Sie sind auf Seite 1von 31

Replication of Asch (1946)

Forming Impressions of Personality:


A Replication and Review of Asch's (1946) Evidence for a Primacy-of-Warmth Effect in Impression
Formation

Sanne Nauts1, Inge Huijsmans1, Oliver Langner2, Roos Vonk1 & Danil H. J. Wigboldus1

Behavioural Science Institute, Radboud University Nijmegen, The Netherlands


2

Friedrich Schiller University Jena, Germany

Sanne Nauts
Montessorilaan 3, 6525 HR Nijmegen, The Netherlands
Phone: + 31 24 3615682
Fax: +31 24 3612677
E-mail: s.nauts@psych.ru.nl

Replication of Asch (1946)

Social psychological laboratories have undergone considerable change since the publication of
Asch's "Forming Impressions of Personality" (1946), leading to the inevitable demise of punch cards
and slide carousels in favor of more technologically advanced experimental equipment. Still, the basic
methodology underlying present-day person perception research is strongly grounded in (and often
remarkably similar to) the methodology first introduced by Asch, now 67 years ago (Gilbert, 1998).
With over 2750 references, Asch's seminal research has been heralded for its paradigm-shifting
methodology, in which impression formation was studied in a controlled laboratory setting, yielding
high internal validity and experimental precision (Fiske, Cuddy, & Glick, 2007; Gilbert, 1998). But
Asch's legacy is by no means limited to the methodological realm: The ten experiments reported in
"Forming Impressions of Personality" have laid much of the groundwork for influential theories about
person perception (e.g., attribution theory; Jones & Davis, 1965; the continuum model of impression
formation; Fiske, Neuberg, Beattie, & Milberg, 1987).
Written long before the dawn of bite-size science (Bertamini & Munafo, 2012) and the advice
to "role-play grandma" to create a clear storyline (Bem, 1987, p. 27), "Forming Impressions of
Personality" is as interesting as it is multi-faceted. Asch (1946) offers a multitude of meandering
messages, frequently wedged between eloquent interludes, engaging observations, and elegant handcrafted figures. Although there may not be one single, unitary message to be taken from his work, the
one message that seems to have strongly resonated with present-day researchers concerns the primacyof-warmth-effect. The primacy-of-warmth effect (e.g., Fiske, Cuddy, & Glick, 2007; Wojciszke,
2005)1 entails that warmth has a primary role in impression formation, in the sense that warmth-related
judgments have a stronger influence on impressions (Wojciszke, Bazinska, & Jaworski, 1998), and
that warmth- related inferences are made more quickly than competence-related inferences (Ybarra,
Chan, & Park, 2001).
The goal of the present research is to critically examine the evidence that Asch's (1946)
research provides for such a primacy-of-warmth effect. Moreover, we will conduct a direct replication
of those studies of Asch's paper that are relevant to this particular effect. Replication attempts of

Replication of Asch (1946)

Asch's work abound (e.g., Ahlering & Parker, 1989; Anderson & Barrios, 1961; Babad, Kaplowitz &
Darley, 1990; Grace & Greenshields, 1960; Hendrick & Constantini, 1970; Kelley, 1950; Luchins,
1948; Luchins & Luchins, 1986; McCarthy & Skowronski, 2011; McKelvie, 1990; Mensh & Wishner,
1947; Pettijohn, Pettijohn & McDermott, 2009; Semin, 1989; Singh, Onglacto, Sriram & Tay, 1997;
Veness & Brierley, 1963; Wishner, 1960), but most are conceptual rather than direct replications,
many are incomplete, few relate to the primacy-of-warmth-effect, and finally, their results do not
necessarily concur with Asch's original findings. Thus, although "forming impressions of personality"
has been regarded as the point of inception for the primacy-of-warmth-effect (e.g., Abele &
Bruckmller, 2011; Abele & Wojciszke, 2007; Cuddy, Fiske & Glick, 2008; Fiske, Cuddy & Glick,
2007; Judd, James-Hawkins, Yzerbyt, & Kashima, 2005; Kervyn, Yzerbyt, & Judd, 2010; Richetin,
Durante, Mari, Perugini & Volpato, 2012; Vonk, 1994), it is unclear to what extent Aschs original
studies represent replicable evidence for such an effect. As one of the founding articles of the person
perception field, and perhaps even of social psychology more generally, we believe it is important to
establish its truth value and provide a replication attempt that provides researchers with sufficient data
to be able to interpret Asch's experiments, something we believe is not possible based on the data
reported in Asch's research due to incomplete reporting of the data, underpowered studies, and a lack
of systematic analysis of the data. Moreover, by conducting additional analyses based on more recent
insights in the person perception literature, we hope to shed more light on the processes that underlie
Asch's classic effects. Before going on to discuss the present replication attempt, we will first give a
short overview of Asch's main findings2.
Overview of Asch (1946)
In "Forming Impressions of Personality" ten studies were reported (total N = 8343) in which Asch
(1946) presented participants with lists of traits containing small manipulations. For example, in
Study I, participants were exposed either to a list of traits containing the word "warm" or a list
containing the word "cold", while all other traits were identical for both groups. After this, they had to
write down what their impression of the target person was (open-ended measure), pick from a list of

Replication of Asch (1946)

trait pairs which trait was most applicable to the target (trait-pair choice measure), and rank the traits
in the stimulus list in order of importance for determining their decisions (ranking measure). From this
study (frequently referred to as the warm-cold experiment) Asch inferred that participants rated
"warm" and "cold" as relatively central in determining their impressions. Moreover, participants more
often picked positive traits as fitting the target in the trait-pair choice measure in the warm condition
compared to the cold condition, except for several traits, such as strength, for which there was no
difference (though Asch could not explain why). Based on this experiment, Asch concluded that
participants' impressions are entirely transformed if "warm" is replaced by "cold". The nine
subsequent studies in "Forming Impressions of Personality" featured variations to this basic paradigm:
other traits were introduced, the order of traits was manipulated, participants were asked to give
synonyms for elements of the trait lists, or were asked to judge how the traits in a list relate to each
other. A short summary of all ten studies is provided in Table 1.
INSERT TABLE I HERE
Based on these ten experiments, Asch (1946) concluded that perceivers form coherent, unitary
impressions of others. To form such a unitary impression, perceivers attribute different meanings and
weights to traits, assigning a central role to some traits (these traits determine the meaning and
function of other traits) and a peripheral role to others (their meaning and function is determined by
the central traits). Which traits become central and which become peripheral is fully determined by the
context of the traits, so that the same trait can be central in one impression but peripheral in another.
As conclusion, Asch presents a Gestalt view on impression formation, emphasizing that traits interact
dynamically to shape each other's interpretation. For example, warmth may play a central role when
accompanied by traits like intelligent, skillful, industrious, determined, practical and cautious (as in
Asch's Study I), but this central role may disappear in the context of traits like obedient, weak,
shallow, unambitious, and vain (as in Asch's Study IV). This shows that impressions of personality are
extremely complex, and that the interpretation and meaning of a trait can completely change
depending on other traits being present (the change-of-meaning-explanation). Or, as Asch says it: "the

