Our facial expressions can trigger
emotional reactions - the so-called "facial feedback hypothesis" - even
when people are unaware that they are making that expression, suggested a 1988 paper that reported two studies
in which the researchers surreptitiously changed participants' facial expressions.
Now, a coordinated replication effort conducted across 17 labs found no
evidence that surreptitiously inducing people to smile or frown affects
their emotional state. The findings of the replication project are
published as part of a Registered Replication Report (RRR) in Perspectives on Psychological Science
, a journal of the Association for Psychological Science.
‘There is no evidence that surreptitiously inducing people to smile or frown affects their emotional state.’
The RRR project, proposed by University of Amsterdam psychology
researchers Eric-Jan Wagenmakers, Titia Beek, Laura Dijkhoff, and
Quentin Gronau, aimed to replicate a 1988 study conducted by
psychological scientists Fritz Strack, Leonard Martin, and Sabine
The 1988 study is cited frequently in the scientific literature and in
introductory courses and textbooks. Although other studies
have tested the facial feedback hypothesis using different methods, this
influential study had not been directly replicated with the same design
and outcome measure. This RRR paper describes a rigorous, multilab
replication of that study, with each lab following a vetted protocol
that was registered online prior to data collection.
The aim was to replicate the original study as closely as possible,
but the RRR differed in several ways from the original. Fritz Strack
provided the materials from the original study, including the original
Far Side cartoons. The RRR study also used a set of Far Side cartoons
after first conducting a study to ensure that they were moderately funny
by today's standards.
The RRR protocol also standardized the
instructions to participants and stipulated that they be delivered via
computer in order to minimize interactions with the experimenter. Based
on guidance from an expert reviewer during the protocol vetting process,
participants were recorded on video during the experiment to ensure
that they were holding the pen correctly on each trial.
All of the materials, the protocol, the data, and the analysis scripts are publicly available on the Open Science Framework.
As in the original study, participants were told they would be
completing different tasks with parts of the body not normally used for
those tasks. Per the instructions provided, they held the pen in their
mouth (between their teeth or between their lips) and completed the
tasks presented in a booklet, which included drawing lines between
various numbers, underlining vowels, and indicating how amused they were
The combined results from 1,894 participants were inconsistent with
the findings reported in the original study. The data provided no
evidence that inducing participants to have particular facial
expressions led them to rate the cartoons differently.
"This RRR did not replicate the [Strack, Martin, Stepper] results
and failed to do so in a statistically compelling fashion," the
contributing researchers write in their report.
"Nevertheless, it should be stressed that the RRR results do not
invalidate the more general facial feedback hypothesis," they conclude.
In a commentary accompanying the RRR report, Strack commends the
efforts of those involved in the RRR. He notes his surprise that the
original finding was not replicated, especially given that his and
colleagues' labs have confirmed the results in "numerous operational and
conceptual replications." Strack speculates about some possible reasons
for the different outcomes, including that the presence of a camera in
the RRR experiments might have affected how participants reacted to the
Daniel Simons, the acting editor for this RRR project, commended the
care taken by the proposing authors, "This team's exceptional rigor and
care in developing the study protocol, teaching other researchers how
to follow it, and fully documenting every step of the process set a
standard that I hope future large-scale studies like this one will