Misrepresenting random sampling? A systematic review of research papers in the Journal of Advanced Nursing

摘要: Aim. This paper discusses the theoretical limitations of use random sampling and probability theory in production a significance level (or P-value) nursing research. Potential alternatives, form randomization tests, are proposed. Background. Research papers nursing, medicine psychology frequently misrepresent their statistical findings, as P-values reported assume sampling. In this systematic review studies published between January 1995 June 2002 Journal Advanced Nursing, 89 (68%) broke assumption because they used convenience samples or entire populations. As result, some findings may be questionable. Discussion. The key ideas for testing (for generating outlined. result research Nursing is then presented, showing how appears to have been misrepresented. Useful alternative techniques that might overcome these discussed. Review limitations. limited scope it applied one journal, so cannot generalized other journals general. However, possible also publishing articles based on misrepresentation several method was not completely clearly stated, circumstance judgment has made employed, indications given by author(s). Conclusion. Quantitative researchers should very careful appropriate design methods studies. If employ appropriate, run risk misinterpreting using inappropriate, unrepresentative biased samples.