Triplication: an important component of the modern scientific method

TL;DR

This paper examines the historical and current use of triplication in plant science, demonstrating its prevalence and methodological advantages for reducing errors and improving scientific reliability.

Contribution

It classifies replication protocols, assesses their prevalence in 2017 plant studies, and explores the theoretical basis for triplication's widespread use.

Findings

Triplication found in ~70% of analyzed studies

Triple-result protocols observed in ~15% of studies

Theoretical analysis supports triplication as optimal for error reduction

Abstract

A scientific-study protocol, as defined here, is designed to deliver results from which inductive inference is allowed. In the nineteenth century, triplication was introduced into the plant sciences and Fisher's p<0.05 rule (1925) was incorporated into a triple-result protocol designed to counter random/systematic errors. Aims here were to: (1) classify replication (including non-replicated) protocols; (2) assess their prevalence in plant-science studies published in 2017 for a defined variable construct; and (3) explore the theoretical rationale for the use of triplication. Methods: the plant sciences were surveyed and a protocol-prevalence report produced; association versus experimental proportions analyzed; and real-world-data proxies were used to show confidence-interval-width patterns with increasing replicate number. Results: triplication was found in ~70% of plant-science studies analyzed, with triple-result protocols observed in ~15%. Theoretical considerations showed that, even if systematic errors predominate, "square-root rules" sometimes apply, contributing to the predominance of triplication (real-world-data proxy examples given). Conclusions: Triplication was extensively applied in the studies analyzed and there are strong methodological reasons why (1) triplication, rather than duplication/quadruplication, is the most appropriate standard; (2) triple-result protocols are preferable to global averaging approaches; and (3) plant science methodological standards remain high, despite immense publication pressures.