running head: evaluative conditioning and impression ...baranany/mb2015.pdf · running head:...

1 RUNNING HEAD: EVALUATIVE CONDITIONING AND IMPRESSION FORMATION

Is Evaluative Conditioning Important? Integration into Impression Formation Research

Would Answer this Question

Tal Moran and Yoav Bar-Anan

Ben-Gurion University of the Negev, Beer-Sheva, Israel

Evaluative conditioning (EC) experiments and impression formation experiments study factors

that influence evaluation. Despite large overlap between these two research traditions, they often

operate in parallel without attention to findings and theories from the other line of research,

risking redundancy. We suggest reintegrating EC research into impression formation as the

branch of research that investigates the effect of stimulus co-occurrence distinct from the effects

of other relational information. Most learning episodes include stimulus co-occurrence and other

relational information (e.g., viewers learn that Batman co-occurs with crime and that he fights

crime). Therefore, EC research is important to the extent that the influence of stimulus co-

occurrence on evaluation is not overridden by the effects of other relational information. We

review initial evidence suggesting that co-occurrence influences evaluation over and above other

relational information, and report two experiments that demonstrate empirical directions for

studying the effects of co-occurrence versus other relational information.

Author’s note: Correspondence should be addressed to: Tal Moran, Department of Psychology, Ben-Gurion University of the

Negev Be’er Sheva, Israel. E-mail: [email protected]. This project was supported by grants from the Israeli Science Foundation

[1012/10], and from Project Implicit Inc. to Y.B.-A.

mailto:[email protected]


Is Evaluative Conditioning Important? Integration into Impression Formation

Research Would Answer this Question

Parallel lines of research in psychology are not rare. For example, Ranganath, Spellman,

and Joy-Gaba (2010) pointed out that theories explaining arguments believability in the social

persuasion literature and in the cognitive category-based induction literature ask similar

questions but use different research paradigms. The integration of the findings from these two

separate literatures provides a more complete picture of the factors that influence argument

believability and inspire new lines of investigation in that area. Overlapping parallel lines of

research could emerge from similar question asked by different disciplines. For example, Steel

and König (2006) pointed out that the fields of economics, decision making, sociology, and

psychology share a common desire to understand human motivation processes, and each field

developed its own motivational theories and paradigms. In other cases, similarities among

concepts and lines of thought that emerge from different research fields contribute to the

understanding of specific phenomena. For example, Carver (2005) argued that work done in

developmental psychology has direct relevance for understanding impulse and constraint in

personality psychology. Another reason for overlooking overlaps between lines of research is a

focus on specific empirical paradigms that were developed to investigate an initial question but

later limit the development of research and theory in that field, and obscure commonalities of

research questions and findings across different paradigms (Meiser, 2011).

In the present article we argue that evaluative conditioning (EC) research and impression

formation research often operate as parallel lines of research, characterized by different

paradigms, while they should be closely integrated. EC research studies the effect of co-

occurrence between a target object and affective stimuli on the evaluation of the target object.

Impression formation research studies how people form impressions of other people and objects.

By this definition, EC should be a branch of impression formation research. However, articles

about EC seldom integrate findings, theories, and concepts from other branches of impression

formation research.

To integrate EC more closely into impression formation research, a main conceptual

challenge is to delineate what differentiates between the typical learning procedures used in EC

research and those used in impression formation research. Most EC procedures present co-


occurrence between the target object and affective stimuli without specifying any relation other

than the co-occurrence. In contrast, most learning paradigms in impression formation research

specify the relation between the target object and affective stimuli. This distinction does not only

allow studying EC as a well-differentiated branch within impression formation research, it would

also help determine the relative importance of EC research. EC research is important to the

extent that the effect of stimulus co-occurrence on evaluation is independent of the effect of

information about other relations between the target and affective stimuli. Below, we explain our

reasoning in more detail, review experiments relevant to the question whether EC is influential

even in the presence of information other than stimulus co-occurrence, report novel experiments

that addressed this question, and discuss the promises of this research direction.

Impression Formation

Impression formation is a rich research tradition that studies how people form

impressions of other people. Asch's (1946) study is often considered the first experimental

attempt to investigate impression formation. Asch's identified three important effects: primacy

effect (the first items on a trait list hold a large influence on impressions), recency effect (the last

items on a trait list hold a large influence on impressions), and centrality effect (central traits

guide the interpretation of other information). Asch’s work inspired two main lines of research.

The first was known as research pertaining to Implicit Personality Theory (Bruner & Tagiuri,

1954; for review see Schneider, 1973). This research studied how perceivers draw inferences

about other people from partial information (e.g., Cantor & Mischel, 1977; Rosenberg & Sedlak,

1972; Wishner, 1960). The second line of research that emerged from Asch's research

investigated Processes of Impression Formation: how perceivers combine different kinds of

information to produce an overall impression (e.g., Anderson, 1965; Cusumano & Richey, 1970;

Rosenkrantz, & Crockett, 1965).

Over the years, impression formation research studied a wide range of questions such as

what factors influence attention during impression formation (e.g., Belmore, 1987; Mayo &

Crockett, 1964), how motivation and mood affect impression formation (e.g., De Dreu & Van

Kleef, 2004; Forgas & Bower, 1987; Neuberg & Fiske, 1987), what factors bias impression

formation (e.g., Skowronski & Carlston, 1989; Ruscher & Hammer, 1966), how category

accessibility affects the influence of ambiguous description on evaluation (e.g., Higgins, Rholes,


& Jones, 1977), how organizational processes affect impression formation (e.g., Hamilton, Katz,

& Leirer, 1996), assimilation versus contrast processes in impression formation (e.g., Martin,

Seta, & Crelia, 1990; Strack et al., 1993), the effects of processing goals (e.g., Chartrand, &

Bargh, 1996), the effects of prior knowledge (e.g., Cohen, 1981), effects of fluency (e.g.,

Sansom-Daly & Forgas, 2010), and more.

Impression formation research is also closely tied to research on persuasion (e.g.,

Hovland, Janis, & Kelly, 1953; Petty & Cacioppo, 1986), a line of research that has flourished

with many empirical investigations and a number of theoretical models. The main difference

between impression formation and persuasion research is that instead of evaluation and

impressions, persuasion research usually studies changes in opinions: whether people support or

oppose certain beliefs, ideologies, and arguments. Still, persuasion research is usually carried out

by presenting information about a target object (e.g., a plan to raise tuition) and studying what

factors influence people favorability toward that object. Because of this overlap, impression

formation and persuasion research are often lumped together as two branches of research on

attitude formation and attitude change, usually without clear boundaries between the two. For

instance, theories about evaluative response are used in research on impression formation and in

persuasion research, on attitudes toward persons and on opinions.

Impression formation and persuasion research subsumes a variety of theories about the

formation and expression of evaluative behavior. For instance, the Information Integration theory

(Anderson, 1981), the Expectancy-value model of attitudes (Fishbein & Ajzen, 1975), the

Elaboration Likelihood Model (ELM; Petty & Cacioppo, 1986), the Motivation and Opportunity

as Determinants of Evaluation (MODE) model (Fazio, 2007), the Associative Propositional

Evaluation (APE) model (Gawronski & Bodenhausen, 2006, 2011), the Meta-Cognitive Model

(MCM; Petty, Briñol & DeMarree, 2007), the systems of reasoning approach to attitude change

(Rydell et al., 2006), and the iterative-reprocessing model of the neural bases of evaluation

(Cunningham & Zelazo, 2007).

Early impression formation research and theories have seldom incorporated research on

conditioning into their models and research questions. However, more recent theories addressed

the effect of co-occurrence on evaluation (e.g., Gawronski & Bodenhausen, 2006; Rydell, et al,

2006). These theories often focused on the differences between automatic and deliberate


evaluation, and, in one way or another, assumed that automatic evaluation is more sensitive than

deliberate evaluation to co-occurrence. Automatic evaluation is an evaluative response that has at

least one of the following features: it is activated without intention, without control, without

awareness or without requiring much cognitive resources (De Houwer, Teige-Mocigemba, &

Spruyt, 2009). Deliberate evaluation is an evaluative response not characterized by any of these

features. Compatible with that theme, several impression formation studies investigated the

effect of co-occurrence on automatic and deliberate evaluation, not always with reference to the

term EC, and usually without reference to previous EC research or theoretical models (Boucher

& Rydell, 2012; Peters & Gawronski, 2011; Petty, Tormala, Briñol, & Jarvis, 2006; Ratliff &

Nosek, 2010; Rydell, McConnell, & Mackie, 2008; Rydell et al., 2006; Siegel, Sigall, & Huber,

2012; Whitfield & Jordan, 2009).

