Experts With A Cause: Proposal To Investigate The Possible Interactive Effect Of Political Expertise And The Dispositional "need To Evaluate" On N400 (erp) Measures Of Automatic Evaluative Processing

Experts with a cause 1 Running head: AUTOMATIC EVALUATIVE PROCESSING IN POLITICAL EXPERTS

Experts with a cause: proposal to investigate the possible interactive effect of political expertise and the dispositional need to evaluate on N400 measures of automatic evaluative processing Brian C. Woolfrey University of Minnesota

Experts with a cause 2 Abstract In 2003 Morris and colleagues conducted an ERP study of automatic evaluative processing in which they found that N400 responses during a lexical decision task were significantly attenuated when a target adjective was affectively congruent with a political prime word. Using this modified N400 attenuation paradigm as a psychophysiological measure of automatic evaluative processing, I propose to further study the relationship between political expertise, the dispositional need to evaluate, and the tendency to automatically evaluate political stimuli. Specifically, I predict to find an interactive effect between political expertise (PE) and the need to evaluate (NE) on ERP measures of automatic evaluative processing, just as Federico and colleagues (2007) found an interactive effect between PE and NE on measures of ideological constraint. If so, this would support the idea that automatic processing tendencies may be one of the mechanisms behind the ideological organization of political information in high-NE political experts.

Experts with a cause 3 Experts with a cause: proposal to investigate the possible interactive effect of political expertise and the dispositional need to evaluate on N400 measures of automatic evaluative processing Over four decades ago Philip Converse conducted a study of the French electorate in which he found considerable disparity in levels of ideological constraint between members of the mass public and a sample of political "elites" (Converse, 1964). Ideological constraint in this context refers to the general consistency of one's ideological beliefs. A person with high levels of constraint, for example, will respond to questions about a variety of political issues in an ideologically consistent manner. As expected, this consistent pattern of responding was common among the group of political elites in Converse's study. Members of the mass public, however, tended to respond in a far less ideologically consistent manner. The average correlation between subjects' stands on ideologically-related policy issues was so low, in fact, that it led to questions in some academic circles about the capacity of average voters to hold ideologically-based beliefs or even understand ideologically-based arguments (Federico, 2007). Since Converse's original study researchers have spent a great deal of time studying the relationship between political involvement and ideological constraint. Not surprisingly, the statistical correlation between the two variables has almost always been found to be strong and positive (Converse, 2000). One of the most concrete examples of this came from an extensive study of National Elections Survey (NES) data by Jennings (1992). In this study Jennings stratified the public into groups according to levels of political activity and found that the higher the group’s level of political activity, the more constrained and stable their political opinions and preferences were. Moreover, the ideological consistency among party elites was found to greatly exceeded even the most politically active stratum of the mass public.

Experts with a cause 4 Awareness of this highly asymmetric distribution of ideological thinking inevitably led researchers to wonder what it was about political involvement that seemed to facilitate the use of ideology. One of the most popular answers to this question was that it may have something to do with the amount of political information one has access to. People who are more involved in politics are naturally exposed to more political information than people who are less involved in politics. Many of them are even described by Krosnick (1990) as political "information junkies" who spend a great deal of time seeking out and reflecting on political information. As consequence, they usually have more complex and well-developed political schemas, or "organized clusters of information about political institutions, actors and abstract political ideas", from which they can draw to use as context when making political decisions (Federico, 2007, 222). In psychological literature, this is referred to as political "expertise", and can be measured fairly accurately by simple tests of relevant political-domain knowledge (Krosnick 1990). Because involvement in politics is closely associated with levels of political expertise, many researchers were not surprised to find that, like political involvement, political knowledge is also positively correlated with ideological constraint (Zaller, 1992). What's more, this relationship was found to exist regardless of a person's actual level of political involvement. Because of this, it was speculated that the acquisition of political information may be a key variable in the development of ideological attitudes. According to this conceptualization, acquiring new political information may inevitably facilitate the use of ideological thinking because complex social information is cognitively organized in terms of schemas, and ideology is the natural of schema by which political information is organized. Evidence to support this idea comes from Zaller (1992) who showed that the mere contemplation of a political issue before making a political decision was more likely to increase ideological constraint in political

