Every profession has some set of norms to follow to achieve its objectives. These norms are called professional ethics which shows the essence of human behaviour. Same way, the field of medical research is expected to follow ethical norms, to obtain reliable and true results. Being a biostatistician, I would like to share my experiences to highlight that how the ethical norms are violated especially during data collection and data analysis, while conducting a trial in a health setup. In a follow-up study, correlated data recorded by the postgraduate students, health workers and field workers is highly biased in the desire of establishing the favourable effectiveness of their intervention. It is also seen that they have the tendency of manipulating the data at the time of analysis to have desirable outcomes. They fix in their mind to prove a drug effective, even when it is ineffective. Students are more prone to this practice. Sometimes, even the statistician apart from the students and health professionals, does not have adequate knowledge to apply the right test, which leads to wrong decisions. All these practices are un-ethical and all the members involved at any level in a trial must be made well aware, and must understand that these results are going to be applied on human beings, where one do not have a second chance to try. So, they should stick to the principles of honesty and truthfulness. In a clinical trial, reliability of data and results is very important. Biostatisticians must also understand that instead of applying wrong test, it is better to consult their seniors and should not indulge themselves in suppressing the true facts. It is their responsibility to translate the statistical interpretation of the results into medical interpretation.

]]>While many coherent fiducial distributions coincide with confidence distributions or Bayesian posterior distributions, there is a general class of coherent fiducial distributions that equates the two-sided p-value with the probability that the null hypothesis is true. The use of that class leads to point estimators and interval estimators that can be derived neither from the dominant frequentist theory nor from Bayesian theories that rule out data-dependent priors. These simple estimators shrink toward the parameter value of the null hypothesis without relying on asymptotics or on prior distributions.

]]>