Replication of Asch (1946)

gaiety of an intelligent man is no more or less than the gaiety of a stupid man: it is different in quality"
(p.287). Thus, Asch suggests that the influence of central traits is not merely due to a general affective
shift (or Halo-effect): instead, he argues that the cognitive content of specific traits changes (e.g., the
gaiety of an intelligent man is not merely more or less positive than the gaiety of a stupid man: it
entails different qualities and behaviors). Though there is evidence that both a change-in-meaningeffect (e.g., Hamilton & Zanna, 1974; Zanna & Hamilton, 1977) and an affective shift (e.g., Kaplan,
1971; 1974) can influence impression formation, it is unclear which processes play a role in Asch's
studies. To find out more about this process, we will perform textual analyses of the open-ended
responses to investigate changes in affective content as well as meaning.

Interpretations of Asch's Work


Much like punch cards and slide carousels, the Gestalt-view of impression formation has slowly but
surely gone out of fashion (partly because there were more simple, parsimonious explanations for
Asch's data, e.g., Anderson, 1981; Rosenberg, Nelson & Vivekananthan 1968; Wishner, 1960).
Introductory textbook authors now generally put more emphasis on Asch's research (1946) as
providing evidence for a primacy-of-warmth-effect than on the Gestalt-model outlined in Asch's work.
For example, many textbooks refer solely to Asch's Study I while ignoring Study IV (that puts the
primacy-of-warmth-effect in a larger perspective4; e.g., Baron & Byrne, 2004; DeLamater & Meyers,
2010; Franzoi, 2009; Hogg & Vaughan, 2011; Kassin, Fein, & Markus, 2011; Pennington, 2000;
Taylor, Peplau, & Sears, 2006; Stainton-Rogers, 2011; Worchel, Cooper, Goethals, & Olson, 2000; for
a notable exception, see Hewstone, Stroebe, & Jonas, 2007). Although Asch acknowledges the
importance of warmth and coldness in impression formation, in his view any trait can be central as
well as peripheral. Thus, no trait is central by design, and even traits that have special importance,
such as warmth and coldness, can become peripheral in a different context because the meaning and
weight of a trait is never fixed or stable. As apparent in Table I, based on the results of Study I, Asch
concluded that warmth could have a central role in impression formation. Importantly, however, he

Replication of Asch (1946)

concluded based on the results of Study IV that warmth could become peripheral in other contexts.
Because Asch discounts a strong version of a primacy-of-warmth-effect in his theorizing, we believe
the only way to establish whether Asch has justifiably been regarded as the progenitor of the primacyof-warmth-effect is by going back to his initial observations. However, as we will describe in the next
section, Asch's initial observations have been sparsely replicated, and it is difficult to determine
whether they provide evidence for a primacy-of-warmth-effect.
Evidence for a Primacy-of-Warmth-Effect in Asch's Original Data (1946)
Asch's initial (1946) observations are based on three types of measures: open-ended questions in
which participants wrote down their general impression of the target, trait-pair choice measures in
which they chose which trait (out of a pair) is most applicable to the target (see Attachment I), and a
ranking measure in which they ranked the stimulus traits from most to least important in forming their
impression. In this section, we will review the extent to which these three measures provide evidence
for a primacy-of-warmth effect, starting with the open-ended questions.
Asch (1946) did not analyze the open-ended questions systematically, so it is difficult to
determine the extent to which they really provide evidence for a primacy-of-warmth effect.
Researchers replicating Asch's research did not include the open-ended questions (e.g., Mensh &
Wishner, 1947; Semin, 1989), did not provide a systematic analysis of this measure (Veness &
Brierley , 1963), or failed to replicate Asch's effects (Luchins, 1948; Gollin, 1954). Thus, so far the
results of the open-ended question data are inconclusive regarding the evidence for primacy-ofwarmth, and they have not been analyzed in ways to clarify whether the change-of-meaning-effect
Asch proposes, or whether other processes best explain the given responses.
The trait-pair choice results in Asch (1946) suggest that changing a trait in a list from positive
(e.g., "warm") to negative (e.g., "cold") makes the overall impression that participants have more
negative (the negative trait of the trait-pairs is chosen more frequently). These effects have been
supported in several direct replications of Study I (Mensh & Wishner, 1947; Veness & Brierley, 1963;
6

Replication of Asch (1946)

Semin ,1989), Study III (Semin, 1989) and Study IV (Mensh &Wishner, 1947). However, the extent to
which these data provide support for a primacy-of-warmth-effect is questionable, as replacing any
positive trait for any negative trait could potentially change the overall valence of the list. Thus, the
data for the trait-pair choice measure seem robust, but they may not provide the most stringent test of
the primacy-of-warmth hypothesis. Moreover, Asch could not explain why changing the central trait
in the list influenced only certain peripheral traits, although this was essential for his argumentation
favoring a change-of-meaning-explanation and discounting simple halo effects, which (in Asch's view)
should have influenced all traits. Later theorizing (e.g., Fiske, Cuddy, & Glick, 2007; Rosenberg,
Nelson, & Vivekananthan, 1968) may provide an explanation for Asch's findings: given that there are
two domains in social cognition (warmth and competence), changing a warmth-related trait should
change the interpretation of other warmth-related traits but not the interpretation of competencerelated traits. We would like to perform additional analyses to test whether, indeed, the influencing
effects of central traits are limited to traits of the same domain.
The ranking measures, in Asch's (1946) view, suggest that warmth is ranked as one of the
most important traits in Study I, but not in Study IV. For Studies I and IV, these results have been
replicated by Mensh and Wishner (1947). The original results of Asch's Study I are depicted in Table
II.
INSERT TABLE II HERE.
The ranking measures of Study I do not provide unequivocal evidence for a primacy-ofwarmth effect; in fact, in Study I, the number of participants indicating that warmth was the most
important trait guiding their impression is exactly as large as the number of participants indicating it
was the least important, a limitation Asch himself acknowledges, but that seems to have been
overlooked in many later references to his work. Moreover, the data are reported incompletely by
Asch so it is unclear which trait perceivers did rank first given that most participants did not rank
warmth as the most important trait in guiding their impression. In Study I, warmth is ranked in first
7

