Seminar Notice

                Statistical Laboratory
                 Iowa State University

DATE AND TIME:     Monday, September 15, 1997, 4:10 p.m.

PLACE:             319 Snedecor Hall

SPEAKER:           Brad J. Bushman
                   Department of Psychology
                   Iowa State University
                   
TITLE:             A Procedure for Combining Sample Standardized
                   Mean Differences and Vote Counts to Estimate the
                   Population Standardized Mean Difference in Fixed
                   Effects Models

                       ABSTRACT

    There are two general approaches to conducting a literature review:  the
narrative (or qualitative) approach and the meta-analytic (or quantitative)
approach.  In the traditional narrative review, the reviewer uses "mental
algebra" to integrate the findings from a collection of studies, and summarizes
the results in a narrative manner.  In the meta-analytic review, the reviewer
uses statistical procedures to integrate the findings from a collection of studi
es,
and summarizes the results using numerical effect-size estimates.  Traditional
narrative reviews are more likely than are meta-analytic reviews to depend on
the subjective judgments, preferences, and biases of the reviewer.  Missing
effect-size estimates pose a difficult problem in meta-analysis.  Conventional
procedures for dealing with this problem include discarding studies with
missing estimates and imputing single values for missing estimates (e.g., 0,
mean).  An alternative procedure, which combines effect-size estimates and
vote-counts, is proposed for handling missing estimates.  The combined
estimator has several desirable features:  (a) it uses all the information
available from studies in a meta-analysis, (b) it is consistent, (c) it is more
efficient than other estimators, (d) it has known variance, and (e) it gives
weight to all studies proportional to the Fisher information they provide.  The
combined procedure is the method of choice in a meta-analysis when some
studies do not provide enough information to compute effect-size estimates, but
do provide information about the direction or statistical significance of
results.
    
COFFEE: 3:45 p.m., 104 Snedecor Hall