Sample size and study interpretation

• Published in: Canadian Medical Association journal (CMAJ) Vol. 170; no. 8; pp. 1207 - 1208
• Main Author: Leung, Frank C
• Format: Journal Article
• Language: English
• Published: Canada CMA Impact, Inc 13.04.2004; Canadian Medical Association
• Subjects: Data Interpretation, Statistical; Letters; Randomized Controlled Trials as Topic - statistics & numerical data; Sample Size
• ISSN: 0820-3946; 1488-2329
• DOI: 10.1503/cmaj.1040023

We welcome the opportunity to restate important limitations of our study, which we discuss in our paper.1 [Frank C. Leung]'s main concern relates to sample size and its implications. The sample size of 16 for a power of 80% was calculated using an estimated standard deviation of 2.0 and an estimated mean difference of 1.5 cm on a 10-cm visual analogue scale. This mean difference was considered adequate in the expected and confirmed sample of moderately hungover individuals.2 Because of the small sample size and the measurement variation, which proved larger than expected, we discuss in our paper the degree of uncertainty relating to the data and state that this might have obscured a possible true effect. Acknowledging the study's limitations and in the absence of any trend in favour of artichoke extract, we stand by our conclusion that "our findings do not suggest that artichoke extract is effective in preventing alcohol-induced hangover."