What Mistakes Disclose: A Process Dissociation Approach to Automatic and Controlled Processes in Social Psychology

Process dissociation is a model for separating automatic and controlled contributions to responses in a single task. Although it was developed to separate conscious and unconscious uses of memory, researchers have increasingly found the model useful to answer questions in social psychology. This article reviews process dissociation studies with a focus on social psychology. It emphasizes the model's conceptual definitions of automatic and controlled processing and how those relate to procedures for estimating them. Process dissociation is contrasted with task dissociation, in which automatic and controlled processes are identified with implicit and explicit tasks. Current trends and future directions are identified, including the use of model testing procedures to compare competing theories of how automatic and controlled processes interact.     

Re-emerging conceptual integration: commentary on Berkowitz's "on the consideration of automatic as well as controlled psychological processes in aggression"

In recent decades, researchers in various areas of psychology have challenged the claims of a single mode of information processing, and developed dual-process models of social behaviors. Although these theories differ on a number of dimensions, they all share the basic assumption that two different modes of information processing operate in making decision and copying behavior. In essence, the common distinction in these perspectives is between controlled vs. automatic, conscious vs. unconscious, and affective vs. cognitive modes of processing. The purpose of Berkowitz's article is to go beyond the notion of automatic processes in order to use classic notions of conditioning and displacement to explain aggressive behavior. I assert that an explanatory framework for psychology of aggression must be anchored not only in the new but also classic theoretical paradigms. However, progress in psychology does not rest solely on the accumulation of theoretical insights. It demands a large body of empirical facts, with attention to incongruities, discordances, and conceptual clarifications.     

Individual differences in working memory capacity and dual-process theories of the mind

Dual-process theories of the mind are ubiquitous in psychology. A central principle of these theories is that behavior is determined by the interplay of automatic and controlled processing. In this article, the authors examine individual differences in the capacity to control attention as a major contributor to differences in working memory capacity (WMC). The authors discuss the enormous implications of this individual difference for a host of dual-process theories in social, personality, cognitive, and clinical psychology. In addition, the authors propose several new areas of investigation that derive directly from applying the concept of WMC to dual-process theories of the mind.     

突发性危机情景下抢购现象的发生及其应对策略

今年1到4月份,在我国一些地区出现了几次因SARS疫情恐慌导致的抢购风潮,其特点主要表现为突发性、群发性和快速平息的短暂性。这些特点与人们的风险认知机制有关。研究表明,情绪唤醒抑制人类认知中的控制性加工部分,但对自动化加工部分起促进作用。人们在危机情景下的快速应对行为主要是自动化的,事前培训和行为习惯决定了自动化反应的策略和方式。由于口传信息的高情绪唤醒水平和操作上的方便易行,恐慌性抢购便被引发出来。     

Dissociating automatic associations: Comparing two implicit measurements of race bias

Weak correspondence across different implicit bias tasks may arise from the contribution of unique forms of automatic and controlled processes to response behavior. Here, we examined the correspondence between estimates of automatic and controlled processing derived from two sequential priming tasks with identical structure and timing designed to separately measure stereotypic (Weapons Identification Task; WIT) and evaluative (Affective Priming Task; APT) associations. Across two studies using predominantly White samples, three consistent patterns emerged in the data: (a) stereotypic bias was stronger for Black targets, whereas evaluative bias was stronger for White targets; (b) overall response accuracy bias correlated modestly across the two tasks; and (c) multinomial processing tree estimates of controlled processing corresponded much more strongly than estimates of automatic processing. These findings support models positing distinct learning and memory systems for different forms of race bias, and suggest that these differing forms contribute to estimates of automatic associations.     

