RESEARCH ARTICLE
Fac o s uc u e o mood o e ime ames
and ci cums ances o measu emen : Two
s udies on he P o ile o Mood S a es
ques ionnai e
Elena And adeID
1
*, Dolo es Rod ı
´guez
2
1Depa men o Social Psychology and Me hodology, Facul y o Psychology, Uni e sidade de San iago de
Compos ela, San iago de Compos ela, A Co uña, Spain, 2Depa men o Clinical Psychology and
Psychobiology, Facul y o Psychology, Uni e sidade de San iago de Compos ela, San iago de Compos ela,
A Co uña, Spain
*elena.and [email protected]
Abs ac
This pape p esen s he esul s o wo s udies on he in a iance o he P o ile o Mood S a es
ques ionnai e ac oss esponse ime ames and ci cums ances o adminis a ion. We
applied Spanish e sions o he ins umen o ga he da a om 1146 a hle es. In he i s
s udy (N = 700), we es ed he ac o s uc u e o he ques ionnai e in aining sessions by
using wo di e en ime ames: ‘ igh now’ (n = 350) and ‘pas week’ (n = 350). In he second
s udy (N = 446), we compa ed he ac o s uc u e o he ques ionnai e wi h da a collec ed
using he ins uc ion ‘ igh now’ a wo di e en ci cums ances: ‘ aining’ (n = 223) and ‘com-
pe i ion’ (n = 223). Da a analysis was simila in bo h s udies. We conduc ed mul i-g oup con-
i ma o y ac o analyses and applied he scaled di e ence chi-squa e s a is ic o examine
whe he disc epancies in successi e cons ained models we e signi ican . We obse ed
con igu al equi alence be ween he wo ime ames. Fu he mo e, we obse ed me ic
equi alence bu no scala in a iance be ween he di e en ci cums ances o measu emen .
The indings highligh he need o s udies o equi alence be o e using a single sel - epo
wi h mo e han one se o ins uc ions, o unde di e se ci cums ances. In a iance o mood
sco es should be examined and aken in o accoun when in e p e ing indi idual and g oup
mood s a e assessmen s.
In oduc ion
Sel - epo measu es o mood a e p edominan in he psychology domain, and ha e been
ex ensi ely used o explo ing subjec i e mood s a es in spo s people. Compa ed o o he sub-
jec i e expe iences like emo ions, moods a e desc ibed as less in ense bu mo e endu ing, di -
use, and o en no a ibu able o a clea cause [1]. Exis ing mood measu es ha e a
mul idimensional cha ac e , and include scales aiming o co e bo h posi i e and nega i e
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0205892 Oc obe 15, 2018 1 / 11
a1111111111
a1111111111
a1111111111
a1111111111
a1111111111
OPEN ACCESS
Ci a ion: And ade E, Rod ı
´guez D (2018) Fac o
s uc u e o mood o e ime ames and
ci cums ances o measu emen : Two s udies on
he P o ile o Mood S a es ques ionnai e. PLoS
ONE 13(10): e0205892. h ps://doi.o g/10.1371/
jou nal.pone.0205892
Edi o : Agus in Ma inez Molina, Uni e si y o
Za agoza, SPAIN
Recei ed: July 18, 2017
Accep ed: Oc obe 3, 2018
Published: Oc obe 15, 2018
Copy igh : ©2018 And ade, Rod ı
´guez. This is an
open access a icle dis ibu ed unde he e ms o
he C ea i e Commons A ibu ion License, which
pe mi s un es ic ed use, dis ibu ion, and
ep oduc ion in any medium, p o ided he o iginal
au ho and sou ce a e c edi ed.
Da a A ailabili y S a emen : The da a iles used in
ou esea ch a e al eady a ailable in an ins i u ional
eposi o y, MINERVA. The pe manen iden i ie is
he ollowing: h p://hdl.handle.ne /10347/17384.
Funding: This esea ch was suppo ed in pa by
he Galician Depa men o Cul u e, Educa ion and
Uni e si ies [g an numbe ED431B2016-017 (wi h
coope a ion and English s yle e iew o he
manusc ip unde g an GPC2014-047)]. The i s
au ho belongs o he CRETUS S a egic
Pa ne ship [g an numbe AGRUP2015/02]. All
componen s [2]. In hese scales indi iduals use adjec i es o desc ibe hei cu en mood,
which is p o iled in e ms o mul iple speci ic s a es.
One o he mos popula mood s a es measu es is he P o ile o Mood S a es (POMS). The
S anda d English e sion [3] comp ises 65 i ems a ed on a i e-poin esponse o ma . Pa ici-
pan s can selec a numbe om 0 (no a all) o 4 (ex emely). Fi y-eigh o he o iginal i ems
yield sco es on six ac o s: Tension, Dep ession (Dep essed mood), Ange , Vigou , Fa igue,
and Con usion. Ano he se en i ems ep esen F iendliness.
