate and standard error variables. If an input degrees of freedom variable is not provided, then theresdfhet()option is ignored.

phet(newvarname)specifies the name of an output variable containing the heterogeneityP-value. If an input degrees of freedom variable is specified, then thisP-value is calculated using theF-statistic output to thefhet()variable, with the numerator degrees of freedom output to thedfhet()variable and the denominator degrees of freedom output to theresdfhet()variable. If an input degrees of freedom variable is not specified, then thisP-value is calculated using the chi-squared statistic output to thechi2het()variable, with the degrees of freedom output to thedfhet()variable.

AuthorRoger Newson, National Heart and Lung Institute, Imperial College London, UK. Email: r.newson@imperial.ac.uk

ReferencesCochrane, W. G. 1954. The combination of estimates from different experiments.

Biometrics10(1): 101-129.Higgins, J. P. T. and Thompson, S. G. 2002. Quantifying heterogeneity in a meta-analysis.

Statistics in Medicine21(11): 1539-1558.Welch, B. L. 1951. On the comparison of several mean values: an alternative approach.

Biometrika36(3/4): 330-336.

Also seeManual:

[R] meta,[R] testOn-line: help forparmhet,parmiv,parmhet_basic_opts,parmhet_resultsset_opts,parmhet_resultssethelp fortesthelp forparmest,parmby,parmcip,metaparm,metanif installed