Replication of Asch (1946)

position by 6 of the 42 subjects, which is the exact number that would be expected based on chance
(given that there are seven options). This could mean one of two things: either every trait was ranked
in first position by an equal number of participants (six, to be exact), which seems at odds with Asch's
position that warmth was a central trait in this list. Alternatively, there may be one trait that was put in
first position more often than warmth (so by more than six participants), which seems to be even more
strongly at odds with Asch's position that warmth was the central trait in this list. Thus, for the
ranking-measure, Asch's data are difficult to interpret because they are reported incompletely and to
the extent that they are interpretable, they do not provide evidence for a primacy-of-warmth-effect.5.
In sum, the open-ended responses seem important for Asch's (1946) theorizing, but they were
not systematically investigated by himself, and his contemporaries failed to replicate his effects on this
measure (Gollin, 1954; Luchins, 1948). The trait-pair choice measure does provide robust results, but
it does not provide a stringent test of the primacy-of-warmth effect. The ranking results do not provide
very clear evidence for a primacy-of-warmth-effect, are only reported partially, have been sparsely
replicated, and are underpowered. For all three measures, it is unclear whether the impressions
perceivers formed based on the different stimulus lists differed significantly from each other, and what
the effect size of these differences is. Because of this, we think a well-powered replication of Asch's
research is in order. In the present research, we will conduct a replication of Asch's Studies I, III and
IV (these are the studies that are most relevant for the primacy-of-warmth effect), aiming to provide
data that, unlike Asch's original work, are sufficiently powered, employ clear coding of the openended questions, provide complete accounts of the data, and provide statistical analysis of the results.
In so doing, we aim to shed light on the replicable evidence that Asch's results provide for the primacy
of warmth in impression formation.
Moreover, we would like to add additional analyses of the trait-pair choice measure and
several additional analyses of the open-ended responses. Asch's ranking measure may be suboptimal,
as it presupposes that perceivers have full access to the factors that determine their impressions (which
is unlikely; Nisbett & Wilson, 1977). We chose to focus our efforts on extra analyses for the trait-pair
8

Replication of Asch (1946)

choice measure and open ended-measure because these measures do not rely on introspection to the
same extent. In our view, the descriptions of targets in particular may contain lots of new information
on the primacy-of-warmth-effect. Asch did not analyze the open ended-responses, and for 67 years,
researchers have focused on easy-to-analyze, top-down measures in impression formation, largely
ignoring the rich information contained in open ended-descriptions. We think it is worthwhile to open
this treasure chest and explore its contents (even though these analyses will be exploratory in nature).
Proposed Method
For the replication, we selected the experiments that have the strongest bearing on the primacy-ofwarmth-effect, namely Studies I, III, and IV of the original publication. Here, we outline the details of
the studies that we plan to conduct. Further, we give an overview of the known differences between
Asch's original work and the planned studies, and indicate how these differences could potentially
influence the results.

Study Outline
In our replication attempt, we will faithfully replicate Asch's (1946) method by exposing
participants to the stimulus lists (as depicted in Table I), and asking them to complete the open-ended
measure, trait-pair choice measure, and ranking measure. Asch only presented a frequency table of
participants' responses, and did not statistically analyze his results (Mensh & Wishner, 1947, do
provide some statistical analysis of their results, but these are very limited, e.g., they do not include
effect sizes). As an addition to his study, we will perform statistical analyses on the ranking measures,
to test if central traits are indeed ranked as more important than other traits, and to establish the effect
size. Moreover, we will examine if any trait is selected as the most important trait determining
perceiver's impressions significantly more often than other traits, something that is not possible based
on Asch's presentation of his data.

Replication of Asch (1946)

In addition to the measures employed by Asch (1946), we will analyze the open-ended responses
based on the literature by Gollin (1954). Gollin distinguishes three types of impressions: unitary or
unified impressions (in which all the trait information is incorporated in a unified manner), simplified
impressions (impression in which part of the trait information is clearly ignored), and aggregated
impressions (impressions in which no unity is achieved at all). We will develop clear coding schemas
based on the research by Gollin, and will have the responses rated by two independent raters.
Moreover, we will add textual analyses of the open ended-responses to further investigate which traits
are central, and how changing one trait in the list can shift the interpretation and/or valence of other
traits in the list.
Participants and Design
This study will employ a between-subject design with seven conditions, as depicted in Table 3. We
will recruit 1050 participants (150 per condition) to participate in our study through MTurk.
INSERT TABLE III HERE
Procedure
Participants will be randomly assigned to one of the seven stimulus lists (as depicted in Table 3) and
will be asked to form an impression of the target described in the list. Next, they will first be asked to
write down what their impression of the target is (open-ended question). After this, they will be
instructed to select the term in each trait pair (as depicted in Attachment I) that is most applicable to
the person (the trait-pair choice measure). Finally, they will once again be exposed to the stimulus list,
and will be asked to rank-order the traits from the most important trait in forming their impression to
the least important trait.
Analysis Plan
Because the hypotheses in Aschs paper (1946) are not stated explicitly, we tried to formulate
hypotheses that fit with Asch's theorizing and are quantifiable and testable. Because there is a
subjective component to our operationalization of Asch's research questions, we formulated multiple
10

Replication of Asch (1946)

hypotheses that cover a range of tests, from very lenient to very stringent. In this section, we will
describe the hypotheses for the three different measures: the ranking measures, the trait pair-choice
measure, and the open-ended measure. Because the study will be run on MTurk, we will remove any
participants who are not native speakers. Moreover, we will remove all participants who fail an
instructional manipulation check (Oppenheimer, Meyvis & Davidenko, 2009).
Ranking Measures
For the ranking measure, we propose the following hypotheses (ordered from most lenient to most
stringent):
Hypothesis 1: average rankings. The first hypothesis concerns the question whether the
average rank administered to a trait (e.g., warmth in Condition 1A) differs from the average rank given
to other traits in the same list. It contains the following sub hypotheses:
1a) Warmth and coldness are expected to be central in Conditions 1A and 1B. Thus, on
average, warmth and coldness should have a higher average rank than any of the other traits in
that condition.
1b) Warmth and coldness are expected to be peripheral in Conditions 2A, 2B, and 2C. Thus,
there should be at least one trait with a higher average rank than warmth and coldness in that
condition.
1c) Polite and blunt are expected to be peripheral in Conditions 3A and 3B. Thus, there should
be at least one trait with a higher average rank than politeness and bluntness in that condition.
Hypotheses 1a, 1b and 1c are quite lenient: in this scenario, warmth could be the most highly ranked
trait on average, even though warmth may not be the most important trait for many (or even any) of
the participants. We will analyze this using two different statistical techniques. First of all, we will use
a Wilcoxon signed rank analysis to investigate if the average rank for the focal trait (warmth and
coldness in Conditions 1 and 2, politeness and bluntness in Condition 3) differs from the average rank
of any of the other traits in a list. Moreover, we will perform an exploded Logit analysis: an analysis