F iendliness was di icul o eplica e in some alida ion s udies conduc ed by he au ho s
o he POMS [3]. This is he eason why esea che s ha e sys ema ically excluded he i ems
ep esen ing he se en h mood s a e, which has os e ed he c i icism o he ine ec i eness o
he POMS o dealing wi h posi i e, pleasan eelings [4]. Recen edi ions o he o iginal es ,
howe e , asse ha F iendliness can be conside ed as a sepa a e indica o o in e pe sonal
unc ioning [5].
The li e a u e on he POMS documen s i s use o measu ing mood s a es in di e se esea ch
and applied ields o psychology, and e lec s he g ea popula i y o his ins umen in he ield
o spo [6]. In he POMS a hle es show a cha ac e is ic pa e n o sco es, widely known as he
icebe g p o ile, a model ha may help p edic spo s pe o mance [1,2]. Lane and Te y [1]
de eloped a concep ual schema o he e ec o mood s a es on pe o mance based on he in e -
ac ions among he POMS’ ac o sco es. Thei heo e ical model emphasised he ela ionships
o dep essed mood s a e wi h o he unpleasan s a es, pa icula ly ension and ange .
The POMS is a ailable o use on bo h adul s and adolescen s, and he ques ionnai e has
been ansla ed in o mo e han i een languages. F om a me hodological pe spec i e, i s in e -
nal s uc u e seems ela i ely es ablished, wi h he excep ion o he Con usion ac o [7–9],
which was ega ded as a cogni i e s a e [2]. Mo eo e , despi e he limi ed ange o pleasan
mood s a es co e ed, o he a o emen ioned easons, ew adap a ions e ained he componen
o F iendliness [4].
Conce ning he ins uc ions, he eques s ‘How ha e you el du ing he pas week includ-
ing oday?’ and ‘How do you eel igh now?’ a e he mos commonly used in he POMS.
McNai e al. [3] chose he e e ence o he pas week since hey conside ed his was a long
enough pe iod o cap u e people’s ypical and pe sis en emo ional eac ions o daily li e
e en s, ye sho enough o assess he acu e e ec s o a ea men . None heless, hey also indi-
ca ed he possibili y o using o he ins uc ions, depending on he pu pose o he s udy. In one
ac o analysis, McNai and co-wo ke s eplica ed he o iginal s uc u e by using he ‘ igh
now’ ins uc ion wi h a sample o college s uden s [3].
Some esea che s ha e examined he consequences o swi ching e e ence imes in mood
assessmen . Acco ding o Wa son [10], he s uc u e o posi i e and nega i e a ec ac o s
eme ged ega dless o he ime ame a ed.
Reg e ully, mos o he subsequen inqui ies ocused on obse ed ac o means, and we e
no p eceded by an analysis o measu emen in a iance. The esul s we e, ne e heless, o
in e es , and se ed o jus i y he selec ion o he ype o ins uc ions. As s a ed by Winkielman,
Knaupe , and Schwa z [11], people can adjus hei e e ence pe iods o a e hei mood s a es.
Thus, he scope o he e e ence pe iod will a ec an indi idual’s pe cep ion o in ensi y, as
well as o se iousness and equency o he episode unde e alua ion [11,12].
This impac on mood s a e a ings was iden i ied in he Anglo-Saxon con ex bo h in adul s
and in adolescen s. Fo ins ance, Te y, S e ens, and Lane [13] ound ha he sco es ob ained
in assessmen s o 135 school child en based on he ‘pas week’ we e highe han he a e age
sco es ob ained in mul iple ‘ igh now’ assessmen s. Recall o mood appea ed o be in luenced
by mood a he ime o ecall and possible signi ican e en s. So Te y e al. [13] sugges ed ha
he ‘ igh now’ esponse ime ame should be he me hod o choice.
Fac o s uc u e o mood o e ime ames and ci cums ances o measu emen
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0205892 Oc obe 15, 2018 2 / 11
hese p og ammes a e co- unded by he Eu opean
Regional De elopmen Fund o he Eu opean
Union.
Compe ing in e es s: The au ho s ha e decla ed
ha no compe ing in e es s exis .
As a complemen a y a gumen suppo ing he ins uc ion ‘ igh now’, some au ho s a i m
ha his empo al e e ence o ‘ igh now’ is mo e sensi i e o demands o he si ua ion. In he
spo s se ing, i is he e o e belie ed mo e use ul o assessing he ela ionship be ween mood
s a es and subsequen pe o mance [14,15].