Dissociating automatic and controlled processes in a memory-search task: Beyond implicit memory

Our goal in this paper was to examine the processes that give rise to action slips. Procedures used to examine implicit memory and automatic processes were found to be unsatisfactory. However, the process-dissociation procedure proved useful for examining the contribution of the automatic and controlled processes underlying performance. The procedure was used in conjunction with a Sternberg memory-search task to examine the effects of set size, response speed, and stimulus-response mapping on controlled and automatic processes. The formulation allowed us to predict accurately how subjects would perform in a varied mapping condition. Moreover, set size and response speed were found to influence the controlled search process, but to leave the automatic influences unaffected. Stimulus-response mapping, on the other hand, was found to lead to probability matching in the automatic processes; this pattern was found to remain constant across changes in set size and response speed.     

Is the moral domain unique? A social influence perspective for the study of moral cognition

The nature of the cognitive processes that give rise to moral judgment and behavior has been a central question of psychology for decades. In this paper, we suggest that an often ignored yet fruitful stream of research for informing current debates on the nature of moral cognition is social influence. We introduce what we call the “social‐moderation‐of‐process” perspective, a methodological framework for leveraging insights from social influence research to inform debates in moral psychology over the mechanisms underlying moral cognition and the moral domains in which those mechanisms operate. We demonstrate the utility of the social‐moderation‐of‐process perspective by providing a detailed example of how research on social influence in behavioral ethics can be utilized to test a research question related to a debate between two prominent theories in moral psychology. We then detail how researchers across the field of moral psychology can utilize our social‐moderation‐of‐process perspective.     

On the meaning of meaning when being mean: commentary on Berkowitz's "on the consideration of automatic as well as controlled psychological processes in aggression"

Berkowitz (this issue) makes a cogent case for his cognitive neo-associationist (CNA) model that some aggressive behaviors occur automatically, emotionally, and through conditioned association with other stimuli. He also proposes that they can occur without "processing," that is, without meaning. He contrasts his position with that of social information processing (SIP) models, which he casts as positing only controlled processing mechanisms for aggressive behavior. However, both CNA and SIP models posit automatic as well as controlled processes in aggressive behavior. Most aggressive behaviors occur through automatic processes, which are nonetheless rule governed. SIP models differ from the CNA model in asserting the essential role of meaning (often through nonconscious, automatic, and emotional processes) in mediating the link between a stimulus and an angry aggressive behavioral response.     

Separating multiple processes in implicit social cognition: the quad model of implicit task performance

The authors argue that implicit measures of social cognition do not reflect only automatic processes but rather the joint contributions of multiple, qualitatively different processes. The quadruple process model proposed and tested in the present article quantitatively disentangles the influences of 4 distinct processes on implicit task performance: the likelihood that automatic bias is activated by a stimulus; that a correct response can be determined; that automatic bias is overcome; and that, in the absence of other information, a guessing bias drives responses. The stochastic and construct validity of the model is confirmed in 5 studies. The model is shown to provide a more nuanced and detailed understanding of the interplay of multiple processes in implicit task performance, including implicit measures of attitudes, prejudice, and stereotyping.     

Moderators of the relationship between implicit and explicit evaluation

Automatic and controlled modes of evaluation sometimes provide conflicting reports of the quality of social objects. This article presents evidence for 4 moderators of the relationship between automatic (implicit) and controlled (explicit) evaluations. Implicit and explicit preferences were measured for a variety of object pairs using a large sample. The average correlation was r=.36, and 52 of the 57 object pairs showed a significant positive correlation. Results of multilevel modeling analyses suggested that (a) implicit and explicit preferences are related, (b) the relationship varies as a function of the objects assessed, and (c) at least 4 variables moderate the relationship: self-presentation, evaluative strength, dimensionality, and distinctiveness. The variables moderated implicit-explicit correspondence across individuals and accounted for much of the observed variation across content domains. The resulting model of the relationship between automatic and controlled evaluative processes is grounded in personal experience with the targets of evaluation.     

