CLINICAL
NEWS IN FOCUS
Editor's note: If you've ever wondered whether the p < 0.05 was a fair marker of statistical
significance, be sure to check out the May cover story in our sister publication, CardioSource
WorldNews, for an interesting look at p values.
An excerpt of the piece is printed below.
The Fading Bright Line of p < 0.05
By Debra L. Beck
T
he innocent p value has been abused to
within an inch of its life. It’s been misused, misinterpreted, miscommunicated,
and misunderstood. But missing it’s not, as researchers, journal editors, journalists, and readers
become ever more dependent on using a simple
p value to make snap judgments of good or bad.
Yet, using the p-value, as it has been in traditional
research, fundamentally miscasts the metric and
undermines the validity and reproducibility of the
science it’s trying to prove.
In an effort to shed light on the plight of the
poor p value and the damage its misuse appears to
be doing to the scientific endeavor, the American
Statistical Association (ASA), for the first time in its
177-year history, released a position statement on a
specific matter of statistical practice.1
“We are really looking to have a conversation
[…] with lots of people about how to take what we
know about what is good using p values and what
is not,” said ASA executive director Ronald L.
Wasserstein, PhD, in an interview with CardioSource WorldNews. What they want, he added, is
to “more effectively pass that along to scientists
everywhere so that we can do a better job of making inferences using statistics.”
After months of debate among a group of experts representing a wide variety of viewpoints, the
ASA released their statement in March 2016, conceding that the hard-fought ‘agreement’ among the
experts breaks no new ground but rather frames
some of the issues that have been debated for years.
The statement is comprised of six guiding principles with accompanying explanations that seek to
bring clarity to the issue.
Ironically, the p value
was never meant to
determine good from
bad or true from false.
12
CardioSource WorldNews: Interventions
1. P values can indicate how incompatible the
data are with a specified statistical model.
2. P values do not measure the probability that
the studied hypothesis is true, or the probability that the data were produced by random
chance alone.
3. Scientific conclusions and business or policy
decisions should not be based only on whether a p value passes a specific threshold.
4. Proper inference requires full reporting and
transparency.
5. A p value, or statistical significance, does not
measure the size of an effect or the importance of a result.
6. By itself, a p value does not provide a good
measure of evidence regarding a model or hypothesis.
According to the ASA, “The issues touched on here
affect not only research, but research funding, journal
practices, career advancement, scientific education,
public policy, journalism, and law.” The question (and
we don’t mean hypothesis-generating): how c