Experts with a cause 5 experts than political novices. Among other things, this suggests that the development of ideological constraint may be "contingent on the complexity of political knowledge structures" (Federico, 2007). The problem with this idea, however, however, is that not all political experts are prone to think of political issues in ideological terms. Although the correlation between political expertise and ideological consistency is moderate to strong, it is by no means perfect. Some professors of philosophy, for example, boast an intimate understanding of abstract political ideas but at the same time refuse to take an ideological stand on many political issues. Although they are certainly a minority among their learned peers, low-constraint political experts do exist. As such, it does not seem very likely that the acquisition of political information alone is responsible for the facilitation of ideologically-organized thinking. Rather, it may be more likely that political expertise is a necessary but not fully sufficient condition for ideological thought. What is it then, if not political expertise alone, that facilitates the use of ideology? Recent findings by Federico and colleagues (2007) suggest that the answer to this question may have something to do with motivation. Specifically, he examined a motivational personality factor that had been shown to be moderately correlated to political activism called "need to evaluate" (NE). Originally developed by Jarvis and Petty (1996), the need to evaluate is a construct that reflects the tendency of a person to spontaneously evaluate objects or experiences in terms of good or bad. People who score high on the NE scale typically experience many "evaluative thoughts" and hold relatively strong attitudes toward a variety of objects, people and ideas (Bizer, 2004). This includes everything from public policies that directly affect a person's life to relatively remote and inconsequential things such as abstract art. In an extensive analysis of National Elections Study (NES) data, Federico found that this trait has a robust mediating effect on the relationship

Experts with a cause 6 between political expertise and attitude constraint. Political experts tended to score higher on measures of ideological constraint overall, but this effect was limited to only those experts who scored high in measures of NE. Such an interaction between NE and political expertise is highly consistent with the idea that the acquisition of political information itself is not sufficient to facilitate ideological thinking, but instead may function as a fuel of sorts that interacts with personality or external motivational factors to facilitate ideological thinking. If so, this may suggest that political experts who have a high level of evaluative motivation, such as party elites, may actually process political information in a fundamentally different way than their low NE counterparts. One distinction that has been made between high and low NE individuals in previous literature involves attitude formation styles. Specifically, recent findings suggest that high-NE individuals tend to form their attitudes in an "on-line" fashion, whereas low-NE individuals tend to form their attitudes in a more "memory-based" way. On-line attitudes, as first described by Hastie and Park (1986), supposedly result from a process in which an individual spontaneously evaluates each new piece of information and integrates each of those evaluations into an overall “working tally” of evaluations. When prompted for an opinion, this working tally is simply reported as the final attitude. Memory-based attitude formation, in contrast, involves the relative deferral of judgment until a judgment is called for and requires on-the-spot integration of immediate context and relevant information that a person is able to recall. Based on this theory, we would expect that people who form their opinions in an on-line fashion would be able to report their attitudes or evaluations more quickly upon request than people who use memory-based attitude formation strategies. Indeed, this was exactly the type of distinction that was originally used by Hastie and Park (1986) as evidence that individual differences in processing styles exist. Thus, when Jarvis and Petty (1996) found that high-NE

Experts with a cause 7 participants reported evaluations of everyday objects significantly faster than low-NE participants, they reasonably concluded that high trait NE is associated with a more on-line process of attitude formation and low trait NE is associated with a more memory-based process (as cited in Bizer, 2004). Further research by Hermans, Houwer, and Eelen (2001) suggest that these apparent differences in attitudinal formation styles may be related to differences in "automatic" processing tendencies. In this study, high-NE participants identified words to be affectively congruent with a positive or negative prime word faster than they identified words to be incongruent with the prime word. Low-NE participants, however, showed no statistical difference in the time it took to identify the congruency of the words pairs. This evidence of affective priming in response to semantic stimuli lends strong support to the idea that high-NE individuals may process affective information in a more "automatic" way than low-NE individuals. According to this line of thinking, otherwise known as the "hot cognition" hypothesis, all socio-political concepts are coded into long-term memory with a corresponding positive or negative affective "charge" that is quickly and automatically activated upon any future presentation of that same stimuli (Morris, 2003). Thus, people who tend to chronically evaluate incoming stimuli should possess particularly strong evaluative associations. Moreover, in a decision-making scenario in which multiple pieces of information are presented to a high-NE individual, each new piece of information is thought to automatically activate an entire network of evaluative associations. If two or more of these networks happen to contain the same attitudinal valence nodes (e.g. "good" or "bad"), these overlapping valences are strengthened and primed through converging activation. It is through this cognitive mechanism by which an evaluative "tally" may be kept during on-line evaluative processing. Additionally, this redundant strengthening and priming of