A complex dynamic systems approach to lasting positive change: The Synergistic Change Model

This paper argues that psychological and social functioning is fundamentally complex, and that this complexity is critical to understanding how lasting positive changes can be achieved. Principles from complex systems theory are integrated into an empirical positive psychological framework to propose a domain-based systems model of positive change called the Synergistic Change Model. This model proposes that enduring positive change depends on the formation of mutually supportive interactions across multiple domains of psychological and social functioning. The paper uses the model to explore three types of response to intervention – relapse, spill-over, and synergy – which have been supported in the existing positive psychology literature. Three practical intervention strategies arising from the model are outlined to inform the design of future positive interventions. The model challenges reductionistic approaches to positive change and offers several approaches to embrace the complexities of lasting positive change in future theory and research.     

How social perception can automatically influence behavior

Do we always know the reasons for our actions? Or is our behavior sometimes unknowingly and unintentionally influenced by what we have recently perceived? It has been traditionally assumed that the automatic influence of knowledge in memory is limited to people's interpretation of the world, and stops short of shaping their actual behavior. Researchers in experimental social psychology have begun to challenge this assumption by documenting how people's behaviors can be unknowingly influenced by knowledge that is incidentally activated in memory during social perception. We review findings that suggest that the social knowledge that is incidentally activated while reading words or imagining events subsequently affects participants' behaviors across a range of ostensibly unrelated domains.     

Cognitive processes in implicit attitude tasks: An experimental validation of the Trip Model

Implicit attitude tasks have become very popular in various areas of psychology. However, little is known about the cognitive processes they involve. To address this issue, we investigated the underlying processes of the Go/No‐go Association Task (GNAT), a go/no‐go variant of the well‐known Implicit Association Test (IAT). More precisely, we tested two alternative multinomial processing tree (MPT) models of GNAT performance, the Trip Model and the generalized Quad Model. Both models assume that GNAT performance is influenced not only by automatic associations but also by response biases and a controlled discrimination process. However, the two models differ with respect to an additional overcoming bias process. In contrast to the Quad Model, the Trip Model assumes that overcoming bias does not play a major role in GNAT performance. Instead, the Trip Model emphasizes the role of response biases. We report three experiments that demonstrate the validity of the Trip Model for GNAT data. Copyright © 2010 John Wiley & Sons, Ltd.     

TARGET ARTICLE: On Building a Better Process Model: It's Not Only How Many,but Which Ones and By Which Means?

The Quad Model proposes that four cognitive processes are pervasive across social behavior. I describe the model and compare it to dual- and uniprocess models. Distinguishing among the four processes can provide a more nuanced and accurate depiction of social behavior than can other models. It also allows the model to be applied to a much broader range of domains. A key problem for many dual-process models is that they confound processing style with either type of content (e.g., category vs. individuating information) or type of task (implicit vs. explicit tasks). As a result, these models have trouble assessing the joint influences of multiple processes. In contrast, the Quad Model provides estimates of the independent and simultaneous contributions of 4 key processes to social behavior.     

Automaticity in clinical psychology.

The authors provide an overview of the literature on the ability of response expectancies to elicit automatic responses in the form of self-fulfilling prophecies and link it to the broader psychological investigation of automatic processes. The authors review 3 areas of research in which response expectancies have been shown to affect experience, behavior, and physiology: placebo effects, the effects of false biofeedback on sexual arousal, and the alteration of perceptual and cognitive functions by hypnotic and nonhypnotic suggestion. Also reviewed are data suggesting that all behavior, including novel and intentional behavior, is initiated automatically. Following this review, the authors summarize some of the ways in which knowledge of response expectancy effects and other automatic processes that influence experience and behavior can enhance clinical practice.     

Simon says: Reliability and the role of working memory and attentional control in the Simon task

TheSimon effect refers to the observation that subjects identify targets (e.g., colors) faster when the irrelevant spatial location of the target corresponds to the location of the response key. Theoretical accounts of the Simon effect typically explain performance in terms of automatic and controlled processes. Furthermore, the relative contributions of automatic and controlled processes are held to change as a function of the proportion of compatible to incompatible trials (compatibility proportion). Data are presented demonstrating that the reliability of the Simon effect, indexed by correlating its magnitude within subjects across blocks of trials, varied substantially as a function of the compatibility proportion. When the compatibility proportion was high, so was reliability. When the compatibility proportion was low, reliability was low as well. The results are discussed in terms of the relative reliability of automatic and controlled processes and the role of working memory and attentional control in goal maintenance.     

