P e egis a ion 1
The S a e o S udy P e egis a ion -
How Science Bene i s F om I and How o Apply I
Hilma B ohme
Depa men o Psychology, G az Open Science Ini ia i e, A qus Open Science Ambassado , Uni e si y o G az,
hilma .b ohme @uni-g az.a , h ps://o cid.o g/0000-0001-7763-4229
Gab iela Ho e
Depa men o Psychology, G az Open Science Ini ia i e, A qus Open Science Ambassado , Uni e si y o G az,
gab iela.ho e @uni-g az.a , h ps://o cid.o g/0000-0003-4407-1487
Sa ah on Gö z
Depa men o Psychology, G az Open Science Ini ia i e, Uni e si y o G az,
sa ah. on-goe z@uni-g az.a , h ps://o cid.o g/0009-0006-9430-8867
Abs ac
S udy p e egis a ion – a w i en plan on how o conduc a u u e s udy – is an impo an , ye
unde app ecia ed Open Science p ac ice. In his a icle, we desc ibe he eme gence o
p e egis a ion in medicine and psychology and how science as a whole may bene i om i . We also
show ha he up ake o p e egis a ion is e y slow hus a among esea che s in he social and
beha io al sciences. The e may be se e al ese a ions and misconcep ions o his, which we
add ess. We hope ha a sho and accessible u o ial may mo i a e esea che s also om mo e
ields o s a using p e egis a ion in hei wo k low.
Keywo ds: p e egis a ion, egis e ed epo , clinical ial, Open Science
This manusc ip is cu en ly unde pee e iew. Ci e a you own isk
P e egis a ion 2
In oduc ion
The concep o a p e egis a ion is simple and s aigh o wa d: I is a w i en and publicly
accessible plan desc ibing why and how a esea che aims a conduc ing and analyzing a scien i ic
s udy in he u u e hus p o iding a le el o anspa ency o one’s esea ch. Al hough his seems like
a i ial ask – and e en one ha laype sons may belie e o be common in science anyways –
p e egis a ions ha e no been a ound o long.
In his a icle, we wan o b ie ly e iew how and unde which ci cums ances p e egis a ion
was in oduced o esea ch on human subjec s, how se e al p e egis a ion o ms de eloped based
on he o iginal idea, and how e ec i e p e egis a ions a e o ackle widesp ead p oblems o
publica ion bias. Nex , we will p esen how p e egis a ion is adop ed by esea che s, pa icula ly
PhD s uden s, in psychology—a ield whe e p e egis a ion has ecen ly become mo e common—
o e ime. Then, we will p o ide a quick u o ial on how o implemen p e egis a ion in one’s esea ch
using he popula Open Science F amewo k eposi o y (OSF). Finally, we will add ess some
common c i iques o p e egis a ion. This a icle aims o gi e a quick o e iew on p e egis a ion and
emo e ba ie s esea che s migh ha e owa ds implemen ing hem by p o iding a quick and easy
u o ial on how o ge s a ed, hus mo i a ing hem o in eg a e p e egis a ion in hei own esea ch.
Eme gence o P e egis a ion
Al hough esea che s p esumably spend a lo o ime wo king ou he de ails on how o se up
hei scien i ic s udies, he idea o a w i en digi al plan ha lays ou he speci ics o a u u e s udy
and ha is publicly accessible is no olde han a qua e cen u y.
In medical esea ch, p e egis a ion was i s implemen ed a ound he yea 2000 in esponse
o an eme ging c edibili y c isis: many clinical s udies p obing new medica ions (usually conduc ed
in andomized con olled ials) seemed o show posi i e e ec s (i.e., pa ien s became heal hy
quickly ollowing a new medica ion compa ed o placebo o an olde medica ion). Howe e , a e
many ini ial s udies, ollow-up s udies o medical ea men s some imes did no show he same heal h
bene i s – despi e seemingly ca e ully conduc ed esea ch. As a esponse, he US Na ional Ins i u e
o Heal h launched he i s egis y o s udies – ClinicalT ials.go – whe e esea che s had o
indica e he aims and se -up o hei planned s udies. I was ecei ed posi i ely and jus i e yea s
la e he In e na ional Commi ee o Medical Jou nal Edi o s o majo scien i ic jou nals made
egis a ion o s udies a manda o y equi emen o la e publica ion (McC ay & Ide, 2000;
CinicalT ials.go , 2025). The e ec i eness o his ea ly o m o a p e egis a ion became e iden in
2010s: ini ial s udies in medical esea ch now o en did no epo exagge a ed heal h bene i s
anymo e. Ins ead, indings became mo e nuanced and o en imes e en showed null e ec s (Kaplan
& I in, 2015). The public egis a ion o he esea ch plan esul ed in ewe esea che s ying o
ci cum en hei ini ial idea o publish “posi i e” esul s ha u ned ou o be luke indings la e .
