www.ssoar.info
What determines a "perspective"? : contrast effects
as a function of the dimension tapped by preceding
questions
Schwarz, Norbert; Münkel, Thomas; Hippler, Hans-Jürgen
Veröffentlichungsversion / Published Version
Arbeitspapier / working paper
Zur Verfügung gestellt in Kooperation mit / provided in cooperation with:
GESIS - Leibniz-Institut für Sozialwissenschaften
Empfohlene Zitierung / Suggested Citation:
Schwarz, Norbert ; Münkel, Thomas ; Hippler, Hans-Jürgen ; Zentrum für Umfragen, Methoden und Analysen ZUMA- (Ed.): What determines a "perspective"? : contrast effects as a function of the dimension tapped by preceding
questions. Mannheim, 1989 (ZUMA-Arbeitsbericht 1989/22). URN: http://nbn-resolving.de/urn:nbn:de:0168ssoar-67053
Nutzungsbedingungen:
Dieser Text wird unter einer Deposit-Lizenz (Keine
Weiterverbreitung - keine Bearbeitung) zur Verfügung gestellt.
Gewährt wird ein nicht exklusives, nicht übertragbares,
persönliches und beschränktes Recht auf Nutzung dieses
Dokuments. Dieses Dokument ist ausschließlich für
den persönlichen, nicht-kommerziellen Gebrauch bestimmt.
Auf sämtlichen Kopien dieses Dokuments müssen alle
Urheberrechtshinweise und sonstigen Hinweise auf gesetzlichen
Schutz beibehalten werden. Sie dürfen dieses Dokument
nicht in irgendeiner Weise abändern, noch dürfen Sie
dieses Dokument für öffentliche oder kommerzielle Zwecke
vervielfältigen, öffentlich ausstellen, aufführen, vertreiben oder
anderweitig nutzen.
Mit der Verwendung dieses Dokuments erkennen Sie die
Nutzungsbedingungen an.
Terms of use:
This document is made available under Deposit Licence (No
Redistribution - no modifications). We grant a non-exclusive, nontransferable, individual and limited right to using this document.
This document is solely intended for your personal, noncommercial use. All of the copies of this documents must retain
all copyright information and other information regarding legal
protection. You are not allowed to alter this document in any
way, to copy it for public or commercial purposes, to exhibit the
document in public, to perform, distribute or otherwise use the
document in public.
By using this particular document, you accept the above-stated
conditions of use.
What determines a 'perspective'?
Contrast effects as a function of the
dimension tapped by preceding questions
Norbert Schwarz, Thomas Munkel
Hans-J. Hippier
ZUMA-Arbeitsbericht N r. 09/22
Zentrum für Umfragen, Methoden und
Analysen e.V. (ZUMA)
Postfach 12 21 55
D-6800 Mannheim 1
Seit Juli 1 983 sind die Z U M A - A r b e i t s b e r i c h t e
in zwei R e i h e n aufgeteilt:
D ie Z U M A - A r b e i t s b e r i c h t e (neue F o l g e ) h a b e n
e in e haus i n t er ne Begutachtung d u r c h l a u f e n und
w e r d e n v o m G e s c h ä f t s führenden D i r e k t o r z u s a m ­
m e n m i t den ü b r i g e n Wissenschaftlichen L e i ­
t e r n h e r a u s g e g e b e n . Die Berichte d i e se r R e i h e
s i n d zur a l l g e m e i n e n Weiterg abe n ach außen
bestimmt.
D i e Z Ü H A - T e c h n i s c h e n Berichte d ien e n zur
h a u s i n t e r n e n K o m m u n i k a t i o n bzw. z u r U n t e r ­
r i c h t u n g e x t e r n e r K o o p e r a t i o n s p a r t n e r . Sie
sind nicht zur allgemeinen W e i t e r g a b e b e ­
stimmt .
The a t t a c h e d r e p r i n t r e p l a c e s Z U M A - A r b e i t s b e r i c h t No.
the s a m e authors.
89/22 by
N o r b e r t Schwarz, T h o m a s Miinkel, & H ans-J. H i p p i e r (1990).