Experts with a cause 8 evaluative judgments may offer a plausible explanation as to why Lodge and Taber (2000) found that affective true/false judgments were made about twice as fast as more "cognitive" true/false judgments regarding a hypothetical congressman (as cited in Morris, 2003), and why Bizer (2006) found that attitudes created through on-line processing were stronger and more resilient than attitudes created through memory-based processing. More importantly, this model may have considerable explanatory power regarding Federico's (2007) findings of a possible interactive effect of NE and political sophistication on ideologically-organized thinking. The more political information high-NE individuals are exposed to, the stronger and more redundant the evaluative networks of these people would theoretically become. Eventually, it may be the case that new political information starts to become cognitively organized in terms of common evaluations rather than simply being organized in terms of "cold" nomological networks (see Figure 1 for a crude hypothetical example). Such a model may provide a reasonable explanation as to why high-NE political experts are more ideologically constrained in their thinking compared to low-NE experts and political novices. Simple behavioral evidence, however, has never been considered sufficient grounds on which to base an entire complex model of information-processing (Stanovich, 2007). In the domain of cognitive psychology, such models have only been considered valid if they are supported by converging evidence from multiple domains of measurements such as self-report data, behavioral-performance measurements, neuropsychological dissociations and, most notably, psychophysiological measurements of brain function. The latter has been proven to be particularly useful in helping to identify distinct neural processes and subsystems involved in information processing. Recently, the tools of cognitive neuroscience have begun to be used as

Experts with a cause 9 converging evidence by political psychologists as well. Although controversial in some circles due to the misuse or oversimplification of psychophysiological evidence, the psychophysiological approach has been particularly successful in the past at identifying automatic cognitive-political processes that have previously only been inferred by behavioral observations (Weston, 2007). With this in mind, Morris and colleagues (2003) took the first step in using psychophysiological methods to provide converging evidence for the "hot cognition" hypothesis of political information processing. In this study, event-related-potential (ERP) recordings of brain activity were interpreted as evidence that participants evaluated political stimuli in a relatively automatic way and that these automatic evaluations affected later processing of political information. To do this, they measured changes in a negatively-deflected ERP wave that occurs about 400 milliseconds after the onset of stimulus presentation. In the domain of cognitive neuroscience, the amplitude of this "N400" wave has been established as a fairly reliable measure of semantic expectancy. When a word is presented to a participant, semantically unexpected words typically produce a greater N400 wave amplitude (greater peak negativity) than semantically primed words (Kutas & Hillyard, 1980, as cited in Morris, 2003). This difference has been interpreted by many to reflect underlying levels of neural processing activity. When a word is semantically primed through inference or the activation of a nomological network, the node that the word occupies in a nomological network is thought to become activated in such a way that neural pathways from this node from long-term memory to working memory are put into a state of "readiness" and thus take relatively less bottom-up processing to access (Morris 2003). It is this type of facilitation that is thought to be reflected in the relative attenuation of the N400 wave during semantic priming or inference tasks.

Experts with a cause 10 Morris and colleagues (2003) adapted this well-established cognitive neuroscience paradigm to test whether affective incongruity in the domain of political information would also produce changes in the N400 wave. In this experiment participants were presented with a political object prime word and then a moment later were presented with an affective target adjective that was either affectively congruent or affectively incongruent with the prime word (depending on evaluative pre-testing and careful selection of these pairs for each participant). When the second word was presented, subjects were told to make a judgment as to whether the target adjective was positive or negative in valence. As predicted, it was found that affectively incongruent political object prime / affective target pairs elicited an enhanced N400 negativity relative to affectively congruent prime/target pairs. This is important because it provides psychophysiological evidence of a "priming" effect of political stimuli on evaluative adjectives and thus supports the hot cognition theory that affective evaluations are stored along with semantic political information. As already discussed, automatic processing styles have already been linked to the need to evaluate (NE) trait through behavioral evidence. Because of this I hypothesize that the difference in N400 peak negativity between affectively congruent and affectively incongruent political object - affective target pairs, or affective facilitation scores, will be greater in high-NE individuals than in low-NE individuals. Also, because the evaluative networks of political experts with high-NE should be even stronger based on the theoretical model discussed earlier, I predict to find an interaction between NE and PE such that political experts with high-NE show a far greater difference in affective facilitation than any other group. If so, this may lend preliminary support to the idea that automatic evaluative processing styles may be at least one of

