I cannot shed any light on how "systematic review" and "meta-analysis"
developed separate meanings - that is, I don't know the precise history with
references. However, the why part of this question seems fairly clear.
After carrying out the systematic searches and quality assessment and
possibly the first steps of statistical analysis, it sometimes happens that
the quality is too poor, or there is statistical heterogeneity that prevents
formal statistical combining. Then one must resort to a "best evidence"
type of analysis of the meaning, reliability and conclusions that can be
drawn from the best of the individual studies. This is then a systematic
approach that has avoided the cherry picking of the traditional narrative
review, but is similar to a review in the study-by-study presentation of
analysis without any statistical combining. Thus, the term "systematic
review" sets it apart both from a traditional selective review and a formal
statistical combining or meta-analysis. By default the term "meta-analysis"
remains to be used for formal statistical combining.
If this informal dichotomy of these terms serves the field, I don't see much
point in worrying about what Glass called things back in the 70s. This is
similar to the situation in which formal English makes no distinction
between pronouns for second person singular and plural (you and you).
Because of this inadequacy of the language, many different ethnic groups
that speak English have developed their own version of the second person
plural pronoun (you all, y'all, you'uns, you'se, you guys, you'se guys,
etc.).
Bye now y'all (you'uns, you'se, whatever...)
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|