W h a t d e t e r m i n e s a " p e r s p e c t i v e " ? C o n t r a s t e f f e c t s as a
f u n c t i o n of t h e d i m e n s i o n t a p p e d by p r e c e d i n g q u es ti o ns .
E u r o p e a n J o u r n a l of S o c ia l P s y c h o l o g y . 2 0 . 357-361.
European Journal o f Social Psychology, Vol. 20, 357-361 (1990)
Short Note
What determines a 'perspective'?
Contrast effects as a function of the
dimension tapped by preceding questions
NORBERT SCHWARZ
Z entrum für Umfragen, M ethoden und Analysen,
ZUM A, Mannheim, FRG
THOMAS MÜNKEL
Universität Heidelberg, Heidelberg, FRG
and
HANS-J. HIPPLER
Zentrum für Umfragen, M ethoden und Analysen,
ZUM A, Mannheim, FRG
Abstract
Subjects estim ated how m any Germans drink vodka or beer, or estim ated the caloric
content o f these drinks. The fo rm er judgm ent, but not the latter, produced contrast
effects on subsequent ratings o f how 'typically German ’ various drinks are. Thus, highly
accessible extrem e stim uli did only affect ratings i f the first judgm ent pertained to
the sam e underlying dimension.
INTRODUCTION
That ratings o f a stimulus along a dimension are a function o f the extremity o f
eonlext stimuli along the same dimension is one o f the best established findings
in social judgment research ( c f . Eiser (in press) for an extensive review o f current
and classic research). Different theories o f social judgment share the assumption
that judges use the range o f stimuli to anchor the response scale provided to them
A ddressee for corresp on d en ce Dr Norbert Schwarz, Z U M A . P.O. Box 122155, D -6800 M annheim. W,
G erm any
T h e reported research is based in pari on the second author's diplom a thesis a t the U niversity o f
H eid elb erg , conducted under the direction or the first author. It was s u p p o rted by gram s SW F004-6
from the B un desm inisterium für F o rsch u n g u n d T echnologie to N orbert Schwarz, and Sir 264/2 from
the D eutsche Forschungsgem einschaft to F ritz Strack and N orbert Schwarz. We thank Tom Ostrom
for his stim ulating com m ents on a previous draft.
0046-2772/90/040357-05505.00
© 1990 by John Wiley & Sons, Ltd.
Received 7 October 1989
A ccepted 20 M arch ¡990
358
A'. Schwarz, T. M unketandH -J. Hippier
(e.g. V olkm an, 1 9 5 l;O stro m a n d U pshaw , 1968; Parducci, 1963). A ccord in gly,in tro­
ducing a m ore extrem e stim ulus in the set o f to-be-evaluated stim uli has been found
to result in contrast effects in ratings o f m oderate stim uli. T his presum ably reflects
that the introduction o f an extrem e stim ulus extends ju d g es’ ‘perspective’ (Ostrom
and U pshaw , 1968), and that judges use the extrem e stim uli represented in their
perspective to anchor the scale.
H owever, il is not well understood what is necessary for an extrem e stim ulus
to becom e part o f a judge's perspective. Is it necessary for the extrem e stim ulus
to be explicitly included in the set o f to-be-evaluated stim uli, as has typically been
the case in previous research, or is il sufficient that the extrem e stim ulus sim ply
com es to mind w hen the jud ge evaluates less extreme stimuli? T he currently available
evidence is restricted to a psychophysical study by Brown (1953). He asked judges
to lift different w eights and to estim ate their heaviness. N o t surprisingly, he found
that a given stim ulus was rated as less heavy when preceded by an extrem ely heavy
on e. This effect w as more pronounced when subjects had to rate the anchor stim ulus
than when they had not, but was still evident under the latter con d ition . M ore im por­
tantly, in other con d ition s o f his experim ent, subjects were passed a tray with weights
equal in total weight to the extrem e stim ulus. Lifting this tray did result in slight
contrast effects when the weight o f the tray had to be judged, but did not influence
subsequent judgem ents w hen no explicit rating o f the tray was required. Brown
(1953, p. 210) concluded that ‘the anchor, to be effective, must be perceived as
a member o f the sam e class o f objects as the other w eigh ts’. Presum ably, a heavy
anchor w eight that resem bled the target stim uli in its appearance was sp ontaneously
categorized as a m ember o f the relevant class, even when it was not explicitly judged,
whereas the tray w as not. A ccordingly, the former, but n ot the latter, w as included
in judges' perspective, resulting in contrast effects on subsequent judgem ents.