11

Replication of Asch (1946)

that qualitatively yields the same results as the Wilcoxon's test, but will give us an opportunity to look
at specific ranks in Hypothesis 2 (Allison & Christakis, 1994).
Hypothesis 2: average rankings for top choices. The second hypothesis concerns the
question whether the average rank administered to a trait differs from the average rank given to other
traits in the same list. Unlike Hypothesis 1, however, we only take the top ranks into account (Rank 1,
2 and 3). If warmth is indeed primary, this should not just be visible in the average rank over all ranks:
it should hold even if we ignore all rankings lower than 3. Thus, this hypothesis rules out the
possibility that warmth and coldness are regarded as central but are not well presented in the top ranks
(Rank 1, 2 and 3). Hypothesis 2 entails the following sub hypotheses:
2a) Warmth and coldness are expected to be central in Conditions 1A and 1B. Thus, on
average, warmth and coldness should have higher average ranks than any of the other traits in
that condition, even if we leave all ranks lower than 3 (4 and up) out of the analysis.
2b) Warmth and coldness are expected to be peripheral in Conditions 2A, 2B, and 2C. Thus,
there should be at least 1 trait that has a higher average rank than any of the other traits in that
condition, even if we leave all ranks lower than 3 (4 and up) out of the analysis.
2c) Politeness and bluntness are expected to be peripheral in Conditions 3A and 3B. Thus,
there should be at least 1 trait that has a higher average rank than any of the other traits in that
condition, even if we leave all ranks lower than 3 (4 and up) out of the analysis.
For these analysis, we will employ an Exploded Logit model. However, we will assign ties to all ranks
higher than 3, so that only the top ranks are analyzed (based on the recommendations of Allison and
Christakis, 1994).
Trait-Pair Choice Measures
For the trait-pair choice measures, the hypotheses are as follows:
3a) Perceivers in Condition 1A should have more positive inferences of traits than participants
in Condition 1B. Thus, there should be significant difference between conditions in the
number of times each trait is selected.

12

Replication of Asch (1946)

3b) Perceivers in Condition 3A should not have more positive inferences of the target than
participants in Condition 3B. Thus, there should be no significant difference between the
conditions in the amount of times each trait is picked. If there is a significant difference, the
size of the effect should be significantly smaller than the difference between Conditions 1A
and 1B.
3c) Only traits that are related to the same domain as the trait that is changed should be
influenced. Thus, if "warm" is changed to "cold" in conditions 1A and 1B, only traits that are
warmth-related should be influenced.
Hypothesis 3a and 3b will be tested with an ANOVA. To test Hypothesis 3c, we will ask two
independent raters to indicate the extent to which the traits of the trait-pair choice lists are related to
warmth and competence. We expect that peripheral traits that are warm-related will differ between
conditions, but competence-related traits will not.
Open-Ended Measures
For the open-ended measures, the hypotheses are as follows:
4a) All perceivers form unified impressions of personality in all conditions.
4b) Warmth and coldness will be central traits in Conditions 1A and 1B, but not in Conditions
2A and 2B. Politeness and bluntness will not be central traits in Conditions 3A and 3C.
4c) Perceivers attribute different meanings to "warm" in Conditions 1A compared to
Conditions 2A and 2B. Perceivers attribute different meanings to "cold" in Condition 1B
compared to 2C, providing evidence for a change-of-meaning-effect.
To test Hypothesis 4a, we will employ a coding scheme based on Gollin (1954), as described above.
Because analysis of these qualitative data is time-consuming, we will randomly select a third of the
responses for every condition, yielding 50 responses per condition (350 responses in total). Two
independent raters will rate 10 percent of these responses to establish inter-rater reliability.
To test Hypothesis 4b, we will perform textual analyses of the data. The traits that are used in
participant's descriptions are extracted (negations such as "not warm" will be coded separately). We

13

Replication of Asch (1946)

expect that central traits will be mentioned by participants more often (without negation) than
peripheral traits and thus should occur more frequently in the descriptions. Second, all traits will be
rated on valence and their relevance to competence and warmth (the extent to which they, cognitively,
refer to the warmth-domain or the competence-domain). We expect that changing a central trait will
lead to an affective shift (e.g., usage of more positive terms when "warmth" is the central trait) as well
as a cognitive shift (e.g., usage of more warmth-related terms if "warmth" is the central trait). These
analyses are new and exploratory in nature, but have the potential to shed new light on the centrality of
traits in perceivers' impressions, as well as on the veracity of a change-of-meaning-explanation for
these effects.
To test Hypothesis 4c, two independent raters will receive a short list with synonyms of
"warm". Next, they will be asked to rate which synonym is most applicable to the description. We will
employ a chi square-test to test if different options are selected for Condition 1A compared to 2A and
2B, and for Condition 1B compared to 2C.
Power Analysis.
Ranking measure. It is difficult to get a good estimate of the effect sizes in Asch's (1946)
studies. Although it is possible to get an idea of Asch's effect sizes based on the results reported in the
paper, most of his studies were likely underpowered (with only 20 to 26 participants per cell for
Studies III and IV). If we calculate effect sizes based on Study I, which may be slightly less
underpowered for the ranking measure than Studies III and IV, effect sizes seem large (around w =
.50). Because of the small sample sizes and potentially smaller effects for Studies III and IV, we
decided to use a slightly more conservative estimate of the effect size as a medium effect instead of a
large effect (w = .30 instead of w = .50). With this medium effects size and an alpha level of .80, this
would mean that we need approximately 151 participants per group (Cohen, 1992). Power analysis
with G*Power (Faul, Erdfelder, Lang & Buchner, 2007) yields a highly similar number of participants
(N = 152). Thus, we expect to need approximately 150 participants per condition, or 1050 participants