The social and beha io al sciences had hei own simme ing c edibili y and eplica ion c isis,
which la ed up e y suddenly in he 2010s. Fi s , conce ns a ose ha some indings in he social
P e egis a ion 3
and cogni i e-psychological domain seemed o be incompa ible wi h physical unde s andings o he
wo ld o jus absu d o be aken o g an ed (Lakens, 2025). Mos p ominen ly, one pape in a
p es igious psychology jou nal demons a ed ha some people could “ eel he u u e” by seemingly
co ec ly guessing he posi ion o e o ic pic u es on an emp y compu e sc een. No only did his
inding cause some media con o e sy (F ench, 2012), bu also p o ed o be un eliable in a ollow-
up s udy (Ri chie e al., 2012).
Mo e o en han no did hese indings show posi i e – s a is ically signi ican – e ec s ha
appea ed o con i m he esea che s’ hypo heses, implying ha hey a e almos always igh wi h
hei p edic ions. This bias owa ds posi i e esul s is common in many esea ch ields (Fanelli,
2010). Howe e , a g oup o me hodologis s con incingly a gued ha many indings could simply be
luke esul s s emming om undisclosed analyses s eps (e.g., esea che s lexibly un se e al
analyses, bu only epo he one’s ha show a posi i e inding). When ca e ully conduc ing
eplica ion s udies, hese esul s would likely anish (Simmons e al., 2011). This was con i med in
a la ge-scale eplica ion p ojec by he Open Science Collabo a ion (2015) o 100 published s udies
om majo jou nals could show indeed many o hese indings (mo e han 60%) did no eplica e –
a majo upse o esea che s in he ield.
Simmons and colleagues (2011) a gued ha he issues o eplicabili y and anspa ency need
o be add essed a beyond esea che s only sha ing hei da a and me hods, which cons i u e
necessa y, bu no su icien Open Science p ac ices o mo e anspa ency. Ins ead, esea che s
would need o hold commi o and uphold impo an pa s o hei esea ch idea be o ehand o
dec ease he chance ha hey would la e de ia e oo much om hei ini ial plan h ough hindsigh
and con i ma ion bias in hei analysis decisions. In o he wo ds, hey needed o plan ahead mo e
ca e ully, which pa o hei esea ch is con i ma o y (hypo heses es ing) and which pa is
explo a o y (hypo heses gene a ing; see also Nosek e al., 2018). Hence, he idea o s udy
p e egis a ion s a ed o become popula in psychology.
Key Poin s o a P e egis a ion
Simmons and colleagues (Da a Colada, 2015) as well as Nosek and his Cen e o Open
Science (n.d.) became a angua d o p e egis a ion by p o iding ea ly online egis ies o popula ize
he concep in psychology and o he social and beha io al sciences and p omo ing he concep
among esea che s. I one compa es popula p e egis a ion o ms on hese eposi o ies (e.g., he
popula “AsP edic ed” o m o he “OSF P e egis a ion” o m), p inciples o wha a p e egis a ion
should en ail o be in o ma i e become e iden (see also McPhe es, 2020; Simmons e al., 2021).
These p inciples can be oughly subsumed unde he ollowing key poin s o a p e egis a ion – pa s
ha should be ixed be o ehand o a oid lexibili y in da a analyses s eps:
a) Resea ch ques ion and hypo heses (i.e., wha ques ion do he esea che s wan o answe
and wha speci ic p edic ion, i any, is being made o he con i ma o y pa ?)
