Uncertainty beyond sampling errorBMJ 2014; 349 doi: https://doi.org/10.1136/bmj.g7065 (Published 25 November 2014) Cite this as: BMJ 2014;349:g7065
All rapid responses
Professors Altman and Bland touched upon yet another time the heart of statistics - the uncertainty and sampling error and beyond. In the twin (1,2 ) communications as statistical notes - former again defines sampling error and compares confidence interval and P-value in decision making for the context of randomized trials while indicating other sources of bias. The later emphasizes the importance of non-sampling errors that attribute to uncertainty beyond sampling error.
In the context of uncertainty beyond sampling error (2) authors beautifully explain how the confidence interval is implicated to sampling variation in terms of precision but not accuracy there by emphasizing the measurement errors or non-sampling errors. Of course random sampling of measurements or randomization is an essential pre-requisite in the measurement of uncertainty through sampling variation. In the light and knowledge of un-quantified uncertainty from other causes that is major fact that influence estimates, the clinical study results have to be carefully consumed by bio-medical researchers. No study is as such 100% full proof of truth and all are clue providers on the truth being investigated?
Under these circumstances what percent of weight-age to be assigned to each assessments of statistical uncertainty and uncertainty due to other causes. First of all it naturally requires to be ensured on measurement or methodological reliability to be maximum extent possible then statistical uncertainty to be used for decision making. The reliability of research findings weighed more in the usefulness of an estimate that depends on the confidence on that estimate as stated by GRADE (3) - the frame work on structured assessment of quality. So, along with the confidence interval estimates obtained from statistics, the confidence in the estimate ensured with quality is essential.
1. Altman DG, Bland JM. Uncertainty and sampling error. BMJ2014;349:g7064.
FREE Full Text
2. Altman DG, Bland JM. Uncertainty and sampling error. BMJ 2014;349:g7065.
FREE Full Text
3. Guyatt GH, Oxman AD, Vist GE, Kunz R, Falck-Ytter Y, Alonso-Coello P, et al. GRADE: an emerging
consensus on rating quality of evidence and strength of recommendations. BMJ2008;336:924-6.
Competing interests: No competing interests