The Three Rs of Good Science
Posted: January 1, 2014
Tags: Research
Reading, writing, and arithmetic — the three "R"s — are the bedrock of a good education. Scientists use these skills every day to learn, communicate, and analyze. But researchers have another set of three Rs: rigor, reproducibility, and relevance. Taxpayers who fund the basic and clinical studies at the National Institutes of Health (NIH) expect nothing less than a firm commitment to these foundational principles. Public trust in science hinges directly upon it.
I am involved with efforts at NIH to shine a brighter spotlight on the importance of the three Rs of good science. The reason for this current NIH emphasis is that recent reports have demonstrated that a significant number of basic, preclinical published studies cannot be reproduced by other scientists. This prompted an urgent call to action among NIH top leadership, as described in a January 2014 Nature commentary from NIH Director Dr. Francis Collins and Deputy Director Dr. Larry Tabak. The article's main thesis states:
"Crucial experimental design elements are all too frequently ignored, including blinding, randomization, replication, sample size calculation and the effect of sex differences."
The phrase "and the effect of sex differences" is where I am ensuring that sex- and gender-based factors inform the biomedical evidence base at many levels. Although even little children recognize the basic differences between boys and girls, this fundamental distinction — which plays out at the level of cells, selves, and societies – remains under-addressed in biomedical research. We need to change that.
Getting on top of this problem need not be complicated or expensive. But we must be proactive. Our neighbor to the north, the Canadian Institutes of Health Research (CIHR), provides inspiration for possible approaches. For the past several years, CIHR research grant applicants must include specific answers to yes-or-no questions about inclusion of sex and gender in their research design. "No" answers require an explanation of why sex and gender has not been included.
The vast majority of irreproducibility in biomedical research is likely unintentional. Rather, it likely stems from a variety of factors such as inadequate emphasis on experimental design and statistical analyses, grant and publication review efforts that fail to catch errors and omissions, or overinterpretation of the results of exploratory experiments.
What part can we play? What can we do right now? Let's build broad coalitions to more fully explore and understand sex and gender influences in health and disease through rigorous NIH funded research. Let's communicate this message widely to grant applicants, reviewers, and other NIH stakeholders. Sex/gender is a fundamental research variable, and thus sex/gender inclusion — from cells to selves — is a necessary component of rigorous science.