In addition to its theoretical interest, the issue o f w hal determ ines inclusion in
a 'perspective' has important im plications for questionnaire construction in social
and psychological research. Suppose, for exam ple, that judges are asked to rale
different beverages according to how typically ‘G erm an’ they are. Suppose further,
that m oderately typical target beverages (such as wine or coffee) are preceded either
by a highly typical context beverage (such as beer), or by a highly atypical one
(such as vodka). In that case, we m ay expect that the m oderately typical targets
are rated as less typically ‘G erm an’ iT preceded by beer, than if preceded by vodka.
In fact, N oelle-N eu m ann (1970) found that a number o f food item s (such as noodles
or potatoes) were considered m ore typically Germ an if preceded by rice, than if
not. H owever, is it necessary for the em ergence o f contrast effects o f this type that
the extrem e stim uli are presented on the sam e list as the m oderate ones? Or is il
sufficient that they com e to mind w hen ihe jud ge evaluates the m oderate stimuli?
In the latter case, a n y preceding question lh al increases the cognitive accessibility
o f the extrem e stim uli m ay be sufficient lo elicit contrast effects even under conditions
where the extrem e stim uli are not included in the list.
Brown's (1953) theorizing is am biguous in this respect, due to im precision o f his
class concept. On the on e hand, the extrem ely typical or atypical beverages are
clearly members o f the natural class o f beverages, as are the m oderate ones. On
the other hand, the extrem e beverages are not m em bers o f the class o f the to-beevaluated beverages if they are n ot presented on the sam e list. In the latter regard,
the finding that lifting the tray elicited slight contrast effects in B row n’s study if
P erspective
359
its weight had to be judged, suggests thal a stim ulus that is not sp ontaneously c o n ­
sidered member o f the sam e class may be included in judges* perspective if it is
linked 10 the dim ension o f judgem ent. I f so, extrem ely typical or atypical beverages,
m entioned in a preceding question, may be included in ju d g es’ perspective if the
preceding question links them to the dim ension alon g which the subsequent stimuli
are to be evaluated. If the preceding question does no! provide linkage, simply increas­
ing the cogn itive accessibility o f the extrem e beverages m ay have little impact on
subsequent ratings. A ccordingly, elTccts o f preceding questions on subsequent jud ge­
m ents m ay only be expected under very specific conditions.
We explored these possibilities in a 2 (beer versus vodka as extrem e stim ulus)
x 3 (list, consum ption question, caloric content question)-faclorial between subjects
design. All subjects were asked lo rate dilTerent target beverages (wine, coffee, and
m ilk) according to h ow ‘typically G erm an' they are. In the list condition, ratings
o f these beverages were either preceded by typicality ratings o f b e e r ’ (a high typicality
drink) or o f ‘vodka' (a low typicality drink). We assum e that contrast effects will
be obtained under this con d ition , with the target drinks being rated us less typically
‘G erm an’ if preceded by beer, than iT preceded by vodka. In the other con d ition s,
the extrem e stim uli were n ot included in the list, but subjects' attention was drawn
lo them by preceding q uestion. Subjects' were either asked to estim ate the frequency
with which G erm ans drink beer or vodka, respectively (consum ption condition), or
lo estim ate the caloric content o f a glass o f beer or vodka, respectively (caloric
content condition). W hile the frequency o f consum ption question taps the dim ension
o f typicality, this is not the case for the caloric content question. A ccordingly, a
com parison o f typicality ratings u n d erlist, consum ption question, and caloric content
question con d ition s allow s an exploration o f the con d ition s under which extreme
stim uli do becom e part o f a jud ge's perspective.