Evaluative Conditioning

Evaluative Conditioning (EC; De Houwer, Thomas, & Baeyens, 2001) refers to changes

in liking due to stimulus co-occurrence (De Houwer, 2007). In a prototypical EC study, a neutral

stimulus (conditioned stimulus; CS) is repeatedly presented with positive or negative stimulus

(unconditioned stimulus; US). The common result is a shift in the evaluation of the formerly

neutral CS such that it becomes similar to the evaluation of the US. Procedures similar to those

studied by modern Evaluative Conditioning research were originally used in research framed as

an investigation of the role of affect and classical conditioning in attitude formation (Zanna,

Kiesler, & Pilkonis, 1970). In the procedures employed by that early research, conditioned

stimuli (usually neutral words or nonsense syllables) co-occurred with positive or negative words

(e.g., Blasford & Sampson, 1964; Cohen, 1964; Staats & Staats, 1958) or with aversive stimulus

such as an electric shock (e.g., Staats, Staats, & Crawford, 1962; Stagner & Britton, 1949). Early

demonstrations of changes in the valence of a natural CS after its co-occurrence with affective

US were criticized as artifacts of demand characteristics (e.g., Kiesler, Collins, & Miller, 1969;

Orne, 1962; Page, 1969). As a consequence, much research effort was invested in demonstrating

that the effect is not confounded by demand characteristics and to determine the role of

awareness (Allen & Madden, 1985; Insko & Oakes, 1966; McGinley & Layton, 1973).

The question of awareness also led researches to make a theoretical distinction between

signal learning—acquisition of propositional-declarative knowledge about stimulus relationships

http://psycnet.apa.org/index.cfm?fa=search.searchResults&latSearchType=a&term=Peters,%20Kurt%20R.


in the environment, mediated by controlled cognitive processes, and evaluative learning—the

process by which an affective reaction evoked by an affective stimulus is transferred to a

previously neutral stimulus. Researchers argued that EC represents the latter and not the former

(e.g., Baeyens, Elen & Van den Bergh, 1990; Levey & Martin, 1987). Further, EC was separated

from classical conditioning with evidence that EC endures conditions that eliminate the effect of

classical conditioning—mainly resistance to extinction (Baeyens, Crombez, Van den Bergh, &

Eelen, 1988; Baeyens, Eelen, Van den Bergh, & Crombez, 1989) and unawareness of

contingency (Baeyens et al., 1990). Although evidence for EC without awareness and for EC’s

resistance to extinction is still controversial (e.g., Gawronski, Gast, & De Houwer, 2015;

Hofmann et al., 2010; Kattner, 2012; Sweldens, Corneille, & Yzerbyt, 2014), EC research has

been established as an independent research tradition (for a review see De Houwer et al., 2001;

Walther, Weil, & Langer, 2011), isolated from research on attitude formation and classical

conditioning—the lines of research from which EC research evolved.

Over the years, EC research mainly investigated the procedural conditions under which

co-occurrence between stimuli results in changes in liking. These conditions include the number

of pairings (e.g., Baeyens, Eelen, Crombez, & Van den Bergh, 1992), the order of stimuli

presentation (e.g., Hammerl, & Grabitz, 1993), extinction (e.g., Baeyens et al., 1988; Gast & De

Houwer, 2013; Gawronski, Gast, & De Houwer, 2015), contingency memory (e.g., Bar-Anan,

De Houwer, & Nosek, 2010; Gawronski & Walther, 2012; Hütter et al., 2012), changes in the

identity or valence of stimuli (e.g., Baeyens, Eelen, Van den Bergh, & Crombez, 1992; Walther,

Gawronski, Blank, & Langer, 2009), and many more.

One turning point in the history of EC research is De Houwer’s (2007) argument that

there is variability in the conceptualization of EC among different EC researchers, and his

suggestion to define EC as an effect. That shift in conceptualization is important because

conditioning is a concept that means to people more than it should. People tend to equate

conditioning with concepts such as automatic, unconscious, associative. De Houwer decoupled

all these implicit assumptions from what is known: the effect observed in numerous studies. De

Houwer defined EC as "a change in the valence of a stimulus that results from pairing the

stimulus with another stimulus” (De Houwer, 2007, p. 230). We prefer to replace pairing with


co-occurrence because there are many methods to pair stimuli other than co-occurrence, yet

virtually all EC research before 2007 investigated co-occurrence and no other pairing methods.

After years of a dominant view of EC as the result of associative processes, De Houwer’s

definition of EC as an effect allowed him to propose the propositional model of EC (De Houwer,

2007; Mitchell, De Houwer, & Lovibond, 2009). The propositional account states that

propositional knowledge about the relation between the stimuli underlies EC. Partly because of

De Houwer’s propositional approach, EC research started to extend beyond stimulus co-

occurrence and investigated other types of information. For example, EC studies investigated the

effect of relational qualifiers that explain the relation between the co-occurring stimuli (Fiedler

& Unkelbach, 2011; Förderer & Unkelbach, 2012; Walther, Langer, Weil, & Komischke, 2011;

Zanon, De Houwer, Gast, & Smith, 2014) and, unsurprisingly, found that people's evaluation

takes the relational information into account. Other studies found that instructions about how to

encode the co-occurrence moderated the effect of co-occurrence on evaluation (Balas &

Gawronski, 2012; Gawronski, Balas, & Creighton, 2014). EC studies also found that context

cues moderated the effect of co-occurrence on evaluation (Bar-Anan & Dahan, 2013; Zanon, De

Houwer, & Gast, 2012).

Why Integrate?

As our review demonstrates, recently, impression formation research has investigated

how co-occurrence affects evaluation. And, lately, EC research started to look at how

information other than co-occurrence affects evaluation. Even more than before, each line of

research reaches into the territory of the other, risking redundancy and ignorance of past relevant

findings and theories. Therefore, we believe that the time is ripe for closer integration between

these two research traditions. On one hand, EC research will benefit from taking into account

theories and findings from impression formation research. Although EC research deals with

evaluative response, general evaluation theories about what influences evaluative response (e.g.,

Anderson, 1981, Fazio, 2007; Fishbein & Ajzen, 1975; Gawronski & Bodenhausen, 2006) are

hardly ever discussed in EC literature. Moreover, impression formation research offers many

theoretical, conceptual and empirical insights that could advance EC research. A relatively recent

example is the distinction between automatic and deliberate evaluation (Fazio, 2007; Gawronski


& Bodenhausen, 2006; Rydell et al., 2006), which EC research hardly used as a tool for

understanding the EC effect (for an exception, see Gawronski, Balas, & Creighton, 2014).

On the other hand, impression formation research has not attempted to integrate EC

research much beyond using the basic EC effect (often without explicitly calling it EC).

Impression formation theories would benefit from taking into account finding about the

conditions under which co-occurrence of stimuli results in changes in liking. Impression

formation theories would increase their power and improve their accuracy if they could explain

EC findings about blocking (e.g., Lipp, Neumann, & Mason, 2001), extinction (e.g., Gawronski

et al., 2015), contingency memory (e.g., Hütter et al., 2012), order of presentation (e.g., Hammerl

& Grabitz, 1993) counter-conditioning (e.g., Walther et al., 2009), and more. In summary,

because both research lines overlap in their questions, they could benefit from more frequent

integration.

EC is a Branch of Impression Formation Research

The key for integrating EC and impression formation is not only to define the overlap

between these two lines of research (both study factors that influence evaluation), but also the

difference between EC research and the typical impression formation research. Most EC research

presented co-occurrence of neutral objects with affective stimuli (mostly images). In contrast, the

typical impression formation research presents traits or behaviors attributed to the target person,

or even more complex information (e.g., the fact that the participant and the target person agree

that drugs should be legalized). In other words, the difference between EC and impression

formation research is that EC paradigms present co-occurrence between the target object and

affective stimuli without specifying any relation other than the co-occurrence. In contrast, in

impression formation research the information about the target usually includes explicit

relational information other than (or additional to) co-occurrence. Most impression formation

research does not only present affective stimuli in spatiotemporal proximity to the target object.

Impression formation research usually specifies the relation between the affective stimuli and the

target object. For instance, impression formation that presented trait information about a person

(e.g., Asch. 1946; Ajzen, 1974; McCarthy & Skowronski, 2011) explicitly indicated that the

traits characterize the target person. In short, EC research is a branch of impression formation

that specializes in the study of the effect of co-occurrence on evaluation.