14

Replication of Asch (1946)

in total. Because the experiment takes just a couple of minutes, and the analysis of the quantitative
data is not labor-intensive, this number seems feasible.
For the Wilcoxon signed rank-analysis, we need 64 participants per cell to pick up on a
medium effect size, so 150 participants per cell will suffice for this analysis, too. For the exploded
logit analysis, we need approximately 100 participants in total-again, 150 participants per cell will
suffice for this analysis.
Trait-pair choice measure. Because Hypothesis 3b is, in essence, a null hypothesis, we
would like to adjust the alpha level for this test to .95, in occurrence with Cohen's (1990)
recommendations for testing null effects. Again, the effect size is hard to determine;if, to be
conservative, we take a low to medium effect size (e.g., Cohen's d = 0.40), power analysis suggests
that we would need 136 participants per condition Thus, the sample size we need for the ranking
measures (150 participants per group) would suffice for the trait-pair choice analysis.
Open ended -measure. For Hypothesis 4a, Asch (1946) stated that every participant will
form a unified impression, and that every participant in his studies did so. Because this universal claim
will likely not hold in our data we would like to give a more qualitative view of the data by merely
presenting descriptives of the data.
To test Hypothesis 4b, we will perform simple t-tests to estimate the difference in occurrence
of traits within the conditions, as well as the difference in valence and domain of the terms mentioned
in the different conditions. We will include all participants in these textual analyses.
To test Hypothesis 4c, we will perform a Chi square test. With 50 participants per cell, the
effect size for the t-test would need to be on the higher end (Cohen's d = 0.70) for us to have enough
power with a sample size of 50 per cell to pick it up. The effect size for the Chi square test would need
to be large (w = .50). Although it would be preferable to have more power for these tests, the costs of
rating the responses of 350 respondents are already quite large, and we believe rating a large
percentage of responses will not be viable from a practical point of view. Moreover, because Asch
describes the effects of the open-ended responses as being very striking, and describes the differences

15

Replication of Asch (1946)

between conditions as large, we believe it is reasonable to predict moderately large effect sizes for
these particular hypotheses.
Known Differences
Although we will try to replicate Asch (1946) as diligently and precisely as possible, there may
nevertheless be differences between our study and the original work. Some of these differences are
advantageous, for example:
1) Compared to the original studies, our research has more statistical power.
2) Unlike the original studies, we provide statistical analysis of the results.
3) Unlike the original studies, we can compare the conditions of Studies I, III and IV against
each other.
4) Unlike the original studies, we systematically analyze open-ended responses.
Other differences may be disadvantageous, for example:
5) Unlike the original studies, the present replication attempt will be carried out online. We
think it is important to carry out the experiments in English, as translating the materials could lead to
important differences in the strength and interpretation of the trait words between Asch's original study
and ours. Unfortunately, as we need many participants, and do not have access to sufficiently large
samples of English speakers in our lab, we would like to conduct the study online (through MTurk). It
is conceivable that Asch's participants (students) were a more homogenous or motivated group than
the participants in our study. Moreover, participants in Asch's study did not read the stimulus lists
themselves: they received instructions verbally and answered on a sheet of paper. It is possible that a
computer-based version of the experiment, in which the experimenter is not present while the
participant completes the experiment, will lead to slightly different responses (e.g., less socially
desirable responses). However, we have no specific reason to believe that a different mode of
administration will profoundly influence the role warmth plays in impression formation.
16

Replication of Asch (1946)

6) Unlike in the original studies, it is 2013. It is possible that words had slightly different
connotations in 1946 than they have in 2013, or that societal changes have brought about changes in
the way people form impressions of others, or how they report those impressions. For example, there
has been a considerable change in racial stereotypes, or people's willingness to report prejudiced views
of racial groups, since 1933 ( Devine & Elliot, 1995; Karlins, Coffman & Walters, 1969), suggesting
that these specific impression formation processes have changed over time. Though it is certainly
possible that the basic impression formation processes described by Asch in 1946 have also changed,
we believe that these processes may be less strongly influenced by social desirability concerns brought
forward by societal changes (compared to other impression formation processes, such as those related
to racial groups). Moreover, we believe that changes in language use may not be particularly
problematic, given that the specific words used in Asch' stimulus lists remain relatively frequent in the
English language.
Coda
Our motivation to replicate Asch's (1946) research might be taken by some as an attempt to discount
his work. In fact, nothing could be further removed from the truth. We greatly admire Asch's work,
which changed the face of person perception research and introduced an entirely new way of studying
first impressions. Exploring his original observations with present-day (statistical) standards in mind,
in our view, does not depreciate the brilliance of Asch's work, which lies in its Gestalt.

17

Replication of Asch (1946)

References
Abele, A. E., & Wojciszke, B. (2007). Agency and communion from the perspective of self
versus others. Journal of Personality and Social Psychology, 93, 751-763.
Abele, A. E., & Bruckmller, S. (2011). The bigger one of the Big Two? Preferential
processing of communal information. Journal of Experimental Social Psychology, 47, 935948.
Ahlering, R.F., & Parker, L.D. (1989). Need for cognition as moderator of the primacy effect. Journal
of Research in Personality, 23, 313-317.
Allison, P.D., & Christakis, N.A. (1994). Logit models for sets of ranked items. Sociological
Methodology, 24, 199-228.
Anderson, N. H. (1981). Foundations of information integration theory. New York, NY: Academic
Press.
Anderson, N.H., & Barrios, A.A. (1961). Primacy effects in personality impression formation. Journal
of Abnormal and Social Psychology, 63, 346-360.
Asch, S.E. (1946). Forming impressions of personality. Journal of Abnormal and Social
Psychology, 41, 258-290.
Babad, E., Kaplowitz, H., & Darley, J. (1999). A classic revisited: Students immediate and
delayed evaluations of a warm/cold instructor. Social Psychology of Education, 3, 81-102.
Baron, R.A., & Byrne, D.E. (2004). Social Psychology. Boston, MA: Allyn Bacon.
Bem, D. J. (1987). Writing the empirical journal article. In M. P. Zanna & J. M. Darley (Eds.), The
complete academic: A practical guide for the beginning social scientist (pp. 171-201). New
York, NY: Random House.
Bertamini, M., & Munafo, M.R. (2012). Bite-size science and its undesired side-effects.
Perspectives on Psychological Science, 7, 67-71.
Cohen, J. (1990). Things I have learned (so far). American Psychologist, 45, 1304-1312.
Cohen, J. (1992). A power primer. Quantitative Methods in Psychology, 112, 155-159.