P e egis a ion 4
b) Resea ch design and a iables (i.e., how is he s udy concep ualized and se up and which
a iables a e measu ed o expe imen ally manipula ed?)
c) Sample size jus i ica ion (i.e., wha is he popula ion and he sample size d awn om i ? Is
his sample size su icien o ind a hypo hesized e ec gi en ha i exis s?)
d) Analysis plan including p epa a ion s eps (i.e., how is he da a being p epa ed o become
eady o he analysis? Wha a e he speci ic analysis s a egies?)
De elopmen o P e egis a ion and Regis e ed Repo s
While he ini ial p e egis a ion o ms ocused on expe imen al s udies in psychology, se e al
g oups o Open Science p oponen s om di e en disciplines ha e since upda ed and e ined he
idea o s udy p e egis a ion, making i sui able o all kinds o esea ch, including quali a i e s udies
(Ha en & an G oo el, 2019), eplica ion s udies (B and e al., 2014), sys ema ic e iews ( an den
Akke e al., 2023), o seconda y analyses o exis ing da a ( an den Akke e al., 2021). These e o s
also e lec ha esea che s do no only ecognize he non- eplicabili y as an issue o social and
cogni i e psychology (in which he eplica ion c isis s a ed), bu as a undamen al issue o all
scien i ic disciplines ha apply some kind o hypo hesis es ing. A e all, excessi e posi i e esul s
a e p esen in mos esea ch disciplines (see Fanelli, 2010), and many o hese deal wi h own
eplica ion issues (e.g., expe imen al economics, Came e e al., 2016, social sciences, Came e e
al., 2018, and p e-clinical cance esea ch, E ing on e al., 2021).
Se e al jou nals e en b ough he p e egis a ion concep one s ep u he : While
p e egis e ing a s udy be o e da a collec ing p o ides a le el o anspa ency, he e is no o mal
e iew p ocess be o ehand. The jou nal o ma o Regis e ed Repo s (Chambe s, 2013, Chambe s
& Tza ella, 2022) add esses his p oblem: his o ma includes a o mal e iew p ocess be o e da a
is collec ed. This way, esea che s w i e a jou nal-s yle p e egis a ion and submi i o a jou nal
be o e he begin o he s udy. The pee - e iew hen c i iques he heo y and me hods ( a he han
he esul s) and can gi e cons uc i e eedback be o e any da a is collec ed. This e ec i ely allows
he esea che s o s ill implemen ex e nal ideas in hei s udy design. Regis e ed Repo s come wi h
he p emise ha he esea ch is e alua ed based on i s ac ual quali y and no whe he i s esul s u n
ou o be posi i e, hus also p o iding a me hod o p e en ing a publica ion bias ha is skewed
owa ds posi i e esul s (Chambe s, 2017).
Simila o medical esea ch, p e egis a ion in he social and beha io al sciences, p o ed o be
e ec i e: e ec sizes in published s udies ha u ilized p e egis a ion a e o en much smalle han in
egula s udies (Schä e & Schwa z, 2019), implying ha ha he li e a u e is less swamped wi h
luke indings. This is pa icula ly ue o Regis e ed Repo s, which also seem o be o highe quali y
in e ms o hei heo izing, me hods, and implica ions (Scheel e al., 2020; Sode be g e al., 2021).
Al hough egis ies like he Open Science F amewo k (h ps://os .io/) p o ide esea che s wi h
a highly s eamlined and low e o way o p e egis e ing hei s udies, he ques ion emains, how
many esea che s ac ually chose o apply i in hei esea ch.
P e egis a ion 5
Up ake o P e egis a ion
Con a y o andomized con olled esea ch in medicine, s udies in he social and beha io al
sciences do no include manda o y p e egis a ion be o e da a collec ion. Howe e , Open Science
communi ies and ini ia i es ha e eme ged ac oss uni e si ies o p omo e he up ake OS p ac ices –
including p e egis a ion. These ini ia i es o en employ a ious me hods such as implemen ing,
jou nal clubs, o ganizing e en s, eaching cou ses e c. o con ince esea che s o he impo ance o
OS p ac ices o imp o e esea ch as a whole as well as p o iding hem wi h he skills hey need o
success ully implemen hem (e.g., O ben, 2019, A meni e al., 2021, B ohme e al. 2025 in his
issue).