Is EC important?

One advantage of integrating EC into impression formation research is that it would help

estimating the importance of EC. In our view, EC research is important if stimulus co-occurrence

has an effect on evaluation that is not easily eliminated by other relational information. We

reason that most evaluative learning episodes include co-occurrence between the targets object

and affective stimuli, but also include other relational information. This is true when people

observe other people’s behaviors, and when they are told about other people’s traits. It is even

true when people watch attractive people in spatiotemporal proximity to a product in commercial

ads, a prototypical example used by EC researchers. Most viewers are aware of the complex

relational information the product and the attractive people were presented together for

marketing purposes. If relational information easily overrides the effect of co-occurrence, then

EC has a limited effect on impression formation in everyday life. For instance, the knowledge

that attractive people were deliberately presented with a product to convince viewers to like the

product might erase any effect of the co-occurrence and render EC irrelevant to this situation.

As a first step in integrating EC into impression formation research we suggest focusing

on the factors that moderate the relative effect of co-occurrence on evaluation versus other

relational information about the target objects. Such research would help determine how

common EC is in evaluative learning episodes. That research direction could also shed light on

the factors that moderate EC in general, providing evidence toward better theoretical models of

EC. This line of investigation would also add valuable evidence about factors that determine

evaluative response, and cognitive processes behind evaluative learning. In the rest of the article,

we consider past relevant evidence, and report two novel relevant experiments.

Evidence that EC is Not Eliminated by the Effects of Other Relational Information

So far, no published research has purposely compared the effect of co-occurrence on

evaluation versus the effect of other relational information. However, some experiments provide

evidence relevant to this question. The first evidence we consider comes from research on source

credibility (Gruder et al., 1978; Hovland & Weiss, 1951). In these studies a topic (e.g., a four-

day work week) is presented with positive or negative arguments communicated by a source.

Participants also receive information about the source credibility (high or low). We construe the

arguments as co-occurrence between the subject matter and positive or negative attributes. The


information about the source credibility is another form of information that, via reasoning

processes, should discount (or enhance) the co-occurrence effect. The common finding is that if

people evaluate the topic immediately after receiving the information, the source credibility

affects their evaluation. However, if people evaluate the topic after a delay, the co-occurring

argument influences the evaluation regardless of the source credibility. With the concept of EC

in mind, one possible reason for this finding might be that co-occurrence influences evaluation

despite discounting information, but only if people are not given the chance to cancel the co-

occurrence effect immediately after the arguments presentation.

More relevant evidence comes from studies that tested the effect of information about the

validity of co-occurrence on evaluation (Boucher & Rydell, 2012; Peters & Gawronski, 2011).

Peters and Gawronski (2011) asked participants to form impressions toward four novel men,

each presented with verbal descriptions of positive or negative behaviors. Participants were also

provided with information whether the behavioral description was characteristic or

uncharacteristic of the target person. One man occurred with positive behaviors that

characterized him, and one occurred with characteristic negative behaviors. The other two men

co-occurred with behaviors that were presented as uncharacteristic of them: one with positive

behaviors and one with negative behaviors. As one could expect, Peters and Gawronski’s (2011)

studies found that the information whether the co-occurring behaviors are characteristic or

uncharacteristic of the men had a strong influence on participants’ deliberate evaluation of the

targets. However, co-occurrence still influenced the evaluation independently of the effect of the

validity information. In all three studies, Peters and Gawronski found a reliable main effect of

co-occurring valence on deliberate evaluation: Participants reported more liking of men who

appeared with positive behaviors over men who appeared with negative behaviors (see

Experiment 1 in Boucher & Rydell, 2012, for similar procedure and results).

More evidence comes from research that tested the effect of relational information on

evaluation (Moran & Bar-Anan, 2013). Participants observed targets characters that started

or ended pleasant or unpleasant sounds. Some characters started pleasant music, some started

unpleasant noise, some ended the pleasant music, and some ended the unpleasant noise.

Thus, participants were exposed to co-occurrence information (the characters co-occurred

with either positive or negative sounds) and also relational information (the characters either


started or ended the sounds) that allowed them to infer the characters’ valence (e.g., the

character is positive because he ended the unpleasant noise). As could be expected, the

characters' relations with the sounds had the dominant influence on deliberate evaluation.

Participants preferred characters that ended the unpleasant noise over characters that ended

the pleasant music. However, co-occurrence influenced evaluation independently of the

effect of inference. Overall, participants evaluated characters that co-occurred with pleasant

music more positively than characters that co-occurred with the noise.

Support that the effect of stimulus co-occurrence is not eliminated by the effect of other

relational information comes also from a few EC studies. Participants in Förderer and

Unkelbach’s (2012) study observed the co-occurrence of CSs (pictures of men) with positive

(e.g., puppies) or negative (e.g., snakes) USs provided simultaneously with a relation qualifier

that was either loves or loathes (e.g., X loathes snake). As could be expected, the relational

qualifiers had a dominant effect on evaluation: People liked targets that loathed snakes and

disliked targets that loathed puppies. However, co-occurrence still influenced evaluation

independently of the effect of relational qualifiers. Participants liked targets that co-occurred

with positive stimuli more than targets that co-occurred with negative stimuli (S. Förderer,

personal communication, June, 6, 2013).

Zanon, De Houwer, and Gast (2012) presented to participants nonwords (CSs) or pairs of

nonwords that co-occurred with a positive outcome (a win) or a negative outcome (a loss). The

study manipulated a patterning rule; in one condition (the opposition condition), two nonwords

that co-occurred with positive US when appearing without other nonwords (CS1-USpos; CS2-

USpos) co-occurred with a negative US when they appeared together (CS1CS2-USneg). In the other

condition (the similarity condition), the nonwords appeared with the same US, even when they

appeared with another nonword (CS1CS2-USpos). After learning, participants evaluated each CS,

when presented alone on the screen. The critical question was the evaluation of nonwords that

appeared only in pairs during the learning phase (e.g., CS3CS4-USpos, CS5CS6-USneg,). These

nonwords co-occurred with US of one valence, but in the opposition condition, participants

could infer that when the CS is alone, its outcome is opposite of the outcome with which it co-

occurred when appearing in pairs. Zanon et al. found support of the assumption that the

hypothetical inference attenuated the evaluation (e.g., smaller preference for CS3 over CS5 in the


opposition than in the similarity condition). More relevant to the present question, Zanon et al.

found that overall, independently of the effect of the patterning rule, participants reported more

positive evaluations of nonwords that co-occurred with positive outcome (CS3 and CS4 in our

example) than nonwords that co-occurred with negative outcome (CS5 and CS6).

Finally, Gawronski, Walther, and Blank (2005) tested the role of cognitive balance in the

formation of interpersonal attitudes. In their study, participants first formed impressions of

individuals based on the behavioral descriptions. Then, these individuals co-occurred with new

individuals with the information that the known individual either likes or dislikes the new

individual. Participants showed positive evaluation of targets liked by positive people and of

targets disliked by negative people. Participants showed negative evaluation of targets liked by

negative people and of targets disliked by positive people. However, a main effect of co-

occurrence still emerged: overall, independently of the effect of the relational qualifier

(likes/dislikes), participants evaluated the new targets that co-occurred with positive known

individuals more positively than targets that co-occurred with negative known individuals.

In summary, some previous research found evidence suggesting that co-occurrence has

an effect on evaluation independent of the effect of other relational information. In these studies,

although relational information other than co-occurrence usually had the dominant influence on

evaluation, co-occurrence still had an independent effect on evaluation.

Evidence that Relational Information Overrides the Effect of Co-occurrence

In contrast to the evidence reviewed above, some previous research did not find any

effect of co-occurrence on evaluation when other explicit relational information was available. In

one study (Petty, Tormala, Briñol, & Jarvis, 2006) participants first observed an EC procedure

that paired an image of one person with positive stimuli and an image of another person with

negative stimuli. Afterwards, participants were exposed to the attitudes held by the two persons:

one of them largely agreed with the participants on important issues, and the other disagreed. In

one condition, the person who agreed with the participant was the CSpos from the pairing

procedure, and in the other condition that person was the CSneg. Participants self-reported a more

positive evaluation of the person who agreed with them than of the person who disagreed with

them. The co-occurrence in the pairing procedure did not moderate that preference.


