OBJECTIVE: To evaluate the quality of research reports published in the chiropractic literature.
DATA SOURCES: Original research articles, defined as those reporting studies that included primary data collection, published in the 13 Chiropractic Research Journal Editors Council member journals from January 1999 through February 2000.
METHODS: A 2-page checklist developed from review criteria used in the biomedical literature. Three doctoral-trained biostatisticians blinded to author identification and affiliation evaluated the articles.
RESULTS: Of the 73 eligible articles, 21% lacked description of the study design, 77% did not provide a sample size justification, 26% reported inappropriate descriptive statistics, 26% reported conclusions not supported by the results, and 21% did not avoid redundancy of information presented in tables, figures, and text. Of the 55 articles with a primary objective of testing, 10 did not report any inferential statistics, 19 reported inappropriate inferential statistics, and 22 did not make the primary comparisons of interest clear.
CONCLUSIONS: In this study, we evaluated the quality of the research report rather than the research itself. However, the value of the research articles is diminished by the poor quality of reporting, as well as the actual errors in data analysis. Yet, this problem is not unique to the chiropractic literature. It is recommended that all investigators submitting manuscripts to chiropractic journals use available guidelines in preparing their research reports and that reviewers use those same guidelines in critiquing the manuscripts.
Click on the above link for the PubMed record for this article; full text by subscription. This abstract is reproduced here with the permission of the publisher.
|