Experts with a cause 11 the underlying factors behind the facilitation of ideologically-organized thinking in high-NE political experts. Method Participants For this study I propose to recruit 40 undergraduate students from the University of Minnesota who will be compensated with their choice of either cash or standardized extra-credit points for an introductory psychology course. To ensure that some students are more politically knowledgeable than others, half of the participants will be recruited from the political science department. All participants will be native English speakers with normal or corrected-to-normal vision and approximately half of them should be female. Measurements There will be several between-groups factors in this experiment. The first is the dispositional need to evaluate, as operationalized by scores on Jarvis & Petty's (1996) original Need To Evaluate self-report questionnaire. Participants who score in the top half of this measure will be considered high-NE and participants in the lower half will be considered low-NE. The second important between-groups factor will be political expertise, as operationalized by a standard multi-dimensional test of political knowledge. As with the NE construct, participants will be divided into groups of high and low political expertise. This will essentially create 4 distinct groups of 10 participants each, including 1) low-NE / political novice 2) high-NE / political novice 3) low-NE / political expert and 4) high-NE / political expert. The last betweengroups factor will be each group's average evoked N400 potentials as measured by standard EEG electrodes placed at three midline scalp locations: FZ, CZ, and PZ (according to the International 10–20 system). Separate N400 averages for congruent prime/target pairs and incongruent

Experts with a cause 12 prime/target pairs will then be calculated for each group and used in the final analysis. Also, because previous behavioral studies have found that there is no statistical difference in affective facilitation between positive and negative stimuli, these differences will be ignored and both positive and negative stimuli will be collapsed into the congruent and incongruent groups. Lastly, a vertical electro-oculogram (EOG) electrode will be used to detect and discard trials with ocular artifacts. Stimuli and Procedure Except for added measures of NE and political expertise, the paradigm used in this experiment will be virtually identical to that used by Morris (2003). Accordingly, participants will come in two days before their scheduled experimental session for pre-testing. During this phase of the experiment participants will complete a computerized "prime-selection" task in which they indicate their attitudes (positive or negative) toward 36 different political "attitude objects" as quickly as possible. Based on decision latency measures, the 5 strongest positive and 5 strongest negative attitude objects will then be selected as the prime words for that individual in the following experimental session. Also during this pre-test session, the NE self-report questionnaire and the test of political knowledge will be administered to participants. During the actual experimental session, each participant will sit in a comfortable chair and all political prime and target words will be presented in 36-point red font against a solid gray background on a high-definition, high-contrast computer screen situated approximately 90 centimeters from the participants' eyes. All florescent lights will be turned off to reduce 60hz EEG noise. To ensure that participants attend to the prime, they will told that the prime words that are presented on the screen are "memory" words that may be used in a memory test later in the study. As in Morris' (2003) study, these prime words will be presented for 150ms each,

Experts with a cause 13 followed by a delay of 100ms before the target word is presented. The target word will remain on the screen for 1 second and then be followed immediately by an "R" to indicate that the participant must make a judgment as to whether the target word was positive or negative in valence. These target words will be from the same list of 15 "clearly positive" and "clearly negative" adjectives used by Morris (2003). Each target word will appear a total of two times in a randomized order for each participant and be preceded by either a positive or negative prime. Thus, each prime word will appear a total of six times and be paired three times with a negative target and three times with a positive target. Results Once all measurements are taken and the N400 components are averaged within each group for both congruent and incongruent prime-target pairs, the differences in average amplitudes between the congruent and incongruent pairs will then be calculated for each group. Based on Morris' (2003) study, these difference scores should reflect the degree of affective priming or facilitation that took place in each group and thus provide a dependant measure of automatic evaluative processing. For statistical analysis, data will be entered into a 2 (NE) X 2 (PE) between-groups analysis of variance with the "affective facilitation" score serving as the dependant measure. If my first hypothesis is true, we should find a main effect for both the need to evaluate and political expertise such that the high-NE and high-expertise groups show greater affective facilitation than the low-NE and low-expertise groups. If my second hypothesis is true, post-hoc analysis will reveal that there is an interactive effect between NE and expertise such that only the high-NE / high-expertise group will show statistically significant affective facilitation scores. No other difference should be significant (See Figure 2 for hypothetical representation).

Experts with a cause 14 Discussion If a main effect of NE was found using Morris' N400 attenuation paradigm as a dependent measure of automatic evaluative priming, this would be consistent with previous behavioral findings that NE is related to automatic processing, or "hot cognition". A more important finding would be the possible interactive effect of NE and political expertise on N400 measures of automatic evaluative processing. This evidence would lend support to the model of overlapping and redundant affective networks in high-NE political experts and, when compared to Federico's (2007) findings of an interactive effect between NE and political expertise on measures of ideological constraint, would suggest the possibility that automatic processing tendencies may be one of the mechanisms behind the apparent ideological organization of information in high-NE political experts. If so, this may have far-reaching implications on how we understand political expertise and motivated decision-making. For this study to be interpreted properly, however, several limitations must first be addressed. First, the variables in this proposed study are all pre-existing factors. Pseudoexperiments like this can only give us information about associations between variables. Even when measures are taken from multiple domains (such as behavioral and psychophysiological), we can still only speculate about underling causal relationships between these variables. There are also other limitations. Morris' (2003) study only provided preliminary support for the hypothesis that N400 attenuation was a valid measure of affective priming. As he cautioned, it could be that the N400 wave reflects or includes some other processes that we do not yet understand. Also, the procedure of dichotomizing continuous variables such as political expertise and NE may not be the best way to analyze an interaction because of the problems it creates with regard to statistical power. A similar problem may also arise out of the fact that the proposed

