1. Anderson, Elizabeth H. Guest Editor
  2. Beck, Cheryl Tatano Guest Co-Editor

Article Content

Nursing has matured as a science. Nurse researchers, clinicians, and administrators are seeking objective synthesis of research findings to advance theory, determine effectiveness of interventions, guide patient care, and develop public policy. Quantitatively this can be accomplished through meta-analysis, a rigorous statistical procedure that synthesizes results from multiple primary research studies on a common clinical problem or issue. Meta-analysis provides the power to detect differences and effects across studies using similar variables in order to answer questions that a single study could not.


Critical to meta-analysis is the incorporation of all relevant primary studies (significant/not significant, published/not published). To enhance the quality and significance of their work, meta-analysts include unpublished studies (prevent the file drawer problem) and doctoral dissertations to reduce the bias toward published studies, as well as contact researchers for required statistics not found in their manuscripts. Researchers can advance nursing science by comprehensively reporting results and/or retaining and providing analyses required for meta-analysis database management systems (Beck, 1996).


Specific data needed are demographic variables relevant to the population and complete reporting of methodology and data analyses. Examples of pertinent population characteristics include: (a) sex, (b) age, (c) ethnicity/race, (d) education, (e) marital status, (f) socioeconomic status, (g) stage of disease, and (h) specific symptoms. Important methodological data are the research design, sample size and specific sample size for each statistical test, type of sampling, data collection techniques, past and current validity and reliability of all instruments, and outcome measures. The quality of each study should be evaluated and entered into the meta-analysis database system. Beck (1997) has developed a Meta-analysis Appraisal Checklist for this purpose.


In reporting data analyses, it is crucial not only to identify each statistical test but equally important is the exact value of the test with the exact p value indicating whether it is one-tailed or two-tailed as well as the degrees of freedom for each test. Likewise, it is valuable to report measures of central tendency (mean, median, and mode), measures of dispersion (standard deviation), and post hoc test values of degrees of freedom less than one for F tests. Incomplete reporting of research data and statistics may preclude the inclusion of a study in a meta-analysis. This could erroneously bias meta-analysis results and its generalizability as studies with missing data may have insignificant findings or be of poor quality (Beck, 1999).


To facilitate future meta-analyses, the Nursing Research Editor's Website at provides the opportunity for primary authors to post detailed supplemental analyses not included in their published article.




1. Beck, C. T. (1996). Use of a meta-analytic database management system. Nursing Research, 45, 181-184. [Context Link]


2. Beck, C. T. (1997). Use of meta-analysis as a teaching strategy in nursing research courses. Journal of Nursing Education, 36, 87-90. [Context Link]


3. Beck, C. T. (1999). Facilitating the work of a meta-analyst. Research in Nursing & Health, 22, 523-530. [Context Link]