In another relevant study (Siegel, Sigall, & Huber, 2012), participants observed co-

occurrence between names of members of one group and positive adjectives, and between names

of members of another group and negative adjectives. In one condition, prior to that procedure,

participants were told that the adjectives accurately described the groups. In the other condition,

participants were told that the adjectives were randomly paired with each group. The evaluation

of participants who were told that the adjectives described the persons who co-occurred with

them was sensitive to the co-occurrence. Although the pattern of results among participants who

were told the adjectives were randomly paired with the social groups reflected an EC effect, the

EC effects were small (ds = 0.31, 0.19, in Experiments 1 and 2, respectively), and did not reach

conventional statistical reliability (the samples were smaller than 30 in each experiment).

The Present Experiments

Our review suggests that the effect of co-occurrence on evaluation often persists even if

other relational information pushes evaluation to a different direction. However, the main effect

of co-occurrence over and above other relational information was an incidental finding in most

studies. Little research efforts were invested in investigating the effect of co-occurrence when

relational information pushes evaluation to valence opposite of the US. As a consequence,

knowledge of what factors moderate this effect is scarce. So far we presented arguments why

investigating those moderators is critical for EC research. Next, we present two experiments,

each demonstrating a different research direction pertinent to the question. One research

direction examines whether assigning specific relational meaning to co-occurrence moderates the

effect of co-occurrence. The other research direction investigates what third variables moderate

the relative effect of co-occurrence versus other relational information. We consider each

experiment a different prototype of a research direction with good potential for advancing

knowledge about the role of EC in impression formation.

There are many possible methods to provide relational information in addition to the co-

occurrence. Relation words such as loathes (Förderer & Unkelbach, 2012) and ends (Moran &

Bar-Anan, 2013) are a direct verbal method. Another method is to “inject” relational meaning

into co-occurrence itself. For instance, Siegel et al. (2012) explicitly informed participants, in

advance, whether the CSs co-occurred with adjectives that described them or not. In our

conceptualization, they specified the relational meaning of co-occurrence. When participants


were informed that co-occurrence does not imply any relation between the CS and the US, Siegel

et al. did not find reliable EC effect, perhaps because of small sample size (the preferences were

consistent with EC). In Experiment 1, we continue in the same research direction, with a much

larger sample, and with different relational meaning attached to co-occurrence. Instead of

specifying that co-occurrence is random, we exposed participants to a context in which stimuli of

opposite valence co-occur. We examined whether such a context would inject an opposition

meaning to co-occurrence and cause reversed EC when participants observe neutral stimuli co-

occur with affective stimuli.

In Experiment 2, we tested focus on valence as a possible moderator of the relative effect

of co-occurrence versus relational information on evaluation. We provided participants with

verbal relational information with each co-occurrence, explicitly stating the relation between the

CS and US. We tested whether participants’ focus during learning on the evaluation of the US

versus the evaluation of the CS as inferred from the specific relation would moderate the relative

effect of co-occurrence on evaluation1.

Experiment 1

The goal of Experiment 1 was to test whether the effect of co-occurrence on evaluation

can be reversed by exposing participants to an environment in which stimuli of opposite valence

tend to co-occur. Participants completed a sequential priming task. They categorized target

images as pleasant and unpleasant. Before each image, one of six strings of letters appeared. Two

letter strings were unpleasant words, two were pleasant words, and two were nonsense words.

The experiment had two conditions. In both conditions, one nonsense words appeared before

pleasant images, and the other before unpleasant images. In previous research, such co-

occurrence induced the typical EC effect (Schmidt & De Houwer, 2012). In the similarity

condition, the four affective words always appeared before images of the same valence (e.g.,

pleasant words always preceded pleasant images). In the opposition condition, the four affective

words always appeared before images of the opposite valence (e.g., pleasant words always

preceded unpleasant images).

1 We report all data exclusions, manipulations, and measures, and how we determined our sample sizes.

To see the materials and data of the whole project (Experiments 1-2) visit https://osf.io/5fnzr/.

https://osf.io/5fnzr/


We hypothesized that co-occurrence has a relational meaning of similarity that is flexible

to temporary changes. Therefore, we assumed that in the opposition condition, EC will reverse.

A rival hypothesis is that the common effect of co-occurrence on evaluation has nothing to do

with the relational meaning of co-occurrence, and therefore changing the rules about which

stimuli tend to co-occur would have no effect on EC.

Method

Participants. In both studies, we planned to collect at least 320 participants for high

statistical power (95% chances of detecting an effect of ηp2 = .02 for the interaction between

EC and the tested moderating factor). Participants in both experiments were volunteers at the

Project Implicit research website (Nosek, 2005). Of the 523 participants who started the

study, 323 (62%) completed all the measures (because the participants at the website are not

obliged or paid for their participation, 50%-70% is currently the range of the typical

completion rate). We excluded from the analyses 13 participants (4%) with more than 40%

trials faster than 300ms on the sequential priming task, and/or more than 40% errors in the

task. The final sample included 310 participants (59% women, Mage = 32.37, SDage = 13.42).

Materials. The CSpos and CSneg were chosen randomly for each participant from a pool of

three nonewords: "BAUBED", "EBELSA", and "QUESTL." The affective words were four

positive words (Love, Pleasure, Pretty and Wonderful), and four negative words (Hate, Pain,

Disgusting and Noxious). Two positive words and two negative words from these lists were

chosen randomly for each participant. The targets on the task were four images of happy

attractive men and women and four images of unhappy unattractive men and women.

Design. The design was 2 (co-occurrence meaning: same, different; between participants)

X 2 (CS type: occurred with positive, co-occurred with negative; within participants).

Learning Procedure. Participants completed an evaluative priming task (Fazio et al.,

1995). In each trial, the prime appeared at the center of the screen for 250ms, then replaced by

the target. The target appeared until response, or 3000ms. After 3000ms with no response, a red

“PLEASE RESPOND FASTER” message appeared on the screen, and the trial ended. In each of

three blocks, each letter string appeared in eight trials (total of 24 pairing), always before images

of one valence.


The instructions informed participants that words and images would appear on the screen.

Participants were instructed to ignore the words and categorize the images as unpleasant (by

pressing the left key) or pleasant (by pressing the right key). In the similarity condition, the

affective words always preceded images of their own valence (e.g., Love always preceded

pleasant people). In the opposition condition, the affective words always preceded images of

opposite valence (e.g., Love always preceded unpleasant people).

Evaluation. Afterwards, participants rated the CSs on three different questions, presented

in random order: "Based on only visual appearance, how pleasant is this word's appearance?",

"Based on what you feel right now, how much do you like this word?", and "If this word had a

meaning, how positive do you feel that the meaning would have been?" The response scale

ranged from 1 (Not at all) to 9 (Extremely).

Results and Discussion

For each CS, we averaged the rating of the three questions (α = .80). A 2 (co-occurrence

meaning: similarity, opposition) x 2 (CS valence: positive, negative) ANOVA on evaluation,

revealed a main effect of valence, F(1, 286) = 16.39, p < .0001, ηp2 = .054, 90% CI [.01, .10],

reflecting the EC effect. This result was against our hypothesis because we assumed that EC will

be canceled across conditions. There was no effect of co-occurrence meaning condition, F < 1,

ηp2 < .01. The EC effect was moderated by co-occurrence meaning, F(1, 286) = 7.06, p = .0083,

ηp2 = .024, 90% CI [.003, .06]. The interaction reflected a substantial difference in EC between

conditions. The average difference between the CSpos evaluation and the CSneg evaluation was M

= 0.76, SD = 1.99, d = 0.38, t(162) = 4.89, p < .0001, in the similarity condition, and only M =

0.15, SD = 2.09, d = 0.07, t(146) = 0.88, p = .38, in the opposition condition.

The results suggest that the meaning of co-occurrence can temporarily change leading to

a diminished EC. However, attaching an opposition meaning to co-occurrence did not reverse its

effect on EC. The lack of reverse fits the initial conclusion suggested by our review of relevant

findings: co-occurrence influence evaluation despite other relational information. In the present

case, although opposite stimuli clearly appeared in 96 (66%) out of 144 trials, and similar stimuli

never co-occurred, the effect of co-occurrence was not reversed. Therefore, it seems that the

common effect of co-occurrence is not easily reversed by reasoning from additional information.