Book Review

Repetition priming is one of the most robust phenomena in cognitive psychology, but participants vary substantially on the amount of priming that they produce. The current experiments assessed the reliability of repetition priming within individuals. The results suggest that observed differences in the size of the repetition priming effect across participants are largely reliable and result primarily from systematic processes. We conclude that the unreliability of semantic priming observed by Stolz, Besner, and Carr (2005) is specific to uncoordinated processes in semantic memory, and that this unreliability does not generalize to other processes in visual word recognition. We consider the implications of these results for theories of automatic and controlled processes that contribute to priming. Finally, we emphasize the importance of reliability for researchers who use similar paradigms to study individual and group differences in cognition.     

The interplay of controlled and automatic processing in the expression of spontaneously inferred traits: A PDP analysis

Inferences made about actors influence subsequent processing about those actors. Three experiments conducted in the context of spontaneous trait inference (STI) making demonstrate that such influences occur can either occur via automatic processes or via controlled processes. Results from Experiment 1 demonstrated that processing goals manipulated prior to encoding actor behavior affected the extent to which STIs automatically influenced subsequent responses but did not alter the extent to which STIs influenced those responses via controlled processes. Results from Experiment 2 showed that the extent to which STIs affected subsequent responding via the action of controlled processes were more affected by a delay between exposure to an actor behavior and the response task than the extent to which STIs affected task performance via the action of automatic processes. Finally, results from Experiment 3 showed that participants' subjective experience of awareness of their trait inferences is related to estimates of the extent to which controlled processing is involved in the production of their future responses but not to estimates of the extent to which those responses are affected by automatic processing.     

The Boulder Model in Academia: Struggling to Integrate the Science and Practice of Psychology

For many professionals, the Boulder model captures the ideals of clinical psychology. Unfortunately, it can be extremely difficult to integrate science and practice in many work settings. The present paper provides several recommendations for encouraging the scientist-practitioner model in academia. Faculty can strive to integrate teaching, research, and clinical services in weekly activities. First, it is important for the professional to retain a strong and clear focus on clinical psychology, without straying into allied fields. Second, it is essential for anyone who works in academia to learn to juggle many different activities that demand the professional’s time. Third, it is important to maintain a consistent focus on specific areas of interest in order to cultivate them into domains of expertise. Fourth, it is helpful to appreciate the synergism whereby work in one area can enhance the other domains. Fifth, it is helpful to integrate several activities into one multi-purpose task. Finally, it can be useful to view professional involvement in many different activities that could help to advance the field.     

Maintained Individual Data Distributed Likelihood Estimation (MIDDLE)

Maintained Individual Data Distributed Likelihood Estimation (MIDDLE) is a novel paradigm for research in the behavioral, social, and health sciences. The MIDDLE approach is based on the seemingly impossible idea that data can be privately maintained by participants and never revealed to researchers, while still enabling statistical models to be fit and scientific hypotheses tested. MIDDLE rests on the assumption that participant data should belong to, be controlled by, and remain in the possession of the participants themselves. Distributed likelihood estimation refers to fitting statistical models by sending an objective function and vector of parameters to each participant’s personal device (e.g., smartphone, tablet, computer), where the likelihood of that individual’s data is calculated locally. Only the likelihood value is returned to the central optimizer. The optimizer aggregates likelihood values from responding participants and chooses new vectors of parameters until the model converges. A MIDDLE study provides significantly greater privacy for participants, automatic management of opt-in and opt-out consent, lower cost for the researcher and funding institute, and faster determination of results. Furthermore, if a participant opts into several studies simultaneously and opts into data sharing, these studies automatically have access to individual-level longitudinal data linked across all studies.