Experts with a cause 15 sample is limited to college students who are unlikely to represent the entire domain of political expertise. A potential solution to both of these problems might be to recruit additional participants from a group of political party elites and then analyze the data in a more parametrically-oriented way. Despite these few weaknesses, the proposed study would take an important step toward clarifying the role of evaluative motivation and political expertise with regard to political decision-making, and would likely encourage new studies to explore the "black box" of attitude cognition in more detail. Future studies would most likely include functional MRI investigations into neural differences in political information processing styles. Additionally, evidence of distinctive processing styles may inspire investigation into the merits and weaknesses of each in relation to real-world political situations. Studies have shown that that although most political experts are better at remembering political information than political novices, they are only slightly better than chance when it comes to predicting actual political and social events (Tetlock, 2005). Could this be because the majority of political experts are influenced more heavily by automatic processing and therefore ignore important contextual information in favor of programmed ideological schemas? Or is this effect pushed by memory-based thinkers who may evaluate political stimuli less efficiently than on-line thinkers? Any future study on this topic would be fascinating and would undoubtedly have a profound effect on the way we think about political expertise and motivation.

Experts with a cause 16 References Bizer, G. Y., Krosnick, J.A., Holbrook, A.L., Wheeler, S.C., Rucker, D.D., Petty, R.E. (2004). The impact of personality on cognitive, behavioral, and affective political processes: The effects of need to evaluate. Journal of Personality, 75, 995-1028. Bizer, G. Y., Tormala, Z.L., Rucker, D.D., Petty, R.E. (2006). Memory-based versus on-line processing: Implications for attitude strength. Journal of Experimental Social Psychology, 42, 646–653. Converse, P. E. (1964). The Nature of Belief Systems in Mass Publics. In D. Apter (Ed.), Ideology and Discontent, New York: Free Press. Converse, P. E. (2000). Assessing the Capacity of Mass Electorates. Annual Review of Political Science, 3, 331-353 Federico, C. M., Schneider, M.C. (2007) Political expertise and the use of ideology: Moderating effects of evaluative motivation. Public Opinion Quarterly, 71, 221-252 Hastie, R., Park, B. (1986). The relationship between memory and judgment depends on whether the judgment task is memory-based or on-line. Psychological Review, 93, 258–268. Hermans, D., Houwer, J., Eelen, P. (2001). A time analysis of the affective priming effect. Cognition and Emotion, 15, 143–165. Jarvis, W. B. G., Petty, R. E. (1996). The need to evaluate. Journal of Personality and Social Psychology, 70, 72–194. Jennings, M. K. (1992). Ideological thinking among mass publics and political elites. Public Opinion Quarterly, 56, 420-440 Krosnick, J. A., (1990). Expertise and political psychology. Social Cognition, 8, 1-8

Experts with a cause 17 Morris, J. P., Squires, N.K., Taber, C. S., Lodge, M. (2003) Activation of Political Attitudes: A Psychophysiological Examination of the Hot Cognition Hypothesis. Political Psychology, 24, 727-745. Stanovich, K. E. (2007). How to think straight about psychology. Boston: Allyn and Bacon. Tetlock, P. E. (2005). Expert Political Judgment: How Good Is It? How Can We Know? New Jersey: Princeton University Press Weston, D. (2007). The political brain: The role of emotion in deciding the fate of the nation. Cambridge: Public Affairs Zaller, J. (1992). The Nature and Origins of Mass Opinion. New York: Cambridge University Press.

Experts with a cause 18 Figure Caption Figure 1. Hypothetical conceptualizations of "cold" (top) vs. "ideological" (bottom) cognitive organization of political information.

Figure 1

-

+

+

-

Political Party

War + -

Democrat

Republican

+

+

-

Death Penalty +

-

Obama

Gun Control -

+

High Taxes +

-

Political Party Democrat

War

Republican -

+

Obama

Bush Gun Control Death Penalty High Taxes

Figure Caption Figure 2. Hypothetical results showing the possible interactive relationship between political expertise and the need to evaluate (NE) on N400 affective facilitation (N4AF) scores.

Figure 2

High-NE N400 Affective Facilitation Score (N4AF) Low-NE

Political Novices

Political Experts