Adap a ion o he ype o ins uc ions o he esea ch objec i e seems easonable. Howe e ,
he impac o changes in e e ence pe iod on psychome ic p ope ies ela ed o he use o he
ins umen mus be es ed. Thus, he i s s ep should be o demons a e he deg ee o in a i-
ance o he measu emen model ac oss ime ames. O he wise, we could judge a pe son’s
sco e o make compa isons among g oup means which a e no suppo ed by he unde lying
s uc u e.
P e ious esea ch [10,11,13,15] has iden i ied addi ional ac o s (e.g., di e se mood s a e
desc ip o s, esponse o ma s, o ci cums ances o assessmen ) ha could a ec he unde -
s anding and compa abili y o he mood s a e esponses. O all he po en ial modula o s, one
a iable s ands ou : he ci cums ances o mood s a e assessmen , i.e., he condi ions o ime
and place unde which he mood s a e esponse is egis e ed. In his espec , he POMS ques-
ionnai e has been adminis e ed be o e and a e compe i ion, and ou side he compe i ion
con ex [15,16].
Te y and Lane [15] specula ed ha mood s a es would di e signi ican ly depending on
he condi ions in which he a hle es we e assessed. Te y, Lane and Foga y [16] examined
whe he he s uc u e o he POMS could be ep oduced in dispa a e samples o adolescen
and adul s in wo di e en si ua ions: be o e compe i ion and in he class oom se ing.
Al hough in he expec ed di ec ion, he esul s o mul i-g oup con i ma o y ac o analyses
we e adequa e only when ac o s uc u e was cons ained o be equal ac oss he di e en sam-
ples. The i s a is ics we e conside ed ma ginal when ac o loadings and ac o co a iances
we e also es ic ed o be equal. To ou knowledge he e we e no u he p obes o in a iance
o he English POMS o e di e en assessmen ci cums ances.
Bo h he ype o ins uc ion and he ci cums ances o adminis a ion o he es a e impo -
an measu emen elemen s.
As ega ds he Spanish con ex , wo e sions o he POMS ha e been published ecen ly;
hese a e adap a ions o he ques ionnai e o use wi h adul a hle es and he gene al popula-
ion [17] and wi h adolescen a hle es [18]. Howe e , wo di e en se s o ins uc ions ha e
been applied, and he ques ionnai e has been adminis e ed du ing aining and jus p io o
compe i ion [19]. In he absence o p obes o in a iance, i will be di icul o es ablish whe he
mood s a e de ia ions be ween ins uc ion g oups o o e ime a e due o eal di e ences o o
changes in he cons uc measu e [20].
Ou aim is o ackle he p oblem by analysing hese wo impo an psychome ic aspec s o
he use o he POMS: he ime- ame ins uc ions, and he ci cums ances o adminis a ion o
he ques ionnai e.
We epo he esul s o wo s udies in which he POMS was adminis e ed o Spanish a h-
le es. In he i s s udy, we compa ed he ac o s uc u e o he ques ionnai e wi h da a
ob ained using wo di e en se s o ins uc ions: ‘ igh now’ and ‘pas week’, du ing aining.
In he second s udy, we compa ed he ac o s uc u e o he ques ionnai e wi h da a ob ained
using he ins uc ion ‘ igh now’ a wo di e en ci cums ances: ‘ aining’ and ‘compe i ion’.
Me hod
Pa icipan s
The da a we e o iginally eco ded o he pu poses o imp o ing he ins umen s. Analysis o
he ime ame was pe o med by compa ing he esponses o wo g oups o Spanish adul
Fac o s uc u e o mood o e ime ames and ci cums ances o measu emen
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0205892 Oc obe 15, 2018 3 / 11
a hle es. The sample comp ised 700 indi iduals: 428 (61.1%) men and 272 (38.9%) women, o
mean age 22.51 yea s (SD = 5.53, ange = 16–53).
Da a on yea s o educa ion, ype o spo and compe i i e le el we e no always a ailable,
because hey we e no pa o he in o ma ion collec ed in hose cases. Howe e , he a ailable
in o ma ion shows he p opo ions o pa icipan s who had comple ed o we e unde aking
he ollowing ypes o educa ion: uni e si y cou ses (65.15%), pos -compulso y schooling
(13.18%), oca ional s udies (11.86%) and compulso y educa ion (9.81%). The pe cen age o
pa icipan s in ol ed in di e en ypes o spo was as ollows: socce (37.4%), baske ball
(9.7%), owing (26.6%), chess (8.6%), a hle ics (4.3%), indoo socce (3.7%) and encing
(3.7%). The emaining 6% o subjec s pa icipa ed in less well ep esen ed spo s such as
hockey, ennis, swimming, wa e polo, cycling and handball. O he sample, 68.6% compe ed
a egional le el and 31.4% a na ional le el.