None heless, in he end, he success o such endea o s may be measu ed in how popula OS
p ac ices ha e become in esea ch ou pu , ha is, in scien i ic publica ions. Fo p e egis a ions, an
inc ease h oughou he las decade may indica e ha esea che s s a ed ecognizing hese
p ac ices being impo an o imp o e anspa ency in gene al. Some ecen da a om Ha dwicke and
colleagues (2024) sugges s ha he e is indeed an up ake in applying p e egis a ion: Whe eas
be ween 2016 o 2018, abou 7% o esea ch a icles in psychology epo ed a unc ional link o a
p e egis a ion o m, his numbe had inc eased o 20% in 2022.
B ohme and Ho mann (2025) did some hing simila , bu ocused on PhD s uden s, because
hey a e assumed o jus ha ing acqui ed hei p ima y esea ch skills, which hey will likely use un il
he end o hei ca ee . In a sample o s udies (N = 379) om 91 disse a ions om psychology
depa men s a wo uni e si ies, hey ound ha 20% (n = 80) con ained a p e egis a ion, con i ming
he esul s by Ha dwicke (see Figu e 1).
Figu e 1. Use o p e egis a ion in s udies o psychology disse a ions 2018-22.

P e egis a ion 6
F om all s udies ha used p e egis a ion, mos we e based on he beginne - iendly
“AsP edic ed” o m (16%, n = 60), ew used he mo e comp ehensi e “OSF P e egis a ion” o m
(3.7%, n = 14), and n = 3 (0.8%) conduc ed a Regis e ed Repo (see also Figu e 1). Hence, ECRs
s a ed using p e egis a ion, bu he e is ce ainly oom o imp o emen .
I one looks a he p e alence o Regis e ed Repo s in jou nals, hey do no appea o be e y
popula ye : we compa ed he p opo ion o Regis e ed Repo s in wo jou nals om psychology:
Collab a: Psychology eme ged as a small jou nal dedica ed o anspa ency and ep oducibili y
h oughou he las couple o yea s, whe eas Na u e Human Beha io quickly became a high-impac
jou nal, ocusing on mo e b oadly in e es ing and high-impac esul s. Bo h jou nals in oduced
Regis e ed Repo s in 2017. Collab a’s a e o Regis e ed Repo s wen up om ze o o 16% in 2024
(14 ou o 88 published esea ch a icles). Fo Na u e Human Beha io only abou 1% o published
s udies in 2024 we e in a Regis e ed Repo o ma ( wo ou o 153 published esea ch a icles).
This may indica e wo hings: I) The ew esea che s, who aim a conduc ing a Regis e ed
Repo may a he “play i sa e” by submi ing i o a jou nal ha alues anspa ency and
ep oducibili y explici ly. A po en ial eason migh be ha hey s uggle wi h e alua ing he impo ance
and impac o hei esea ch be o e knowing hei esul s.
1
II) Fo a simila eason, he edi o s o high-
impac jou nals may ejec Regis e ed Repo s mo e o en as hey ea po en ial null indings (which
a e mo e common in Regis e ed Repo s), which hey hink a e less ci able.
Un o una ely, he e is li le da a on he adop ion o p e egis a ion in esea ch ields beyond
psychology. Bu as he awa eness o he impo ance o OS p ac ices only ecen ly gained ac ion in
ields like sociology, consume esea ch, o educa ion esea ch, a slow and delayed up ake o
p e egis a ion is a he plausible. Wi h ou ollowing u o ial, we hope o s imula e a u he up ake
o p e egis a ion ac oss ields.
Fou Easy S eps o P e egis a ion – A B ie Tu o ial
Ce ainly, science would bene i om a quicke up ake o OS p ac ices and pa icula ly om
he up ake o s udy p e egis a ion. Howe e , i can be di icul o know whe e o s a as his
me hodological skill is o en no augh in uni e si y cou ses. Mo eo e , esea che s a la e ca ee
s ages migh ha e ecei ed hei aining be o e he eme gence o p e egis a ion. In his sec ion, we
will p o ide a ou -s ep u o ial o p e egis a ion in he hopes ha his will make i easy o in e es ed
esea che s o ge s a ed.