On the other hand, we found that EC was eliminated by attaching an opposition meaning

to co-occurrence. These results suggest that there might have been two forces that pushed

evaluation to two different ends: co-occurrence had an EC effect that was balanced by the

inference from the new relational meaning attached to co-occurrence. In summary, the results of

Experiment 1 refute two opposing extreme hypotheses. The effect of co-occurrence is not

constant, but it is also not completely constructed from inference about the type of relation co-

occurring stimuli tend to have. Further research is needed to reveal more about what happens

when people are exposed to an environment in which opposite stimuli co-occur. From our

results, it seems that people are either affected by two opposing forces, or that there are

individual differences in people’s reaction to such situations. One clear next step is to measure

individual differences in people’s explicit knowledge about the co-occurrence rules introduced in

each condition (e.g., the opposition rule), and examine whether the knowledge predicts EC.

Experiment 2

From the incidental evidence reviewed earlier and from the results of Experiment 1, it

seems difficult to completely erase the typical effect of co-occurrence on evaluation.

However, it is clear that the relative effect of co-occurrence versus the effects of other

relational information is inconsistent. Whereas some experiments found that clear explicit

relational information is more dominant than the effect of co-occurrence (e.g., Moran & Bar-

Anan, 2013; Förderer, & Unkelbach, 2012), others found that the effect of co-occurrence was

equal to the effect of other relational information (e.g., Experiment 1 in the present article;

Zanon et al., 2012), and some were not able to detect any effect of co-occurrence above and

beyond the effect of other relational information (e.g., Petty et al., 2006). In Experiment 2 we

directly test a possible moderator of the co-occurrence effect: whether people focus on

evaluating the US or on inferring the evaluative meaning of the specific CS-US relation.

Participants observed men that either gave them or took away from them positive or

negative stimuli. The men who took away affective stimuli co-occurred with stimuli of one

valence but performed an action that had an outcome of opposite valence. For instance, a

man who took from the participants positive objects performed a negative behavior but co-

occurred with positive valence. We tested whether co-occurrence has a stronger influence on

the evaluation of the target men when participants focus on the evaluation of the stimuli with


which the men co-occurred than when participants focus on the evaluation of the men’s

behavior.

Method

Participants. Of the 2,098 participants who started the experiment, 1,092 (52%)

participants completed all the measures. We inadvertently obtained many more participants

than planned before stopping collection and analyzing the data. 119 participants (11%) who

did not pay attention to the manipulation were excluded2. The final sample included 973

participants (63% women, Mage = 41.76, SDage = 18.02).

Materials. The four targets were pictures of young adult white males (Minear & Park,

2004; pre-tested by Bar-Anan & Amzaleg-David, 2014). We named them Chris, James, Michael

and David. The affective stimuli were eight images of cute animals (e.g., a puppy) and eight

images of nasty animals (e.g., a cockroach).

Design. The design was 2 (focus: on men’s action, on US valence; between participants)

X 2 (US valence: positive, negative; within participants) X 2 (action’s valence: positive,

negative; within participants).

Learning procedure. The experiments started with the learning procedure. Before the

procedure started, participants were told that they would meet four men and that they should

learn what actions each of them performed and form impressions of the men.

Each trial started with a 400ms fixation (different images of forests). Afterwards, a man

image and name appeared on the left side of the screen next to an image of cute or nasty animals

on the right side of the screen for 400ms. Then, text was displayed between the man and the

animals, indicating either gives or takes away. That was the display of the full action and it

2 We included in the analyses only participants whose performance in the learning task indicated that they

paid attention. Among participants who had to evaluate the US, we included only participants whose

mean rating of the positive animals was more positive than the mean rating of the negative animals (96%

of the participants who completed this condition). Among participants who had to evaluate the men’s

actions, we included only participants whose mean rating of the "give positive animals" action was more

positive than the mean rating of the "give negative animals" action, and that their mean rating of the "take

away negative animals" action was more positive than the mean rating of the "take away positive

animals" action (83% of the participants who completed that condition).


remained on the screen for 1600ms, before disappearing. The task had 32 trials (eight trials of

each man) presented randomly. Instructions informed participants in advance that each man

always performed the same action: one man gave the participants positive animals, one gave

negative animals, one took negative animals from the participant, and one took positive animals.

To manipulate focus, after each trial we presented a rating task. In the US focus condition

we presented the animal image that appeared on the preceding trial with the instructions “Please

rate how positive or negative this animal is.” In the action focus condition we presented again

the relation word and the animal image that appeared on the preceding trial with the instructions

“Please rate how positive or negative this action is.” In both conditions the rating scale was 1

(very negative), 2 (negative), 3 (positive), and 4 (very positive).

Evaluation. After the learning task, participants reported on a 7-point scale how much

they liked each man (1=dislike extremely, 7=like extremely)3.

Results and Discussion

The upper part of Table 1 presents the evaluation scores for each target type, for each

focus condition. We submitted the evaluation scores to a 2 (focus) x 2 (US valence) x 2 (action’s

valence) ANOVA. A main effect of action’s valence, F(1, 971) = 269.20, p < .001, ηp2

= .23,

90% CI [.18, .25], reflected more liking of targets that helped participants (M = 4.46, SD = 1.12),

than targets that harmed participants (M = 3.44, SD = 1.10). The action effect was moderated by

focus condition, F(1, 971) = 7.85, p = .005, ηp2 = .008, 90% CI [.001, .01], reflecting a stronger

effect of action’s valence among participants who focused on the valence of the action, F(1, 458)

= 163.3, p < .001, ηp2 = .26, 90% CI [.20, .31], than among participants who focused on the US

valence, F(1, 513) = 128.37, p < .001, ηp2 = .20, 90% CI [.15, .24].

Replicating the reviewed evidence that co-occurrence influence evaluation independently

of the effect of other relational information, the ANOVA found a main effect of US valence, F(1,

971) = 37.54, p < .001, ηp2

= .03, 90% CI [.02, .05], reflecting more liking of targets who co-

3 For exploratory reasons, all participants completed a Brief IAT (Sriram & Greenwald, 2009) after the

deliberate evaluation measure. The results were very similar to those found with deliberate evaluation

(see full details in https://osf.io/9nr35/).

https://osf.io/9nr35/


occurred with positive animals (M = 4.08, SD = 0.87) than of targets who co-occurred with

negative animals (M = 3.82, SD = 0.90).

Table 1

Experiment 2: Evaluation as a function of co-occurring valence and role valence in each focus

condition Target role Positive Positive Negative Negative

Co-occurring valence Positive Negative Positive Negative

All participants

Focus on men’s action N = 459 4.57a (1.44) 4.47a (1.47) 3.45c (1.35) 3.18d (1.46)

Focus on US valence N = 514 4.60a (1.42) 4.23b (1.44) 3.70c (1.42) 3.40d (1.42)

Only participants with perfect comprehension

Focus on men’s action N = 185 5.60a (1.22) 5.11b (1.68) 2.88c (1.48) 2.12d (1.18)

Focus on US valence N = 187 5.63a (1.20) 4.47b (1.79) 3.42c (1.83) 2.34d (1.24)

Notes. Participants with perfect comprehension evaluated each of the two men who helped them more

positively than each of the two men who harmed them. On each row, different subscripts indicate a significant

difference, tested as contrasts in the ANOVA.

Finally, we found very weak evidence in support of our prediction that focus would

moderate the effect of US valence, F(1, 971) = 3.25, p = .071, ηp2 = .003, 90% CI [.00, .01]. As

expected, the interaction reflected stronger co-occurrence effect when participants focused on US

valence, F(1, 513) = 27.21, p < .001, ηp2 = .05, 90% CI [.02, .08], than when they focused on the

action’s valence, F(1, 458) = 11.83, p = .006, ηp2 = .02, 90% CI [.006, .05].

The results replicated the reviewed evidence that co-occurrence influences evaluation

even when other relational information pushed evaluation away from the US valence. As for our

hypothesized moderation, the results were in the predicted pattern but very weak. This could be

considered more evidence in support of the sustainability of the co-occurrence effect. Even when

participants focused on the valence that should be inferred from the explicit relational

information (e.g., taking away positive animals is a negative behavior), the valence of the co-

occurring valence influenced the evaluation. On the other hand, it could be that the moderator

effect is stronger in reality but our learning procedure was too difficult for some participants who

failed learning the role of each man. To investigate that possibility, we repeated the same

analysis with participants who rated each of the two men who helped them more positively than

the two men who harmed them (40% of the participants in the focus on men’s action condition,

and 36% of the participants on the focus on US valence condition). These participants showed

strong evidence that they understood the roles of each man correctly. That is, these people

understood the relational information well and correctly picked up on what each target man has


done during the game. The lower part of Table 1 presents the evaluation scores of this

subsample. The results remained similar, but the effects were usually stronger than those found

with the whole sample.