Specifically, if extrem e stim uli are only included in the perspective if they are
presented on the sam e list as the target stim uli, differences in ratings o f the target
stim uli as a function o f the context stim uli should only be obtained under list co n ­
ditions. I f it is sufficient that the extrem e stim uli com e to m ind, on the other hand,
differences in ratings o f the target stimuli should be obtained under all conditions.
Finally, it may n ot be necessary that (he extrem e stimuli are presented on the sam e
list, but that they are thought about with regard to the sam e underlying dim ension
as the target stim uli, as assumed in the above 'linkage’ discussion. If so, differences
in ratings o f the target stim uli should em erge under consum ption question con d ition s,
but not under caloric content question con d ition s.
METHOD
O ne hundred and fifty-six students at the University o f M annheim and the University
o f K onstanz, W est G erm any, random ly assigned lo con d ition s, participated in a
survey adm inistered on PC (using the IBIS interviewing softw are, cf. H ippier, M eier
and Schwarz, 1988). A s part o f this survey, all respondents were asked to rate wine,
cofTee, and milk according lo 'how typically Germ an' they are (I = not at all typical;
9 - very typical). T hese ratings constituted the dependent variable.
In the list conditions, these m oderately typical beverages were either preceded by
'beer' (a prototypically G erm an drink), or by ‘vod k a’ (an atypical drink), which
360
N . Schw arz. T. M u n kcl anil H -J. H ippier
were lo be rated alon g the sam e typicality scale. In the consum ption question conditions,
respondents were asked to rale the frequency with which G erm ans drink beer (or
vod k a, respectively), alon g a scale from 1 = rarely, to 7 = very frequently, whereas
in the caloric content question conditions ihey rated the caloric content o f beer (or
vod k a, respectively) alon g a scale from 1 = low , to 9 = high in caloric content.
R E SU L T S
In the list con d ition , beer was rated as typically G erm an ( M = S.7), whereas vodka
was rated as atypical (M = 1.4), thus establishing the adequacy o f the extrem e stim uli.
T able 1 show s the mean ratings o f ihe target beverages (averaged over wine, colfee,
and m ilk)1 as a function o f experim ental con d ilion s. A s predicted, a significant
interaction o f the nature o f the context stim uli (beer versus vodka) and (heir m ode
o f introduction em erged, F ( 2, 150) = 4 .5 4 , p < 0.02.
Table 1. Mean ratings of turgei beverages as a function o f context
stimuli and presentation mode
Presentation mode
Context
stimulus
Lisl
Consumption
question
Calorie content
question
Beer
4.42
{ S D . * \ .0)
4.85
{S D = 0 6<J)
4.47
[5 D . ■= 1.15)
Vodka
5,40
(S.& =0.97)
5.29
( 5 D - 1.0)
4.28
(i'.O. = 1.03)
Mean ralmg averaged over three beverages is given
9 ™ very typical- N = 25 to 27 per cell.
I ” not at all npie.il.
Planned com parisons indicate thal the target beverages w ere rated as less typically
German if beer rather than vodka was presented as the lirsi stim ulus on the same
list, /(150 = 3.56, p < 0.001, providing a conceptual replication o f num erous previous
findings. The sam e holds Irue when subjects were asked to estim ate the percentage
o f Germ ans w h o drink beer or vodka in a preceding q uestion, t(150) = 1.94, p < 0.06
H ow ever, the m ean difference is som ew hat less pronounced than under list con d ilion s,
as is reflected in a m arginally significant result o f the respective interaction contrast,
f(l5 0 ) = 1.61, p < 0.10. Finally, estim ating the caloric content o f a glass o f beer
or vodka, respectively, did n oi alTect the ratings o f Ihe targel beverages, t < I,
DISCUSSION
T hese findings are consistent with the assum ption thal judges' p erspectives arc orga­
nized along dim ensions, and they indicate thal inclusion o f a stim u lu s requires sim ul­
taneous aclivation o f the stim ulus a mi the respective dim ension. W hereas extreme
stim uli d o not need to be presented on the sam e list as the targel stim uli to affect
jud ges’ perspective, it is also not sufficient that the extreme stim uli are sim ply highly
Individual analyses o f each largci drink sho * the same pattern.