We ha e chosen he Open Science F amewo k (OSF), as his is in ou opinion he mos
accessible egis y and esea che s migh al eady be amilia wi h i o o he OS- ela ed pu poses
(e.g., as a eposi o y o da a sha ing). The e o e, a p e equisi e is ha in e es ed eade s c ea e an
accoun on he OSF (h ps://os .io/). I is also impo an ha you al eady ha e a ela i ely conc e e
1
Fo esea che s, who alue OS a lo , ano he eason migh be ha he smalle , anspa ency-o ien ed jou nal is mo e in
line wi h hei alues and ha he open-access publica ion cha ges in ha jou nal a e no as high.
P e egis a ion 7
s udy idea in hei mind, including he esea ch ques ions, hypo heses and me hodology. The OSF
p o ides many a ailable p e egis a ion o ma s, which di e in le el o de ail and ype o esea ch
hey can be applied o (see h ps://help.os .io/a icle/229-selec -a- egis a ion- empla e). As his is a
beginne ’s u o ial, we will ocus on he “AsP edic ed” o ma which is ela i ely simple and can be
applied o mos expe imen al s udies. The ou s eps a e illus a ed in Figu e 2.
2
S ep 1: C ea e a New P ojec
“P ojec s” on he OSF a e supe o dina e olde s, which can be c ea ed o single s udies o
whole esea ch p ojec s (e.g., a disse a ion o a pos -doc g an p ojec ). Upon clicking on he bu on
“C ea e New P ojec ” (Figu e 2a), one should selec a sho , bu clea ly desc ip i e i le (e.g., “S udy
1: Coope a i e Beha io ”). When selec ing a se e , i is ad isable o keep local da a p o ec ion
egula ions in mind (e.g., Eu opean esea che s may selec he F ank u se e ).
Figu e 2. Fou S eps o a P e egis a ion ia he Open Science F amewo k
No e: ed ellipses and a ows wi h anno a ion a e added o cla i ica ion; o m is a ailable on
h ps://os .io/; sc eensho s we e aken in July 2025
2
A ideo u o ial in Ge man language is p o ided on he Uni ube po al o he Uni e si y o G az: h ps://uni ube.uni-
g az.a /po al/au zeichnungen.h ml?id=e1a d6a9-c7c -4 aa-a77d-4953c501c67
P e egis a ion 8
S ep 2: Na iga e o he P e egis a ion Fo m
The landing p ojec page is now c ea ed and should look like as in Figu e 2b. No e ha his
p ojec is “p i a e”, which means i is only isible o he au ho hemsel es and will emain p i a e
un il he au ho s chose o sha e he p ojec (e.g., du ing publica ion o he esea ch pape ). Following
he abs on he op, one could la e add in o ma ion on he p ojec by clicking on he Wiki ab, add
supplemen al documen s, me hods, and da a iles ia he Files ab, o add co-au ho s and
collabo a o s o he p ojec ia he Con ibu o s ab. Fo adding a p e egis a ion, one should simply
click on he Regis a ions ab and hen on “c ea e new egis a ion”.
S ep 3: Choose a P e egis a ion Fo m
As we men ioned abo e, he e a e many o ms a ailable, bu we selec he “AsP edic ed” o m
om he lis (Figu e 2c).
3
Resea che s wi h mo e p e egis a ion expe ience may choose he “OSF
P e egis a ion” o m, which is mo e in o ma i e in ega d o some de ails (i.e., op ion 1, see Bakke
e al., 2020) o ano he o m ha closely aligns wi h hei s udy design (e.g., Replica ion Recipe).
A e selec ion, click on “c ea e d a ”.
