"To consult the statistician after an experiment is finished is often merely to ask him to conduct a post mortem examination. He can perhaps say what the experiment died of"

Ronald Fisher’s metaphor about consulting the statistician after an experiment is finished evokes the irreversible consequences of approaching data analysis as an afterthought. He likens the post-experimental involvement of a statistician to a medical examiner who can only uncover the cause of death rather than save the patient. The experiment, in this analogy, is already deceased, its fate sealed by early irreversible decisions that could render any subsequent analysis futile or misleading.

This perspective highlights the critical importance of statistical expertise in the initial planning stages of research. Experimental design is foundational, not merely a technicality to be patched up later. Elements such as the method of randomization, control group selection, sample size, and measurement techniques are all deeply consequential. Statistical methods are not mere arithmetic performed after the fact; they fundamentally shape the validity, reliability, and interpretability of the study’s findings. Engaging with a statistician only after the experiment has run its course often uncovers basic design flaws, confounding factors, or insufficient power to detect an effect, none of which can be remedied once data collection is complete.

Fisher’s observation serves as a warning: retrospective statistical consultation may illuminate where and how an experiment went wrong, but it cannot rescue an investigation from its foundational shortcomings. The implication is clear, statistical literacy and planning cannot be a postscript but must inform the earliest moments of the research process. For empirical research to yield useful, trustworthy knowledge, methodology and analysis must be seamlessly interwoven from inception. Ultimately, Fisher appeals for a collaborative approach, where statisticians and researchers work side by side, designing experiments that can withstand rigorous scrutiny so that the answers they seek are both discoverable and defensible.