Specifically, we found a main effect of the action’s valence, F(1, 370) = 584.97, p < .001,

ηp2

= .61, 90% CI [.56, .65], that was moderated by focus condition, F(1, 370) = 10.97, p = .001,

ηp2 = .02, 90% CI [.007, .06], reflecting a stronger action valence effect among participants who

rated the actions during learning, F(1, 184) = 358.01, p < .001, ηp2 = .66, 90% CI [.59, .70], than

among participants who rated the US during learning, F(1, 186) = 230.53, p < .001, ηp2 = .55,

90% CI [.47, .61]. Importantly, although these participants were selected because they showed

good understanding of the additional relational information (the men’s actions), they still showed

a strong effect of US valence, F(1, 370) = 142.79, p < .001, ηp2

= .27, 90% CI [.21, .33],

reflecting more liking of men who co-occurred with positive animals (M = 4.38, SD = 0.85) than

men who co-occurred with negative animals (M = 3.51, SD = 0.94). Importantly, with this

subsample the interaction between co-occurring valence and focus was stronger, F(1, 370) =

11.87, p = .0006, ηp2 = .03, 90% CI [.008, .06], again reflecting a stronger co-occurrence effect

when participants evaluated the US during learning, F(1, 186) = 92.68, p < .001, ηp2 = .33, 90%

CI [.24, .41], than when they evaluated the target’s actions, F(1, 184) = 50.56, p < .001, ηp2 = .21,

90% CI [.13, .29].

The analysis with a subsample that showed excellent comprehension of the whole

information provided in the learning procedure suggests that focus on evaluation during learning

is probably a reliable moderator of the effect of co-occurrence on evaluation, independently of

the effect of other relational information. In all conditions and all samples we found, again,

strong evidence that co-occurrence pushes the evaluation toward the US valence regardless of

the CS-US specific relation. Experiment 2 suggests that this effect is stronger when, during

learning, people focus on the evaluation of the US and that it is weaker when they focus on the

evaluative implication of the specific CS-US relation.

General Discussion

The main message of the present article is a call for closer integration of evaluative

conditioning (EC) research into impression formation research. The two research traditions study

how people form impressions of other people and objects. We consider EC research a branch of


impression formation research that specializes in the effect of co-occurrence between a target

object and affective stimuli on the evaluation of the target object. Yet, research about EC has

seldom borrowed any knowledge from other branches of impression formation research.

Although EC research investigates the causal chain from stimulus co-occurrence to evaluative

response, present EC theories focus mainly on processes during learning with no clear

assumptions about the processes that lead to the evaluative response. It stands to reason that

theoretical models and findings pertaining to how evaluative information is stored in memory

and how that information is activated and influence behavior are relevant to the question of what

memory forms upon exposure to novel information (i.e., the CS-US co-occurrence).

EC research has largely ignored knowledge accumulated in impression formation

research. For instance, there is little or no attention in EC research to findings and theories about

how evaluative information is integrated (Anderson, 1981), how different goals and motivations

influence evaluative learning (Chaiken, Liberman, & Eagly, 1989; Hamilton, Katz, & Leirer,

1980; Neuberg & Fiske, 1987; Petty & Cacioppo, 1986), the formation of ambivalent objects

(Priester, Petty, & Park, 2007; DeMarree et al., 2014), and the difference between automatic and

deliberate evaluative response (Gawronski & Bodenhausen, 2006). This neglect becomes more

costly now that research under the EC tradition has started investigating factors such as

motivation and explicit relational information (e.g., Balas & Gawronski, 2012; Förderer &

Unkelbach, 2012). With such additions, EC researchers might soon establish a new line of

impression formation research, parallel to current impression formation research. Such a

development is inefficient and might hinder scientific progress. Instead, we believe that the

integration of findings, theories and concepts from impression formation research into EC

research and vice versa will make a valuable contribution to scientific progress.

Integrating EC more closely into impression formation requires defining the unique

features of the procedures studied so far under that line of research. In our perspective, the main

distinction is that EC paradigms present co-occurrence between a target object and affective

stimuli without specifying any relation other than the co-occurrence. In contrast, impression

formation paradigms specify the relation between the target object and affective stimuli.

The integration of EC into impression formation research does not only entail considering

theories and findings from impression formation research. The integration would also stir


empirical investigations into one of the most critical questions about the EC effect: how

important is EC relatively to other evaluative learning effects? Typical EC paradigms might

seem contrived, exposing participants to situations that rarely occur in everyday life. In reality,

people usually do not perceive stimulus co-occurrence without having any information about the

reason for the co-occurrence. In other words, co-occurrence is rarely the only relation known

between the CS and the US. Therefore, it is crucial to study whether the effect of CS-US co-

occurrence on evaluation persists even when people receive information about other relations.

For that reason, the integration that we propose is not only of theoretical knowledge. We propose

conducting studies that add relational information to co-occurrence, test the effects of different

types of relational information, and find what factors moderate the sensitivity of evaluation to

co-occurrence versus other relational information.

In this paper we presented two possible research directions to investigate these important

questions. In Experiment 1, we tested whether it is possible to change the relational meaning of

co-occurrence. We exposed participants to a context in which stimuli of opposite valence co-

occur and examined whether such context would reverse EC. Experiment 1 was an example for

research about whether a specific method to convey relational information eliminates EC. In

Experiment 1, the information was about the meaning of co-occurrence, provided with examples

what types of stimuli tend to co-occur in a particular environment. There could be many other

types of relations conveyed in many different methods. For instance, our lab started investigating

whether Gestalt principles that determine whether people perceive different stimuli as members

of the same or different groups serve as relational information that enhances, diminishes, or even

completely reverses EC (e.g., would EC occur when the CS and the US appear moving into

different directions?).

In Experiment 2, we provided participants verbal relational information with each co-

occurrence, explicitly stating the relation between the CS and US. We found that participants'

focus during learning on the evaluation of the US versus the evaluation of the CS as inferred

from the specific relation moderates the relative effect of co-occurrence on evaluation. The

experiment was an example for investigation of third variables that might moderate the relative

effect of co-occurrence versus other relations. There are numerous other candidates for

moderators. For instance, perhaps because the effect of co-occurrence is illogical when other


relational information suggests a different evaluation, people who tend to favor feelings over

reason would show a stronger EC than people who favor reason over feelings.

In both experiments we found strong evidence that co-occurrence affects evaluation

despite other relational information that pushes evaluation to valence opposite of the US. In

Experiment 1, attaching an opposition meaning to co-occurrence did not reverse EC. In

Experiment 2, the main effect of co-occurring valence was strong despite the effect of relational

information. These results suggest that the effect of co-occurrence is a unique effect, independent

of the effect of explicit relational information. This evidence is compatible with the incidental

findings we reviewed from previous experiments. Thus, critically, initial evidence allows

optimism regarding EC: it is an important effect, possibly present in most episodes of evaluative

learning in everyday life.

References

Allen, C. T., & Madden, T. J. (1985). A closer look at classical conditioning. Journal of consumer

research, 12(3), 301-315.

Anderson, N. H. (1965). Averaging versus adding as a stimulus-combination rule in impression

formation. Journal of Experimental Psychology, 70(4), 394-400.

Anderson, N. H. (1981). Foundations of information integration theory. New York: Academic Press.

Asch, S. E. (1946). Forming impressions of personality. The Journal of Abnormal and Social

Psychology, 41(3), 258-290.

Bar-Anan, Y., & Amzaleg-David, E. (2014). The effect of evaluation on co-occurrence memory

judgment. Cognition & Emotion, 28(6), 1030-1046.

Bar-Anan, Y., & Dahan, N. (2013). The effect of comparative context on evaluative

conditioning. Cognition & emotion, 27(2), 367-375.

Bar-Anan, Y., De Houwer, J., & Nosek, B. A. (2010). Evaluative conditioning and conscious knowledge

of contingencies: A correlational investigation with large samples. The Quarterly Journal of

Experimental Psychology, 63(12), 2313-2335.

Baeyens, F., Crombez, G., Van den Bergh, O., & Eelen, P. (1988). Once in contact always in contact:

Evaluative conditioning is resistant to extinction. Advances in behavior research and

therapy, 10(4), 179-199.

Baeyens, F., Eelen, P., Crombez, G., & Van den Bergh, O. (1992). Human evaluative conditioning:

Acquisition trials, presentation schedule, evaluative style and contingency awareness. Behaviour

research and therapy, 30(2), 133-142.