Perspective
.161
accessible in memory. Rather, ihcy must have been thought about witli regard tu
the dim ension o f judgem ent to affect ju d g es’ perspective. II' the preceding question
d oes n oi tap the relevant dim ension, sim ply drawing attention to the extrem e stim uli
will not influence subsequent ratings, as a com parison o f the con su m p lion und caloric
content con d ition s illustrates.
T hese findings also suggest that incidental exposure to extreme stimuli (e.g. Herr.
1986; Kenrick and G ulierres, 1980) will only result in contrast effects on subsequent
judgem ents if it sp ontaneously evokes the relevant dim ension. For exam ple, Kenrick
and G ulierres (1980) observed that w atching a m ovie with strikingly attractive female
actors decreased m ales’ ratings o f the attractiveness o f potential dales. The current
findings suggest that such effects should be limited to judgem ents along dim ensions
that are considered sp on tan eou sly, and should not be obtained along less salient
dim ensions o f judgem ent.
From an applied point o f view, the current findings bear on context effects in
questionaires (<;/. H ippier, Schw arz, Sudm an (1987), Schwarz and Sudman (in press),
for reviews). W hile researchers are well aware that ratings o f a stim ulus m ay depend
on the nature o f the context stim uli presented in the same list ( c j . N oelle-N eum ann,
197U; Sudm an and Bradburn, 1983), the present results dem onstrate that the impact
o f extrem e stim uli on subsequent judgem ents is not limited to this w ell-know n c o n ­
dition. Rather, contrast effects may also emerge if extrem e stim uli are addressed
in pri'ceiling questions, provided that these questions tap the sam e underlying dim en­
sion. If the preceding questions lap an unrelated dim ension, how ever, they seem
unlikely to introduce system atic biases.
R E F ER E N C ES
Unmii, D. R 1195.1) 'Sumulus-similiirity and the anchoring of subjective scales’, AmerUun
Journal oj Psychology. 66 199 214.
Eiser, J R tin pressl. Social Judgment, Open University Press, London.
Herr, P M (1486). ■(.otisequeiices o f priming: Judgment and behavtur'. Journal uj Personality
unit Sin ml Psychology, 51. 1106-1115.
Hippier. H J„ Meier. I; and Schwarz, N. (I9HH), 'Erfahrungsbericlu zur Erprobung des
imeriikiiven Beiragungvund Insirukuunssysiems ” 11)15'“ , Z O M A-Naihricliieii, No. 23,
7V-V1.
Hippier. H. J . Schwarz, N. and Sudman. S. lEds) (1987) Social Injarmalioll Prwremmg
anil Survey Methodology, Springer Verhig, New York.
Kendrick, D. T. and Gulierres. S. E (tVSt)). 'Contrast clfects und judgments o f physical
iillracliveness' When beauty becomes a social problem*. Journal of Personality unit Social
P m huh,«.i, 38. 1.11-140.
Noelle-Neumann. E. (1970). ‘Wanted: Rules Tor wording questions'. Public Opinion Quarterly,
34 191-201.
Oslrom, T M. and L'pshaw. H. S, (1968). 'Psychological perspective and attitude change'.
In (Jreenwdld, A. C., Brock. T. C. und Ostrom, T. M. (Eds) Psychological Foundations
uj Attitudes, A endemic Press, New York.
Parducci, A. (196.1). 'Range-frcqueney compromise in judgment’, Psychological Monugriip/tt,
77 (2. whole No. 565).
Schwarz, N. and Sudman. S. (Eds) (111 press) Context Effeits in Soiial and Psychological
Research. Springer Veriag. New York.
Sudman, S and Bradburn, N M (1983). Asking Questions. Jossey-Bass, San Francisco.
Volkman, J (1951), 'Scales of judgment and their implications for social psychology', In.
Rohrer, J H. and Sherif, M. (E d sli’urio/ Psychology at the Crossroads, Harper. New York