L. Held (University of Zurich)

**‘A new standard for the ana
lysis and design of replication studies’**

A new standard is p
roposed for the evidential assessment of replication studies. The approach
combines a specific reverse Bayes technique with prior-predictive tail prob
abilities to define replication success. The method gives rise to a quantit
ative measure for replication success, called the sceptical p-value. The sc
eptical p-value integrates traditional significance of both the original an
d the replication study with a comparison of the respective effect sizes. I
t incorporates the uncertainty of both the original and the replication eff
ect estimates and reduces to the ordinary *p*-value of the replicati
on study if the uncertainty of the original effect estimate is ignored. The
framework proposed can also be used to determine the power or the required
replication sample size to achieve replication success. Numerical calculat
ions highlight the difficulty of achieving replication success if the evide
nce from the original study is only suggestive. An application to data from
the Open Science Collaboration project on the replicability of psychologic
al science illustrates the methodology proposed.

To be published in&n bsp;Series A; for more information go to the Wil ey Online Library.

K. Rice (University of Washington, Seattle), T. Bonnett (National Laboratory for Cancer Research, Frederick) and C. Kra kauer (University of Washington, Seattle)

**‘Knowing the signs:
a direct and generalizable motivation of two-sided tests’**

M
any well-known problems with two-sided *p*-values are due to their u
se in hypothesis tests, with ‘reject–accept’ conclusions about point null h
ypotheses. We present an alternative motivation for *p*-value-based
tests, viewing them as assessments of only the *sign* of an underlyi
ng parameter, where we can conclude that the parameter is positive, negativ
e or simply say nothing either way. Our approach is decision theoretic, but
—unusually—we consider the whole set of possible utility functions availabl
e. Doing this we show how, in a specific sense, close analogues of familiar
one- and two-sided tests are *always* the optimal decision. We argu
e that this simplicity could aid non-experts’ understanding and use of test
s—and help them to think critically about whether or not tests are appropri
ate tools for answering their questions of interest. Several extensions are
also considered, showing that the simple idea of determining the signs of
parameters yields a rich framework for inference.

To be published in& nbsp;Series A; for more information go to the Wi ley Online Library

CONTACT:https://events.rss.org.uk/rss/270/home X-EXTRAINFO:This meeting takes place as part of the RSS 2019 Conference.\n\nIf you are attending the conference please register as normal via the link above.\n\nI f you just wish to attend the Discussion Meeting please email This email ad dress is being protected from spambots. You need JavaScript enabled to view it. document.getElementById('cloaka3919803446d6498d868b162dfc6b71c').i nnerHTML = ''; var prefix = 'ma' + 'il' + 'to'; var path = 'hr' + 'ef' + '='; var addya3919803446d6498d868b162dfc6b71c = 'c onference' + '@'; addya3919803446d6498d868b162d fc6b71c = addya3919803446d6498d868b162dfc6b71c + 'rss' + '.' + 'or g' + '.' + 'uk'; var addy_texta3919803446d6498d868b162dfc6b71c = 'conference' + '@' + 'rss' + '.' + 'org ' + '.' + 'uk';document.getElementById('cloaka3919803446d6498d868b 162dfc6b71c').innerHTML += ''+addy_texta3919803446d6498d868b162dfc6b71c+''; DTSTAMP:20190919T060749Z DTSTART;TZID=Europe/London:20190904T172000 DTEND;TZID=Europe/London:20190904T192000 SEQUENCE:0 TRANSP:OPAQUE END:VEVENT END:VCALENDAR