A small numbe o pa icipan s we e s ill in adolescence. I migh be app op ia e o cla i y
ha , adop ing he same a gumen gi en by he au ho s o he ins umen ha was used [17],
hose a hle es we e kep in he s udy, since hey ained and compe ed wi hin he op ca ego y
eams, and hey did no show ex eme alues in he a iables o in e es .
The g oup who ecei ed he ins uc ion ‘ igh now’ comp ised 350 pa icipan s: 182 (52%)
men and 168 (48%) women. Thei mean age was 24.04 yea s (SD = 7.25, ange = 16–53).
The g oup who ecei ed he ins uc ion ‘pas week’ comp ised 350 pa icipan s: 246
(70.3%) men and 104 (29.7%) women. Thei mean age was 21.03 yea s (SD = 2.20,
ange = 17–33).
We es ima ed he di e ences be ween he g oups in e ms o sex and age. This di e ences
we e signi ican , wi h low o mode a e e ec sizes: o sex, X
2
(1) = 24.63, p <.01, η
2
= .035; o
age, F(1, 687) = 55.36, p <.01, η
2
= .075.
Analysis o he ci cums ances o measu emen was pe o med wi h da a om wo g oups
o Spanish adolescen a hle es who ecei ed he same ins uc ion: ‘ igh now’. The g oups com-
p ised a o al o 446 pa icipan s: 344 (77.1%) men and 102 (22.9%) women, aged be ween 13
and 18 yea s (M = 15.13, SD = 1.20). The dis ibu ion o pa icipan s pe spo was as ollows:
baske ball (50.67%), socce (14.35%), indoo socce (14.80%), olleyball (9.19%) and chess
(4.71%). The emaining 6.28% pa icipa ed in handball, wa e polo, encing and a hle ics. O
he sample, 48.5% compe ed in egional league di isions, and 51.5% compe ed in local
ca ego ies.
The i s g oup comp ised 223 a hle es measu ed du ing aining. One hund ed and se en
(71.7%) we e men and 63 (28.3%) we e women. Thei mean age was 14.86 yea s (SD = 1.20,
ange = 13–18).
The second g oup included 223 a hle es measu ed jus be o e compe i ion. One hund ed
and eigh y- ou (82.5%) we e men and 39 (17.5%) women. Thei mean age was 15.39 yea s
(SD = 1.14, ange = 13–18).
The di e ences be ween he g oups is e ms o sex and age we e signi ican , wi h low e ec
sizes: o sex, X
2
(1) = 7.32, p <.01, η
2
= .02; o age, F(1, 440) = 22.11, p <.01, η
2
= .05.
Ins umen s
We used a Spanish o m o he POMS ques ionnai e designed o adul a hle es o assess he
ype o ins uc ion [17]. This comp ised 30 e enly dis ibu ed i ems ( i e pe ac o ) in six
i s -o de ac o s: ension, dep ession, ange , igou , a igue and iendliness. Examples o
i ems included in he ques ionnai e we e he adjec i es ang y,ene ge ic o sad.
The esponse o ma consis ed o i e o de ed ca ego ies sco ed be ween 0 and 4. The psy-
chome ic p ope ies o his measu emen model in a hle es we e documen ed by And ade
Fac o s uc u e o mood o e ime ames and ci cums ances o measu emen
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0205892 Oc obe 15, 2018 4 / 11
e al. [17], who epo ed sa is ac o y alues o eliabili y (wi h consis ency es ima ions be ween
.79 and .87) and he ollowing goodness-o - i s a is ics: X
2
(390, N = 400) = 803.12, p <.01,
CFI = .97, NNFI = .97, RMSEA = .051, and SRMR = .057.
We used a Spanish e sion o he POMS ques ionnai e designed o adolescen a hle es o
compa e he ci cums ances o measu emen [18]. This comp ised 29 i ems, ou o a igue,
and i e each o he ac o s ension, dep ession, ange , igou and iendliness.
The i ems we e a ed in i e o de ed ca ego ies ha ecei ed sco es be ween 0 and 4.
And ade e al. [18] epo ed adequa e alues on he in e nal s uc u e o he es , which was
subjec ed o con i ma o y ac o analysis ( i alues o X
2
(361, N = 320) = 684.38, p <.001,
CFI = .96, NNFI = .96, RMSEA = .053, and SRMR = .061.), and on i s eliabili y, es ima ed by
he C onbach’s alpha s a is ic (wi h alues be ween .77 and .87).
P ocedu e
The esea ch in ol ed he collec i e, anonymous adminis a ion o es s. I did no equi e he
app o al o he Uni e si y bioe hics commi ee, which would be necessa y in case o ‘pe sonal
cha ac e da a subjec ed o a ese ed ea men , which a ec ed he igh s and libe y o peo-
ple, he in e es s linked o p ese a ion o en i onmen o o o he legally p o ec ed goods.’