Baeyens, F., Eelen, P., & Van den Bergh, O. (1990). Contingency awareness in evaluative conditioning:

A case for unaware affective-evaluative learning. Cognition and Emotion, 4(1), 3-18.


Baeyens, F., Eelen, P., Van den Bergh, O., & Crombez, G. (1989). Acquired affective-evaluative value:

Conservative but not unchangeable. Behaviour Research and Therapy, 27, 279-287.

Baeyens, F., Eelen, P., Van den Bergh, O., & Crombez, G. (1992). The content of learning in human

evaluative conditioning: Acquired valence is sensitive to US revaluation.

Learning and Motivation, 23(2), 200-224.

Balas, R., & Gawronski, B. (2012). On the intentional control of conditioned evaluative

responses. Learning and Motivation, 43(3), 89-98.

Belmore, S. M. (1987). Determinants of attention during impression formation. Journal of Experimental

Psychology: Learning, Memory, and Cognition, 13(3), 480-489.

Blandford, D. H., & Sampson, E. E. (1964). Induction of prestige suggestion through classical

conditioning. The Journal of Abnormal and Social Psychology, 69(3), 332-337.

Boucher, K. L., & Rydell, R. J., (2012). Impact of Negation Salience and Cognitive Resources on

Negation During Attitude Formation. Personality and Social Psychology Bulletin, 38, 1329 –

1342.

Bruner, J. S., & Tagiuri, R. (1954). The perception of people. In G. Lindzey (Ed.), Handbook of social

psychology (Vol. 2). Cambridge, MA: Addison Wesley.

Cantor, N., & Mischel, W. (1977). Traits as prototypes: Effects on recognition memory. Journal of

Personality and Social Psychology, 35(1), 38-48.

Carver, C. S. (2005). Impulse and constraint: Perspectives from personality psychology, convergence

with theory in other areas, and potential for integration. Personality and social psychology

review, 9(4), 312-333.

Chaiken, S., Liberman, A., & Eagly, A. H. (1989). Heuristic and systematic information processing

within and beyond the persuasion context. In J. S. Uleman & J. A Bargh (Eds.), Unintended

thought (pp. 212-252). New York: Guilford Press.

Chartrand, T. L., & Bargh, J. A. (1996). Automatic activation of impression formation and memorization

goals: Nonconscious goal priming reproduces effects of explicit task instructions. Journal of


Cohen, B. H. (1964). Role of awareness in meaning established by classical conditioning. Journal of

Experimental Psychology, 67(4), 373-378.

Cohen, C. E. (1981). Person categories and social perception: Testing some boundaries of the processing

effect of prior knowledge. Journal of Personality and Social Psychology, 40(3), 441-452.

Cunningham, W. A., & Zelazo, P. D. (2007). Attitudes and evaluations: A social cognitive neuroscience

perspective. Trends in cognitive sciences, 11(3), 97-104.

Cusumano, D. R., & Richey, M. H. (1970). Negative salience in impressions of character: Effects of

extremeness of stimulus information. Psychonomic Science, 20(2), 81-83.

De Dreu, C. K., & Van Kleef, G. A. (2004). The influence of power on the information search,

impression formation, and demands in negotiation. Journal of Experimental Social

Psychology, 40(3), 303-319.

De Houwer, J. (2007). A conceptual and theoretical analysis of evaluative conditioning. Spanish Journal

of Psychology, 10, 230-241.

De Houwer, J., Teige-Mocigemba, S., Spruyt, A., & Moors, A. (2009). Implicit measures: A normative

analysis and review. Psychological bulletin, 135(3), 347-368.


De Houwer, J., Thomas, S., & Baeyens, F. (2001). Association learning of Likes and dislikes: A review

of 25 years of research on human evaluative conditioning. Psychological Bulletin, 127, 853–869.

DeMarree, K. G., Wheeler, S. C., Briñol, P., & Petty, R. E. (2014). Wanting other attitudes: Actual–

desired attitude discrepancies predict feelings of ambivalence and ambivalence

consequences. Journal of Experimental Social Psychology, 53, 5-18.

Fazio, R. H. (2007). Attitudes as object-evaluation associations of varying strength. Social Cognition, 25,

603-637.

Fazio, R. H., Jackson, J. R., Dunton, B. C., & Williams, C. J. (1995). Variability in automatic activation

as an unobtrusive measure of racial attitudes: a bona fide pipeline?. Journal of personality and

social psychology, 69(6), 1013-1027.

Fiedler, K., & Unkelbach, C. (2011). Evaluative conditioning depends on higher order encoding

processes. Cognition and Emotion, 25(4), 639-656.

Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory and

research. Reading, MA: Addison-Wesley.

Förderer, S., & Unkelbach, C. (2012). Hating the cute kitten or loving the aggressive pit-bull: EC effects

depend on CS–US relations. Cognition & emotion, 26(3), 534-540.

Forgas, J. P., & Bower, G. H. (1987). Mood effects on person-perception judgments. Journal of

personality and social psychology, 53(1), 53-60.

Gast, A., & De Houwer, J. (2013). The influence of extinction and counterconditioning instructions on

evaluative conditioning effects. Learning and Motivation, 44(4), 312-325.

Gawronski, B., Balas, R., & Creighton, L. A. (2014). Can the formation of conditioned attitudes be

intentionally controlled? Personality and Social Psychology Bulletin, 40, 419- 432.

Gawronski, B., & Bodenhausen, G. V. (2006). Associative and prepositional processes in evaluation: An

integrative review of implicit and explicit attitude change. Psychological Bulletin, 132(5), 692-

731.

Gawronski, B., & Bodenhausen, G. V. (2011). The associative-propositional evaluation model: Theory,

evidence, and open questions. Advances in Experimental Social Psychology, 44, 59-127.

Gawronski, B., Gast, A., & De Houwer, J. (2015). Is evaluative conditioning really resistant to

extinction? Evidence for changes in evaluative judgements without changes in evaluative

representations. Cognition and Emotion, 29(5), 816-830.

Gawronski, B., & Walther, E. (2012). What do memory data tell us about the role of contingency

awareness in evaluative conditioning?. Journal of Experimental Social Psychology, 48(3), 617-

623.

Gawronski, B., Walther, E., & Blank, H. (2005). Cognitive consistency and the formation of interpersonal

attitudes: Cognitive balance affects the encoding of social information. Journal of Experimental

Social Psychology, 41(6), 618-626.

Gruder, C. L., Cook, T. D., Hennigan, K. M., Flay, B. R., Alessis, C., & Halamaj, J. (1978). Empirical

tests of the absolute sleeper effect predicted from the discounting cue hypothesis. Journal of


Hamilton, D. L., Katz, L. B., & Leirer, V. O. (1980). Cognitive representation of personality impressions:

Organizational processes in first impression formation. Journal of Personality and Social

Psychology, 39(6), 1050-1063.


Hammerl, M., & Grabitz, H. J. (1993). Human evaluative conditioning: Order of stimulus

presentation. Integrative Physiological and Behavioral Science, 28(2), 191-194.

Higgins, E. T., Rholes, W. S., & Jones, C. R. (1977). Category accessibility and impression

formation. Journal of Experimental Social Psychology, 13(2), 141-154.

Hofmann, W., De Houwer, J., Perugini, M., Baeyens, F., & Crombez, G. (2010). Evaluative conditioning

in humans: a meta-analysis. Psychological bulletin, 136(3), 390-421.

Hovland, C. I., I. L. Janis, and H. H. Kelley (1953), Communication and Persuasion. New Haven, CT:

Yale University Press.

Hovland, C. I., & Weiss, W. (1951). The influence of source credibility on communication

effectiveness. Public opinion quarterly, 15(4), 635-650.

Hütter, M., Sweldens, S., Stahl, C., Unkelbach, C., & Klauer, K. C. (2012). Dissociating contingency

awareness and conditioned attitudes: evidence of contingency-unaware evaluative

conditioning. Journal of Experimental Psychology: General, 141(3), 539-557.

Insko, C. A., & Oakes, W. F. (1966). Awareness and the" conditioning" of attitudes. Journal of

Personality and Social Psychology, 4(5), 487-.496.

Kattner, F. (2012). Revisiting the relation between contingency awareness and attention: Evaluative

conditioning relies on a contingency focus. Cognition & emotion, 26(1), 166-175.

Kiesler, C. A., Collins, B. E., & Miller, N. (1969). Attitude change: A critical analysis of theoretical

approaches (pp. 133-142). New York: Wiley.

