Juana Gómez Benito

Research Synthesis Coding and Conjectures

In meta-analyses the extraction and coding of information from primary research reports has to be completed in a competent way because these tasks implicate most of the decisions that determine the usefulness of thefinal product. The authors offer guidelines that make it more likely that high-quality information is reliably extracted and codedfrom primary research reports. These guidelines address issues ranging from the selection of items and construction of coding materials to sustaining reliability and vigilance across extended periods of coding. Thereafter; the authors note how the methodology of metaanalysis results in pressure to change the type of information that appears in primary research reports, and close by offering a few conjectures about the future of meta-analysis.

Estimates of the Internal Consistency of a Factorially Complex Composite.

In this study of 852 subjects between four and nine years of age an endeavor was made to estimate the degree of consistency among the elements comprising Borelli-Oldrons Performance Scale by taking into account two levels of analysis: item clusters and subtest clusters. On the first level, handling as variables the items of each of the subtests, alpha coefficients were calculated to estimate the internal consistency of each cluster. On the second level, in which variables constituted subtests that conformed to hypothesized dimensions, every one of the factors was confirmed in an earlier publication. The estimation procedure of multicomponent reliability put forward by Werts, Rock, Linn, and Joreskog (1978) was used. Given that this technique makes it possible to obtain consistency estimates of a composite scale made up of tests clustered in different factors, it was also applied to the full scale.