The p ojec obse ed he e hical p inciples ega ding pa icipan s [21]. People esponsible o
he eams and pa icipan s ga e e bal consen and we e in o med abou : (a) he objec i e o
he s udy, expec ed du a ion and p ocedu es; (b) hei igh o decline o pa icipa e and o
wi hd aw om he esea ch once pa icipa ion had begun ( ha is, he pa icipa ion in he
s udy was olun a y); (c) he p o ec ion o con iden iali y; and (d) he oppo uni y o ask ques-
ions and ecei e answe s.
The POMS ques ionnai e was adminis e ed o bo h g oups o adul a hle es jus be o e a
usual aining session. Th ee spo psychologis s ga he ed he da a in si u. The ques ionnai e
was adminis e ed in a simila way in bo h g oups by ollowing a s anda dized p o ocol. How-
e e , he ype o ins uc ion was di e en . In one g oup, pa icipan s we e asked o espond
acco ding o how hey el a ha p ecise momen (‘ igh now’ ins uc ion). Pa icipan s in he
o he g oup we e asked o espond acco ding o how hey had el du ing he pas week,
including he day o he es (‘pas week’ ins uc ion).
Ano he h ee psychologis s adminis e ed he POMS ques ionnai e o he sample o adoles-
cen a hle es. The ype o ins uc ion was iden ical o he wo g oups, bu he ci cums ances o
adminis a ion o he ques ionnai e we e di e en . The ques ionnai e was gi en o he i s
g oup immedia ely be o e a aining session (‘ aining’ g oup); he same o m o he ques ion-
nai e was gi en o he second g oup wi hin he 30 minu es be o e a compe i ion (‘compe i ion’
g oup).
Da a analysis
The esul s a e p esen ed pe ques ion unde s udy, al hough he analy ical sequence was simi-
la . Fi s ly, we es ed he hypo hesised measu emen model in each g oup. The es ima ions
we e pe o med wi h LISREL 8.80, and sa is ac o y single-sample model- i was ep esen ed
abo e all by he con en ional alues o CFI and NNFI >.95, and RMSEA <.06. Secondly, we
conduc ed mul i-g oup con i ma o y ac o analyses by imposing successi e es ic ions on
he model. We planned o pe o m he ollowing es s: con igu al in a iance (equi alen pa -
e n o ixed and ee ac o loadings ac oss all g oups), me ic in a iance (equali y o ac o
loadings), scala in a iance (equali y o in e cep s), la en means in a iance, and in a iance o
ela ions among la en ac o s. We applied he scaled di e ence chi-squa e s a is ic o examine
whe he disc epancies in he successi e cons ained models we e signi ican [22,23].
Fac o s uc u e o mood o e ime ames and ci cums ances o measu emen
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0205892 Oc obe 15, 2018 5 / 11
Resul s
S udy on he ime ames
The da a used o compa e ime ames did no ul il he assump ion o mul i a ia e no mali y,
wi h ela i e ku osis alues o 1.216 in he ‘ igh now’ g oup and 1.120 in he ‘pas week’ g oup.
The obus maximum likelihood me hod was he e o e chosen as he es ima ion me hod [24].
The ac o s uc u e included 30 i ems, i e o each o he pos ula ed i s -o de ac o s.
The model was o e -iden i ied wi h 390 deg ees o eedom.
In he ‘ igh now’ g oup, he esul s o he goodness-o - i indices we e as ollows:
SBX
2
(390, N = 350) = 607.65, p <.001, CFI = .98, NNFI = .98, RMSEA = .040 (CI 90% [.034,
.046]) and SRMR = .058. The ange o he s anda dized ac o loadings was be ween .43 and
.88, and be ween .044 and .076 o ypical e o s. The in e - ac o co ela ions anged om .09
( ela ionship among igou and ension) o .82 ( ela ionship be ween ange and dep ession).
In he ‘pas week’ g oup, he esul s o he mos usual indices we e as ollows: SBX
2
(390,
N = 350) = 734.66, p <.001, CFI = .97, NNFI = .97, RMSEA = .050 (CI 90% [.045, .056]), and
SRMR = .074. S anda dized ac o loadings anged be ween .52 and .91, wi h ypical e o s
be ween .040 and .075. Wi h espec o in e - ac o co ela ions, he alues anged be ween .04
( o iendliness and a igue) and .64 ( o ange and ension).
As he posi ed s uc u e was accep able in bo h g oups, we conduc ed a simul aneous analy-
sis o con igu al equi alence. Bo h he numbe o ac o s and he ac o -indica o co espon-
dence a e he same. This solu ion (see Table 1) will se e as baseline o pos e io in a iance
p oo s.
We also es ed he mo e es ic i e model o me ic equi alence. We used he scaled di e -
ence in chi-squa es o compa e bo h models. The di e ence was signi ican , which p e en s us
om in e ing weak in a iance.
