Adequate data regarding disease burden is critical when justifying the need for prevention or intervention programs and health policies. Indigenous peoples may make up a small proportion of the overall participants in health studies, but altogether removing Indigenous peoples from results or collapsing Indigenous populations into an “other” category obscures important and needed health data.
• Studies should report and publish results for numerically small groups, regardless of statistical significance of estimates.17 For a discussion of similar dynamics in higher education research, see Shotton et al.23 • When sample sizes do not allow for statistical significance to be achieved, studies should avoid the urge to group numerically small groups, including Indigenous communities, into an “other” category. Instead, report results by group, regardless of sample size.17 Indeed, when practitioners attempt to use existing data to justify the need for programs or to develop prevention or intervention projects, estimates that categorize Indigenous communities as part of an “other” group are of little utility.17 • Limited data, when limitations are clearly specified, are better than no data. Withholding Indigenous-specific estimates makes it impossible for future meta-analyses or pooling projects to make use of existing study results. Further, Indigenous communities have noted that although they participate in studies, they often do not see results that are relevant to them. This weakens relationships necessary for the research enterprise to be sustained. Additional useful information includes sample size (raw numbers), percentage of total population, and confidence intervals (with explanations of their meanings). Include issues that stem from sample size in the limitations section. • Small sample size should not be considered an obstacle to high-quality research, particularly since findings can yield useful information for local practitioners. Include the results of qualitative techniques, such as focus groups or talking circles, to provide important information otherwise obscured by small sample sizes.18
|