Levey, A. B., & Martin, I. (1987). Evaluative conditioning a case for hedonic transfer. In Theoretical

foundations of behavior therapy (pp. 113-131). Springer US.

Lipp, O. V., Neumann, D. L., & Mason, V. (2001). Stimulus competition in affective and relational

learning. Learning and Motivation, 32(3), 306-331.

Martin, L. L., Seta, J. J., & Crelia, R. A. (1990). Assimilation and contrast as a function of people's

willingness and ability to expend effort in forming an impression. Journal of Personality and

Social Psychology, 59(1), 27-37.

Mayo, C. W., & Crockett, W. H. (1964). Cognitive complexity and primacy-recency effects in impression

formation. The Journal of Abnormal and Social Psychology, 68(3), 335-338.

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(6),

1304-1307.

McGinley, H., & Layton, B. (1973). Conditioning of evaluative meaning without contingency or demand

awareness. The Journal of General Psychology, 89(1), 81-90.

Meiser, T. (2011). Much pain, little gain? Paradigm-specific models and methods in experimental

psychology. Perspectives on Psychological Science, 6(2), 183-191.

Minear, M., & Park, D. C. (2004). A lifespan database of adult facial stimuli. Behavior Research

Methods, Instruments, & Computers, 36(4), 630-633.

Mitchell, C. J., De Houwer, J., & Lovibond, P. F. (2009). The propositional nature of human associative

learning. Behavioral and Brain Sciences, 32, 183-198.

Moran, T., & Bar-Anan, Y. (2013). The effect of object–valence relations on automatic

evaluation. Cognition & emotion, 27(4), 743-752.


Neuberg, S. L., & Fiske, S. T. (1987). Motivational influences on impression formation: outcome

dependency, accuracy-driven attention, and individuating processes. Journal of personality and

social psychology, 53(3), 431-444.

Nosek, B. A. (2005). Moderators of the relationship between implicit and explicit evaluation. Journal of

Experimental Psychology: General, 134(4), 565-584.

Orne, M. T. (1962). On the social psychology of the psychological experiment: With particular reference

to demand characteristics and their implications. American psychologist, 17(11), 776-783.

Page, M. M. (1969). Social psychology of a classical conditioning of attitudes experiment. Journal of

Personality and Social Psychology, 11, 177-186.

Peters, K. R., & Gawronski, B. (2011). Are We Puppets on a String? Comparing the Impact of

Contingency and Validity on Implicit and Explicit Evaluations. Personality and Social

Psychology Bulletin, 37, 557-569.

Petty, R. E. Briñol, P., & DeMarree, K. G. (2007). The Meta-Cognitive Model (MCM) of attitudes:

Implications for attitude measurement, change, and strength. Social Cognition, 25(5), 657-686.

Petty, R. E., & Cacioppo, J. T. (1986). The elaboration likelihood model of persuasion. Advances in

experimental social psychology, 19, 123-205.

Petty, R. E., Tormala, Z. L., Briñol, P., & Jarvis, W. B. G. (2006). Implicit ambivalence from attitude

change: An exploration of the PAST model. Journal of Personality and Social Psychology, 90,

21–41.

Priester, J. R., Petty, R. E., & Park, K. (2007). Whence univalent ambivalence? From the anticipation of

conflicting reactions. Journal of Consumer Research, 34(1), 11-21.

Ranganath, K. A., Spellman, B. A., & Joy-Gaba, J. A. (2010). Cognitive “Category-Based Induction”

Research and Social “Persuasion” Research Are Each About What Makes Arguments Believable

A Tale of Two Literatures. Perspectives on Psychological Science, 5(2), 115-122..

Ratliff, A., & Nosek, B. A. (2010). Creating distinct implicit and explicit attitudes with an illusory

correlation paradigm. Journal of Experimental Social Psychology, 46, 721-728.

Rosenberg, S., & Sedlak, A. (1972). Structural representations of implicit personality theory. Advances in


Rosenkrantz, P. S., & Crockett, W. H. (1965). Some factors influencing the assimilation of disparate

information in impression formation. Journal of Personality and Social Psychology, 2(3), 397-

402.

Ruscher, J. B., & Hammer, E. Y. (1996). Choosing to sever or maintain association induces biased

impression formation. Journal of personality and social psychology, 70(4), 701-712.

Rydell, R. J., & McConnell, A. R. (2006). Understanding implicit and explicit attitude change: A systems

of reasoning analysis. Journal of Personality and Social Psychology, 91(6), 995–1008.

Rydell, R. J., McConnell, A. R., & Mackie, D. M. (2008). Consequences of discrepant explicit and

implicit attitudes: Cognitive dissonance and increased information processing. Journal of

Experimental Social Psychology, 44, 1526–1532.

Rydell, R. J., McConnell, A. R., Mackie, D. M., & Strain, L. M. (2006). Of two minds: Forming and

changing valence-inconsistent implicit and explicit attitudes. Psychological Science, 17(11), 954-

958.

http://psycnet.apa.org/index.cfm?fa=search.searchResults&latSearchType=a&term=Peters,%20Kurt%20R.


Sansom-Daly, U. M., & Forgas, J. P. (2010). Do blurred faces magnify priming effects? The interactive

effects of perceptual fluency and priming on impression formation. Social Cognition, 28(5), 630-

640.

Schmidt, J. R., & De Houwer, J. (2012). Contingency learning with evaluative stimuli: Testing the

generality of contingency learning in a performance paradigm. Experimental Psychology, 59,

175–182.

Schneider, D. J. (1973). Implicit personality theory: A review. Psychological Bulletin, 79(5), 294-309.

Siegel, E., Sigall, H., & Huber, D. E. (2012). The IAT is sensitive to the perceived accuracy of newly

learned associations. European Journal of Social Psychology, 42(2), 189-199.

Skowronski, J. J., & Carlston, D. E. (1989). Negativity and extremity biases in impression formation: A

review of explanations. Psychological bulletin, 105(1), 131-142.

Sriram, N., & Greenwald, A. G. (2009). The Brief Implicit Association Test. Experimental Psychology,

56, 283–294.

Staats, A. W., & Staats, C. K. (1958). Attitudes established by classical conditioning. The Journal of

Abnormal and Social Psychology, 57(1), 37-40.

Staats, A. W., Staats, C. K., & Crawford, H. L. (1962). First-order conditioning of meaning and the

parallel conditioning of a GSR. The Journal of general psychology, 67(1), 159-167.

Stagner, R., & Britton Jr, R. H. (1949). The conditioning technique applied to a public opinion

problem. The Journal of social psychology, 29(1), 103-111.

Steel, P., & König, C. J. (2006). Integrating theories of motivation. Academy of Management Review, 31,

889–913.

Strack, F., Schwarz, N., Bless, H., Kübler, A., & Wänke, M. (1993). Awareness of the influence as a

determinant of assimilation versus contrast. European journal of social psychology, 23(1), 53-62.

Sweldens, S., Corneille, O., & Yzerbyt, V. (2014). The role of awareness in attitude formation through

evaluative conditioning. Personality and Social Psychology Review, 18(2), 187-209.

Walther, E., Gawronski, B., Blank, H., & Langer, T. (2009). Changing likes and dislikes through the back

door: The US-revaluation effect. Cognition and Emotion, 23(5), 889-917.

Walther, E., Langer, T., Weil, R., & Komischke, M. (2011). Preferences surf on the currents of words:

Implicit verb causality influences evaluative conditioning. European Journal of Social

Psychology, 41(1), 17-22.

Walther, E., Weil, R., & Langer, T. (2011). Why do we like the iPhone? The role of evaluative

conditioning in attitude formation. Social and Personality Psychology Compass, 5(7), 473-486.

Wishner, J. (1960). Reanalysis of "impressions of personality". Psychological Review, 67(2), 96-112.

Whitfield, M., Jordan, C. H., (2009). Mutual Influence of implicit and explicit attitudes. Journal of


Zanna, M. P., Kiesler, C. A., & Pilkonis, P. A. (1970). Positive and negative attitudinal affect established

by classical conditioning. Journal of personality and social psychology, 14(4), 321-328.

Zanon, R., De Houwer, J., Gast, A., & Smith, C. T. (2014). When does relational information influence

evaluative conditioning?. The Quarterly Journal of Experimental Psychology, 67(11), 2105-2122.

Zanon, R., De Houwer, J., & Gast, A. (2012). Context effects in evaluative conditioning of implicit

evaluations. Learning and Motivation, 43(3), 155-165.


Footnotes

running head: evaluative conditioning and impression ...baranany/mb2015.pdf · running head:...

Documents