By compa ing he non-s anda dized ac o loadings o bo h g oups, oge he wi h modi i-
ca ion indices, we iden i ied he mos p oblema ic i ems (e.g. he i em weak, ep esen ing he
ac o a igue). We decided o p oceed in a con ex o pa ial in a iance. Howe e , any plausi-
ble nes ed model implied a signi ican dec ease in goodness-o - i .
S udy o he ci cums ances o measu emen
The CFA was conduc ed in he g oups ha we e assessed a wo dis inc ci cums ances wi h
he me hod o obus maximum likelihood, because he da a dis ibu ion was no mul i a ia e
Table 1. Resul s om single and mul i-g oup con i ma o y ac o analyses wi h espec o he ime ames.
Type o analysis SB X
2
NT X
2
ML X
2
d NT X
2
di
ML X
2
di
Δd CFI NNFI RMSEA (90% CI)
Single g oup
Righ now (n= 350) 607.65 738.78 757.16 390 .98 .98 .040 (.034, .046)
Las week (n= 350) 734.66 834.61 859.50 390 .97 .97 .050 (.045, .056)
Measu emen In a iance
Con igu a ion 1338.00 1573.40 1616.66 780 .98 .97 .045 (.041, .049)
Me ic in a iance 1437.76 1691.66 1730.08 810 100.22
a
30 .97 .97 .047 (.043, .051)
96.12
a
SB X
2
, Sa o a and Ben le Chi-Squa e; NT, No mal Theo y Weigh ed Leas Squa es S a is ic; ML, o dina y Maximum Likelihood S a is ic (Minimum Fi Func ion
Value); CFI, Compa a i e Fi Index; NNFI, Non-No med Fi Index; RMSEA, Roo Mean Squa e E o o App oxima ion; SRMR, S anda dized Roo Mean Squa e
Residual.
a
p<.05
h ps://doi.o g/10.1371/jou nal.pone.0205892. 001
Fac o s uc u e o mood o e ime ames and ci cums ances o measu emen
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0205892 Oc obe 15, 2018 6 / 11
no mal. The alues o ela i e mul i a ia e ku osis we e 1.223 in he ‘ aining’ g oup and
1.221 in he ‘compe i ion’ g oup.
The measu emen model, wi h 29 i ems and six i s -o de ac o s, was also o e -iden i ied,
wi h 362 deg ees o eedom.
In he ‘ aining’ g oup, he esul s o he goodness-o - i indices we e as ollows: SBX
2
(362,
N = 223) = 535.47, p <.001, CFI = .97, NNFI = .97, RMSEA = .046 (IC 90% [.038, .055]) and
SRMR = .070. S anda dized ac o loadings anged be ween .34 and .90, wi h ypical e o s
be ween .062 and .091. The co ela ions be ween la en ac o s adop ed absolu e alues
be ween .10 ( o iendliness and ension) and .63 ( o ange and ension).
In he ‘compe i ion’ g oup, he esul s o he i indices we e as ollows: SBX
2
(362,
N = 223) = 558.02, p <.001, CFI = .95, NNFI = .94, RMSEA = .049 (IC 90% [.041, .057]) and
SRMR = .080. The ac o loadings anged be ween .13 ( o ca e ee) and .86 ( i ed), wi h ypi-
cal e o s be ween .063 and .096. The co ela ions be ween la en ac o s yielded alues
be ween .03 ( o iendliness and igou wi h ension) and .51 ( o iendliness and igou ).
The simul aneous analysis o con igu al equi alence (Table 2) showed an accep able da a
i . The solu ion o equali y o ac o loadings did no p oduce a signi ican dec ease in he
goodness-o - i , which allows us o assume me ic equi alence. Responden s ac oss g oups
a ibu e he same meaning o he la en cons uc unde s udy.
Howe e , he mo e es ic i e model o scala equi alence indica ed a signi ican inc ease
in he di e ence be ween chi-squa es. The e o e, scala in a iance was no obse ed wi h
hese samples.
Gi en ha ac o weigh s and he indica o s in e cep s a e no in a iable in bo h samples,
he di e ence in he la en means be ween hem canno be in e p e ed.
Discussion
The aim o his pape was o examine he in a iance o he ac o s uc u e o he POMS ques-
ionnai e ac oss he ype o ins uc ion and he ci cums ances in which he measu e was aken.
Bo h aspec s a e ele an o meaning ul in e p e a ion o sco es in sel - epo s.
To assess mood s a es, esea che s ask pa icipan s o a e he in ensi y o equency wi h
which hey had expe ienced ce ain eelings du ing a pe iod o ime. Howe e , in hese
Table 2. Resul s om single and mul i-g oup con i ma o y ac o analyses wi h espec o he ci cums ances o measu emen .