18

Replication of Asch (1946)

Cuddy, A. J., Fiske, S. T., & Glick, P. (2008). Warmth and competence as universal
dimensions of social perception: The stereotype content model and the BIAS map. Advances
in Experimental Social Psychology, 40, 61-149.
DeLamater, J.D., & Myers, D.J. (2010). Social Psychology. Belmont, CA: Wadsworth Publishing.
Devine, P.G., & Elliot, A.J. (1995). Are racial stereotypes really fading: The Princeton trilogy
revisited. Personality and Social Psychology Bulletin, 11, 1139-1150.
Fiske, S.T., Neuberg, S.L., Beattie, A.E., & Milberg, S.J. (1987). Category-based and
attribute-based reactions to others: Some informational conditions of stereotyping and
individuating processes. Journal of Experimental Social Psychology, 23, 399-427.
Fiske, S.T., Cuddy, A.J.C., & Glick, P. (2007). Universal dimensions of social cognition: Warmth and
competence. Trends in Cognitive Sciences, 11, 77-83.
Faul, F., Erdfelder, E., Lang, A.G., & Buchner, A. (2007). G*Power 3: A flexible statistical power
analysis program for the social, behavioral, and biomedical sciences. Behavior Research
Methods, 39, 175-191.
Franzoi, S.L. (2009). Social Psychology (5th ed.). New York, NY: McGraw-Hill.
Gilbert, D.T. (1998). Ordinary Personology. In D.T.Gilbert, S.T.Fiske, & L. Gardner (Eds.), The
Handbook of Social Psychology (4th ed., Vol.2, pp. 89-150). New York, NY: McGraw-Hill.
Gollin, E.S. (1954). Forming impressions of personality. Journal of Personality, 23, 65-76.
Grace, H.A., & Geenshields, C.M. (1960). Effect of closure on formation of impressions.
Psychological reports, 6, 94.

Hamilton, D. L., & Zanna, M. P. (1974). Context effects in impression formation: Changes in
connotative meaning. Journal of Personality and Social Psychology, 29, 649-654.
Hendrick, C., & Constantini, A.V. (1970). Effects of varying trait inconsistency and response
requirements on the primacy effect in impression formation. Journal of Personality and Social
Psychology, 18, 158-164.
Hewstone, M., Stroebe, W., & Jonas, K. (2007). Introduction to Social psychology: A European
19

Replication of Asch (1946)

Perspective (4th ed.). Hoboken, NJ: John Wiley & Sons.


Hogg, M., & Vaughan, G. (2011). Social Psychology (6th ed.). Upper Saddle River, NJ: Prentice Hall.
Jones, E. E., & Davis, K. E. (1965). From acts to dispositions: The attribution process in person
perception. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 2, pp.
219-266). New York, NY: Academic Press.
Judd, C. M., James-Hawkins, L., Yzerbyt, V., & Kashima, Y. (2005). Fundamental
dimensions of social judgment: understanding the relations between judgments of competence
and warmth. Journal of Personality and Social Psychology, 89, 899-913.

Kaplan, M. F. (1971). Context effects in impression formation: The weighted average versus
the meaning-change formulation. Journal of Personality and Social Psychology, 19,
92-99.
Kaplan, M. F. (1974). Context-induced shifts in personality trait evaluation: A comment on
the evaluative halo effect and meaning change interpretations. Psychological Bulletin,
81, 891-895.
Karlins, M., Coffman, T.L., & Walters, G. (1969). On the fading of social stereotypes: Studies in three
generations of college students. Journal of Personality and Social Psychology, 13, 1-16.
Kassin, S., Fein, S., & Markus, H. R. (2011). Social psychology (8th ed.). Stamford, CT: Cengage
Learning.
Kelley, H.H. (1950). The warm-cold variable in impressions of persons. Journal of Personality, 18,
431-439.
Kervyn, N., Yzerbyt, V., & Judd, C. M. (2010). Compensation between warmth and
competence: Antecedents and consequences of a negative relation between the two
fundamental dimensions of social perception. European Review of Social Psychology,
21, 155-187.
Luchins, A.S. (1948). Forming impressions of personality: A critique. Journal of Abnormal and Social
Psychology, 43, 318-325.
20

Replication of Asch (1946)

Luchins, A.S., & Luchins, E.H. (1986). Primacy and recency effects with descriptions of moral and
immoral behavior. The Journal of General Psychology, 113, 159-177.
Pennington, D.C. (2000). Social Cognition. (1st ed.). London, Great Britain: Routledge.
McCarthy, R.J., & Skowronski, J.J. (2011). You're getting warmer: Level of construal affects the
impact of central traits on impression formation. Journal of Experimental Social Psychology,
47, 1304-1307.
McKelvie, S.J. (1990). The Asch primacy effect: Robust but not infallible. Journal of Social
Behavior & Personality, 5, 135-150.
Mensh, I.N., & Wishner, J. (1947). Asch on "Forming impressions of personality": Further evidence.
Journal of Personality, 16, 188-191.
Nisbett, R. E., & Wilson, T. D. (1977). Telling more than we can know: Verbal reports on
mental processes. Psychological Review, 84, 231-259.
Oppenheimer, D.M., Meyvis, T., & Davidenko, N. (2009). Instructional manipulation checks:
Detecting satisficing to increase statistical power. Journal of Experimental Social Psychology,
45, 867-872.
Peeters, G., & Czapinski, J. (1990). Positive-negative asymmetry in evaluations: The
distinction between affective and informational negativity effects. European Review of
Social Psychology, 1, 33-60.
Pettijohn, T.F. II, Pettijon, T.F., McDermott, L.A. (2009). Active learning exercises for teaching
classic research on impression formation in social psychology courses.
The Open Education Journal, 2, 78-81.
Richetin, J., Durante, F., Mari, S., Perugini, M., & Volpato, C. (2012). Primacy of warmth versus
competence: A motivated bias? The Journal of Social Psychology, 152, 417-435.
Rosenberg, S., Nelson, C., & Vivekananthan, P. S. (1968). A multidimensional approach to
the structure of personality impressions. Journal of Personality and Social Psychology, 9, 283294.