S ep 4: Fill Ou he Fo m and Submi I
This inal s ep equi es some ime, depending on he s udy’s complexi y and how well hough -
h ough all s udy aspec s al eady a e. Fo a simple s udy, whe e all me hodological and analy ical
decisions ha e al eady been made, one may plan o in es abou 20 o 30 minu es. Bu he mo e
de ails one can add, he be e o he pu pose o he anspa ency. Following a sho s udy
desc ip ion (see Figu e 2d), one has o ill ou eigh mo e ques ions (o headings), which con ain
in o ma ion on he abo e-men ioned ou key poin s A o D. He e we desc ibe hese poin s o he
“AsP edic ed” o m o p o ide esea che s wi h some guidance o anspa ency:
1. Ti le and Desc ip ion. He e, he pu pose o he s udy should be con ex ualized and he
esea ch should be men ioned (e.g., online expe imen ia a su ey pla o m). (Co esponds
wi h key poin b)
2. Has he Da a Collec ion S a ed? Choose he app op ia e op ion – usually i is “no”. I
“yes” is chosen, elabo a e in poin 9 why you s ill belie e ha his s udy can be
p e egis e ed.
3. Hypo heses. Fo con i ma o y s udies, one should lis all hypo heses, which will be es ed
in he s udy. Al hough he heading asks o hypo heses, one could also lis u he
explo a o y esea ch ques ions. (Co esponds wi h key poin a)
3
E en i he planned s udy a hand is no an expe imen , bu a he an obse a ional / co ela ional s udy, one can use
cus omize he con en o i he pu pose. E.g., ins ead o “condi ions”, one may simply desc ibe all p edic o a iables.
P e egis a ion 9
4. Dependen Va iable. This is he a iable (o a iables) o in e es ha is measu ed o d aw
a conclusion om (i.e., he ou come a iable on which wo g oups a e compa ed).
(Co esponds wi h key poin b)
5. Condi ions. In a simple expe imen , his poin cons i u es he expe imen al and con ol
g oup. I is bene icial o include some in o ma ion wha he ea men o he expe imen al
g oup en ails (e.g., a new medica ion). Also, one could add ele an con ol a iables (e.g.,
pa icipan gende ). Fo co ela ional s udies, one should add he e any independen ,
p edic o o mode a ion a iable which is ele an o he s udy. (Co esponds wi h poin b)
6. Analyses. How is he da a analyzed o d aw conclusions om i and wha is he in e en ial
c i e ium (e.g., he p- alue o a es s a is ic)? He e, one should also add impo an
p ep ocessing in o ma ion, such as, how he dependen a iable is ans o med o how
composi e sco es a e calcula ed. (Co esponds wi h key poin d)
7. Ou lie s and Exclusion. Some imes, cases ha e o be excluded be o e da a analysis (e.g.,
because o ex eme alues on he dependen a iable, unme inclusion c i e ia (e.g., ela ed
o a minimum o maximum age, language equi emen s, e c.), o doub s abou da a quali y).
These exclusion ules should be de ined he e. (Co esponds wi h key poin d)
8. Sample Size. One should indica e an an icipa ed sample size ha one aims o collec . I is
ad isable o also p o ide a sample size jus i ica ion (e.g., ia a s a is ical powe analysis)
and dis inguish be ween collec ed ( aw) sample size and analyzed sample size (e.g., a e
exclusion, see poin 7). (Co esponds wi h key poin c)
9. O he . I could be ha da a on addi ional a iables is collec ed o he s udy is pa o a la ge
esea ch p ojec . Such speci ics, including u he explo a o y analysis ideas, can be
op ionally indica ed he e.
Up un il his poin , he p e egis a ion is s ill in “p i a e” mode. Upon p essing he “submi ”
bu on, which o icially egis e s he s udy, esea che s can se an emba go o when he
p e egis a ion becomes isible and indable This can be up o ou yea s in he u u e and gi es
esea che s ime o conduc hei s udy and publish hei indings. The link o he p e egis a ion
should hen be made a ailable as pa o he inal publica ion. This will make hei s udy plan
anspa en in he con ex o he published s udy.
4
Rese a ions Agains P e egis a ions
Finally, we wan o add ess some equen ly aised c i icisms o p e egis a ion ha a e
commonly oiced and ha we ha e encoun e ed du ing ou wo k a he Open Science ini ia i es (see
4
No e ha o he pee - e iew p ocess, esea che s can also p o ide an anonymized iew-only link o hei
p e egis a ion, as a pee e iew is usually double-blind. To achie e his, one can ollow he in o ma ion on his link:
h ps://help.os .io/a icle/155-c ea e-a- iew-only-link- o -a- egis a ion