Type o analysis SB X
2
NT X
2
ML X
2
d NT X
2
di
ML X
2
di
Δd CFI NNFI RMSEA (90% CI)
Single g oup
T aining (n= 223) 535.47 679.20 657.09 362 .97 .97 .046 (.038, .055)
Compe i ion (n= 223) 558.02 703.14 716.69 362 .95 .94 .049 (.041, .057)
Measu emen In a iance
Con igu a ion 1093.41 1382.34 1373.78 724 .96 .96 .048 (.042, .054)
Me ic in a iance 1138.22 1452.21 1434.25 753 39.70
a
29 .96 .96 .048 (.042, .054)
34.36
a
Scala in a iance 1195.10 1503.34 1483.16 776 83.82
b
23 .95 .95 .049 (.044, .055)
80.18
b
SB X
2
, Sa o a and Ben le Chi-Squa e; NT, No mal Theo y Weigh ed Leas Squa es S a is ic; ML, o dina y Maximum Likelihood S a is ic (Minimum Fi Func ion
Value); CFI, Compa a i e Fi Index; NNFI, Non-No med Fi Index; RMSEA, Roo Mean Squa e E o o App oxima ion; SRMR, S anda dized Roo Mean Squa e
Residual.
a
p>.05.
h ps://doi.o g/10.1371/jou nal.pone.0205892. 002
Fac o s uc u e o mood o e ime ames and ci cums ances o measu emen
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0205892 Oc obe 15, 2018 7 / 11
si ua ions he same wo d may ep esen a wide ange o dis inc expe iences. I he meaning o
he wo d is no speci ied by he esea che , pa icipan s migh in e i om some cha ac e is-
ics o he measu emen ins umen . One such cha ac e is ic is he leng h o he e e ence
pe iod [11,12].
F om he esul s o p e ious s udies wi h he POMS in Spanish [17,18], i seems ha he
use o one o ano he ype o ins uc ion does no comp omise he gene al psychome ic qual-
i y o he measu e. None heless, his does no imply ha he cha ac e o he sco es and he
in o ma ion hey p o ide o esea che s and p o essionals is he same in all cases.
In ou i s analysis, we compa ed he concu en (‘ igh now’ ins uc ion) and e ospec-
i e (‘pas week’ ins uc ion) epo s p o ided by he Spanish POMS. By ollowing s ic good-
ness-o - i c i e ia, we ound ha he e was con igu al equi alence o e he wo se s o
ins uc ions. Howe e me ic in a iance was no endo sed. In o he wo ds, bo h he numbe
o ac o s and he ac o -indica o co espondence a e he same, bu he uns anda dized ac o
loadings o each indica o a e no equal ac oss he compa ed g oups.
F om a me hodological poin o iew, Te y e al. [16] ci ed among he possible causes o
mis i some adjec i es ep esen ing a igue. Simila ly, he i em weak was one o he mos p ob-
lema ic ma ke s when es ing me ic in a iance in ou s udy.
I con igu al in a iance is e i ied bu me ic in a iance is no , one can add ess i by iden i-
ying and eleasing p oblema ic indica o s i he e a e empi ical and subs an i e easons o do
so [25]. We conside ed his possibili y, bu did no ind a plausible model which s a is ically
sa is ied he c i e ia o goodness-o - i . On he o he hand, pa ial measu emen in a iance
could be highly dependen on chance cha ac e is ics o he obse ed samples, and i s conse-
quences on eliabili y and es ima ion o la en pa ame e s a e ye o be de e mined [20].
As pe ou i s examina ion, we mus conclude ha he esponden s ac oss g oups did
no a ibu e he same meaning o he la en cons uc unde s udy. This ou come suppo s he
need o sepa a e no ma i e alues o obse ed sco es ga he ed wi h each se o ins uc ions.
F om a concep ual pe spec i e, a possible explana ion o he lack o me ic equi alence could
be ha he ‘pas week’ e e ence is a measu e o long-las ing mood s a es [26], while he ‘ igh
now’ e e ence ep esen s si ua ionally-sensi i e s a es [15,27]. Al hough adhe ence o disc e e
s a e o ai concep s is an o e simpli ied app oach o he measu emen o mood s a e dimen-
sions, i should be one o he conside a ions du ing decision on which ime ame o choose.
Ano he aspec o g ea impo ance ega ding de use o he POMS a e he ci cums ances o
adminis a ion o he ques ionnai e. Depending on he su ey objec i es, he si ua ion would
some imes equi e p oximi y be ween he measu e o mood s a es and he s ess ul e en ; a
o he imes he esea che would be in e es ed in assessing mood s a es du ing he aining
phase, o moni o mood s a es a he han o making decisions abou immedia e in e en ion.