21

Replication of Asch (1946)

Semin, G.R. (1989). The contribution of linguistic factors to attribute inferences and semantic
similarity judgements. European Journal of Social Psychology, 19, 85-100.
Singh, R., Onglacto, M.L.U., Sriram, N., & Tay, A.B.G. (1997) The warm-cold variable in impression
formation: Evidence for the positive-negative asymmetry. British Journal of Social
Psychology, 36, 457-477.
Stainton-Rogers, W. (2011). Social Psychology (2nd ed.). Philadelphia, PA: Open University Press.
Taylor, S. E., Peplau, L. A., & Sears, D. O. (2006). Social psychology (12th ed.). Upper Saddle River,
NJ: Prentice-Hall.
Todorov, A., Said, C.P., Engell, A.D., & Oosterhof, N.N. (2008). Understanding evaluation of faces
on social dimensions. Trends in Cognitive Sciences, 12, 455-460.
Ybarra, O., Chan, E., & Park, D. (2001). Young and old adults' concerns about morality
and competence. Motivation and Emotion, 25, 85-100.
Veness, T., & Brierley, D.W. (1963). Forming impressions of personality: Two experiments. British
Journal of Social and Clinical Psychology, 2, 11-19.
Vonk, R. (1994). Trait inferences, impression formation, and person memory: Strategies in
processing inconsistent information about persons. European Review of Social Psychology, 5,
111-149.
Wishner, J. (1960). Reanalysis of "impressions of personality". Psychological Review, 67, 96-112.
Wojciszke, B., Bazinska, R., & Jaworski, M. (1998a). On the dominance of moral categories
in impression formation. Personality and Social Psychology Bulletin, 24, 1245-1257.
Wojciszke, B., Dowhyluk, M., & Jaworski, M. (1998). Moral competence-related traits: How
do they differ? Polish Psychological Bulletin, 29, 283-294.
Wojciszke, B. (2005). Morality and competence in person- and self-perception. European Review of
Social Psychology, 16, 155-188.
Worchel, S., Cooper, J., Goethals, G., & Olson, J. (2000). Social psychology. Belmont, CA:
Wadsworth.

22

Replication of Asch (1946)

Zanna, M. P., & Hamilton, D. L. (1977). Further evidence for meaning change in impression
formation. Journal of Experimental Social Psychology, 13, 224-238.

23

Replication of Asch (1946)

Footnotes
1

In the present research, in line with the recommendations by Fiske, Cuddy and Glick (2007), warmth

is used as an omnibus term that includes dimensions such as other-profitability (Peeters & Czapinski,
1990), morality (Wojciszke, 2005), and trustworthiness (Todorov, Said, Engell, & Oosterhof, 2008).
2

We do not claim to provide a complete and unbiased view of Asch's (1946) complex work; in trying

to provide a readable overview, we have omitted countless details that are only accessible in the
original work. For a truly complete appreciation of Asch, please read Asch.
3

A well-informed reader may notice that Asch writes in his introduction that he tested over a 1000

participants, but the results of 834 (not 1000) are reported in Asch's method-sections.
4

Although some authors additionally refer to Study VI or VII about primacy-effects.

We would like to note that our goal is not to question the veracity of the primacy-of-warmth effect

per se (there is a lot of research suggesting that warmth plays a central role in impression formation;
for a review, see Fiske, Cuddy, & Glick, 2007; Wojciszke, 2005). Instead, the aim is to investigate to
what extent Asch's initial observations can be taken as evidence for a primacy-of-warmth-effect, and
to what extent they suggest that there may be limitations or boundary conditions to such an effect (as
Asch hypothesized). We believe this is not merely important from a historical perspective: person
perception researchers have been strongly inspired by Asch's research, and we hope that a more
complete, interpretable overview of his results, combined with some additional analyses that explain
more about the processes underlying the effects can inform new research.

24

Replication of Asch (1946)

Tables
Table 1.
Short Overview of the Methods of the Ten Experiments Reported in Asch (1946), Including the
Conclusions that Asch Draws Based on the Experiments.
Study

Included?

Stimulus lists

Method and Conclusions

yes

group A (N = 90):
intelligent, skillful,
industrious, warm,
determined, practical,
cautious

Participants read the stimulus lists, and wrote down the impression they had of the target
(open-ended measure). Next, they had to indicate for a list of traits (see Attachment I)
which was most in accordance with their impression (trait-pair choice measure). Next,
they had to rank order which traits of the stimulus list were most important in
determining their impressions (ranking measure).

group B (N = 76):
intelligent, skillful,
industrious, cold,
determined, practical,
cautious

Summary of Asch's main conclusions:


a) Based on the open-ended measure, Asch concluded that perceivers formed a unitary
impression, that changes radically when "warm" is changed to "cold".
b) Based on the trait-pair choice measure, Asch concluded that impressions reversed for
certain qualities (e.g., generous, happy), but not others (e.g., honest, strong) when
"warm" is changed to "cold".
c) Based on the ranking measures, Asch concludes that some characteristics (such as
warmth) were more central than others.

II

no

(N = 56): intelligent,
skillful, industrious,
determined, practical,
cautious

Participants read the stimulus list and completed the open-ended measure and ranking
measure.
Summary of Asch's main conclusions:
a) Based on the ranking measure, Asch concluded that omitting warm/cold from the
stimulus list does not function entirely as an omission. Instead, approximately half of the
subjects spontaneously forms a warm impression, the other half a cold one.

III

yes

group A (N = 20):
intelligent, skillful,
industrious, polite,
determined, practical,
cautious
group B (N = 26):
intelligent, skillful,
industrious, blunt,
determined, practical,
cautious

IV

yes

group A (N = 23):
obedient, weak, shallow,

Participants read the stimulus list and completed the open-ended measure and ranking
measure.
Summary of Asch's main conclusions:
a) Based on the ranking measure, Asch concluded that changing a peripheral quality,
such as "polite" to "blunt", influences participant's impression less strongly than
changing a central quality, such as "warm" to "cold"

Participants read the stimulus list and completed the open-ended measure and ranking
measure.