As o he esul s ob ained in ou second analysis, i appea s ha e en in ela i ely small sam-
ples he e is me ic equi alence ac oss he wo measu emen momen s, aining and compe i-
ion. Thus, he cons uc s a e mani es ed he same way in each g oup.
Howe e , scala in a iance was no ul illed, so we canno assume ha we would ob ain he
same obse ed sco e o an indica o a a gi en le el o he la en ac o . G oups canno be
compa ed on hei sco es on he la en a iable.
We judge his inding cong uen wi h pe sonali y psychology which concei es beha iou
o e ime as consis ing o densi y dis ibu ions o s a es [28]. Di e ences in beha iou a e also
bes desc ibed as densi y dis ibu ions, and he indi idual di e ences ha show egula i y a e
hose dis ibu ions o s a es a he han single sco es o only he mean o ano he unique
pa ame e o he dis ibu ions.
De ini ely, new sys ema ic in es iga ion is needed o es hypo heses abou he aspec s
which a ec mood s a es measu emen in gene al and abou in a iance ela i e o he
Fac o s uc u e o mood o e ime ames and ci cums ances o measu emen
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0205892 Oc obe 15, 2018 8 / 11
applica ion o he POMS ques ionnai e in pa icula . This pape ep esen s a i s app oach,
and is no in ended o o e a inal wo d on he ma e .
I also holds some sho comings ha should ecei e a en ion in u u e esea ch. One weak
poin o ou wo k is ha mood s a es sco es based on he ‘pas week’ cons i u e a e ospec i e
measu e, and ac o s po en ially causing memo y dis o ions could no be con olled [13].
Resea ch would also bene i om designs in which he same g oup o pa icipan s assessed
hei mood s a es unde he wo di e en se s o ins uc ions o o e ime. The e a e ew an e-
ceden s in he li e a u e, and hei esul s a e di e se. Galinha, Pe ei a, and Es e es [29] es ed
he empo al in a iance o he PANAS (Posi i e and Nega i e A ec Schedule), using he
s uc u al equa ion modelling analysis, in a g oup o adul s. As an in e es ing esul , al hough
hey used i s s a e e sion, he PANAS scale showed s abili y o e a wo-mon h in e al. In
con as , F ied e al. [30] in es iga ed unidimensionali y and measu emen in a iance o ou
common dep ession a ing scales (one sel - epo , h ee clinician- epo s) in wo la ge s udies.
Fo all ins umen s, nei he unidimensionali y no measu emen in a iance we e judged as
accep able.
Addi ionally, we could ha e conside ed less es ic i e c i e ia o judge each mul i-g oup
con i ma o y ac o analysis model. As Cheung and Rens old [31] poin ed ou , he ou come o
he chi-squa e di e ence es could indica e he lack o measu emen in a iance e en when
he imposed equali y cons ain s lead o mino dec eases in model i . One way o handle his
is o compa e he uns anda dized pa ame e es ima es ac oss he solu ions. Ano he is o
e iew changes in al e na i e i indices. Fo simplici y, only app oxima e i indices we e
shown in he ables. Cheung and Rens old [31] sugges ed ha a change o jus .01 in CFI alues
mean ha he null hypo hesis o in a iance should no be ejec ed. Meade, Johnson, and
B addy [32] sugges ed ha a change o .002 in CFI alues implied jus i ial de ia ions om
pe ec measu emen in a iance. Kang, McNeish, and Hancock [33] suppo ed he inc emen
in McDonald’s NCI as he index o choice.
The samples in ou s udies we e no as la ge as hose simula ed in, o ins ance, Meade
e al.’s pape . So we based ou conclusions on he di e ence be ween chi-squa es a he han
on al e na i e i indica o s, hus adop ing a mo e conse a i e a i ude a his s age o esea ch
on he opic.
Conclusion
This pape emphasizes he need o s udies o equi alence be o e using a single sel - epo
wi h mo e han one se o ins uc ions, o unde di e se ci cums ances. Pa icula ly wi h
espec o ins uc ions, i shows ha i we adminis e a mood s a es es wi h wo dis inc ime
ames (e.g. ‘ igh now’ and ‘pas week’), we may be measu ing he same concep ual dimen-
sions, bu wi h no gua an ee o doing i wi h he same meaning.
Au ho Con ibu ions
Concep ualiza ion: Elena And ade, Dolo es Rod ı
´guez.
Fo mal analysis: Elena And ade, Dolo es Rod ı
´guez.
Resou ces: Elena And ade, Dolo es Rod ı
´guez.
Supe ision: Elena And ade.
Visualiza ion: Elena And ade, Dolo es Rod ı
´guez.
W i ing – o iginal d a : Elena And ade, Dolo es Rod ı
´guez.
Fac o s uc u e o mood o e ime ames and ci cums ances o measu emen
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0205892 Oc obe 15, 2018 9 / 11