25

Replication of Asch (1946)

warm, unambitious, vain

Summary of Asch's main conclusions:

group B (N = 21): vain,


shrewd, unscrupulous,
warm, shallow, envious

a) Based on the open-ended measure, Asch concluded that "warm" is interpreted


differently in this context than in the context of Study I.

group C (N = 20):
intelligent, skillful, sincere,
cold, conscientious,
helpful, modest
V

VI

VII

no

no

no

b) Based on the ranking measure, Asch concluded that warm/cold is much less important
in determining impression than it is in Study 1: it has become peripheral and is no longer
central. Thus, a trait can be central or peripheral depending on the context of the
impression.

group A (N = 38): kind,


wise, honest, calm, strong

Participants read the stimulus list and completed the open-ended measure and give
synonyms for calm and strong (synonym measure).

group B (N = 41): cruel,


shrewd, unscrupulous,
calm, strong

Summary of Asch's main conclusions:

group A (N = 34):
intelligent, industrious,
impulsive, critical,
stubborn, envious

Participants read the stimulus list and completed the open-ended measure, trait-pair
choice measure and ranking measure.

group B (N = 24):
envious stubborn, critical,
impulsive, industrious,
intelligent

a) Based on the trait-pair choice measure, Asch concluded that more positive qualities
are chosen for group A compared to group B, suggesting that traits early in the list have a
stronger influence on impressions (evidence for a primacy-effect).

group A (N = 27):
intelligent, skillful,
industrious, determined,
practical, cautious, evasive

Participants read the stimulus list and completed the open-ended measure, trait-pair
choice measure and ranking measure.

group B (N = 30): evasive


skillful, industrious,
determined, practical,
cautious, intelligent

a) Based on the trait-pair choice measure, Asch concludes that more positive qualities are
chosen for group A compared to group B, suggesting that traits early in the list have a
stronger influence on impressions (evidence for a primacy-effect).

a) Based on the synonym measure, Asch concluded that people attribute a different
(more positive) meaning to calm and strong in group A than group B. Thus, the
interpretation of a trait depends on the context.

Summary of Asch's main conclusions:

b) Based on the ranking measure, Asch concluded that traits early in the series are ranked
as more important in determining participant's impressions (also evidence for a primacyeffect).

Summary of Asch's main conclusions:

b) Based on the ranking measure, Asch concluded that traits early in the series are ranked
as more important in determining participant's impressions (also evidence for a primacyeffect).
This is a replication of Study VI with similar (though slightly weaker) effects.

VIII

no

Group A (N=24):
intelligent, industrious,
impulsive, critical,
stubborn, envious

Participants either read the stimulus list as if it described one person (group A) or as if it
described two people (group B) and had to complete the open-ended and trait-pair choice
measure.
Participants in group B later heard that the lists described one single person instead of

26

Replication of Asch (1946)

IX

no

Group B (N = 52):

two, and were asked to again form an impression (second open-ended measure).

person1:intelligent,
industrious, impulsive

Summary of Asch's main conclusions:

person 2: critical, stubborn,


envious

a) Based on the open-ended questions, Asch concluded that perceivers have trouble
forming a unified impression in group B, but not in group A, suggesting that the
interpretation given to traits depended on the context of other traits, With a sudden
change in context, participants' initial interpretations were no longer fitting.

IX part1: intelligentskillful-warm

Participants read the stimulus list and completed the open-ended measure and trait-pair
choice measure.

IX part2: group A (N =
22): warm

Summary of Asch's main conclusions:

group B (N = 33): cold

no

Participants saw all of the


following four sets:
Set 1: quick, skillful,
helpful
Set 2: quick , clumsy,
helpful
Set 3: slow, skillful,
helpful

a) Based on the trait-pair choice measure, Asch concluded that warmth is central, but that
it has more weight than in Study I, suggesting that even if warmth may have been central
in Study I, it was influenced by the context of other, peripheral traits. Thus, the weight
and meaning of central traits depends on the context.
Participants read all four stimulus list and completed the open-ended measure and
indicated which traits they thought resemble each other (e.g., does "quick" from Set 1
resemble "quick" from Set 2, or "slow" from Set 3?; the similarity measure).
Summary of Asch's main conclusions:
a)Based on the similarity measure, Asch concluded that different traits are sometimes
judged as more similar (e.g., quick in Set 1 and slow in Set 3) than similar traits (e.g.,
quick in Set 1 compared to quick in Set 2). This suggests that the meaning and
interpretation of traits depends on the context.

Set 4: slow, clumsy,


helpful
He also uses a similar
methodology with
different stimulus lists but
similar results.

Note: Studies that will be included in our replication attempt are marked with "yes" under "included?".
"Stimulus lists" features the lists of stimuli that participants were exposed to.

27

Replication of Asch (1946)

Table II
Original Data of Experiment I (Asch, 1946); Distribution of Ranking of the Warm-Cold Qualities in
Order of Their Importance in Forming Impressions.
Warm

Cold

Rank

Percentage

Percentage

14

12

27

15

35

21

10

10

10

14

13

33

Total

42

100

41

100

28

Replication of Asch (1946)

Table III
Conditions Included in our Replication and the Stimulus List Participants are Exposed To.
Condition in our Replication

Condition in Asch (1946)

Stimulus List

Condition 1 A

Study I, Group A

intelligent, skillful,
industrious, warm,
determined, practical, cautious

Condition 1 B

Study I, Group B

intelligent, skillful,
industrious, cold, determined,
practical, cautious

Condition 2 A

Study IV, Group A

obedient, weak, shallow,


warm, unambitious, vain

Condition 2 B

Study IV, Group B

vain, shrewd, unscrupulous,


warm, shallow, envious

Condition 2 C

Study IV, Group C

intelligent, skillful, sincere,


cold, conscientious, helpful,
modest

Condition 3 A

Study III, Group A

intelligent, skillful,
industrious, polite,
determined, practical, cautious

Condition 3 B

Study III, Group B

intelligent, skillful,
industrious, blunt,
determined, practical, cautious

29

Replication of Asch (1946)

Attachment I

Checklist (Trait-Pair Choice Measure) as Used in Asch (1946)


"Choose the characteristic that is most in accordance with the view you have formed."
1. generous-ungenerous
2. shrewd-wise
3. unhappy-happy
4. irritable-good natured
5. humorous-humorless
6. sociable-unsociable
7. popular-unpopular
8. unreliable-reliable
9. important-insignificant
10. ruthless-humane
11. good looking-unattractive
12. persistent-unstable
13. frivolous-serious
14. restrained-talkative
15. self centered-altruistic
16. imaginative-hard headed
17. strong-weak
18. dishonest-honest

30

Replication of Asch (1946)

Attachment II
Grant Application
Activity

Expected costs

Recruiting 1050 participants through MTurk

1050 * $0.70 = $735

Hiring an RA to rate 350 open-ended responses*

$1200

total costs

$1935

*Undergraduate Research Assistants at Radboud University Nijmegen costs 21.21 per hour
(including taxes, social security and health insurance requirements) and may not be paid less due to
University regulations. We expect that an RA will be able to rate 10 responses per hour on the
dimensions specified in the analysis plan (35 hours in total), and that we will need approximately
seven hours extra to have 10% of the responses by an independent rater (another RA), and have the
raters discuss cases on which they do not agree. This means that we will need approximately 42 hours
in total (890, 82, which is approximately $1200).

31

Das könnte Ihnen auch gefallen