How do we know that we know? The accessibility model of the feeling of knowing

Even when Ss fail to recall a solicited target, they can provide feeling-of-knowing (FOK) judgments about its availability in memory. Most previous studies addressed the question of FOK accuracy, only a few examined how FOK itself is determined, and none asked how the processes assumed to underlie FOK also account for its accuracy. The present work examined all 3 questions within a unified model, with the aim of demystifying the FOK phenomenon. The model postulates that the computation of FOK is parasitic on the processes involved in attempting to retrieve the target, relying on the accessibility of pertinent information. It specifies the links between memory strength, accessibility of correct and incorrect information about the target, FOK judgments, and recognition memory. Evidence from 3 experiments is presented. The results challenge the view that FOK is based on a direct, privileged access to an internal monitor.     

Belief and feeling: Evidence for an accessibility model of emotional self-report.

This review organizes a variety of phenomena related to emotional self-report. In doing so, the authors offer an accessibility model that specifies the types of factors that contribute to emotional self-reports under different reporting conditions. One important distinction is between emotion, which is episodic, experiential, and contextual, and beliefs about emotion, which are semantic, conceptual, and decontextualized. This distinction is important in understanding the discrepancies that often occur when people are asked to report on feelings they are currently experiencing versus those that they are not currently experiencing. The accessibility model provides an organizing framework for understanding self-reports of emotion and suggests some new directions for research. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Developmental aspects of the feeling of knowing.

Prior reports that feeling-of-knowing (FOK) accuracy increases through childhood and remains constant during adulthood are inconclusive because (a) they used absolute FOK judgments that can be contaminated by differences in the threshold for claiming to know, (b) they used measures of FOK accuracy that are not invariant across different degrees of recognition-test difficulty, and (c) they did not rule out the possibility that age-related differences in FOK accuracy may be caused by changes in the reliability of FOK judgments. We avoided these methodological problems in two studies by using relative FOK judgments, by computing Goodman-Kruskal gamma coefficients to assess FOK accuracy, and by assessing the test–retest reliability of the subjects' FOK judgments. We found that 6-year-olds had less reliable FOK judgments but greater FOK accuracy than 10-year-olds or 18-year-olds. Moreover, 18- and 70-year-olds had equally reliable FOK judgments and equivalent FOK accuracy. Possible reasons for the greater FOK accuracy of youngsters included their lower rate of commission errors and more frequent recognition of correct answers for their commissions. Implications are drawn for the study of metacognitive development. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

A comparison of three predictors of an individual's memory performance: The individual's feeling of knowing versus the normative feeling of knowing versus base-rate item difficulty.

Compared the memory performance predictions of 168 undergraduates with normative item-difficulty predictions and with normative feeling-of-knowing (FOK) predictions. Ss attempted to recall the answers to general-information questions, made FOK judgments on nonrecalled items, and were subsequently administered a criterion test (relearning, perceptual identification, or 1 of 2 versions of recognition). Results indicate that, for predicting an S's criterion performance, the S's own FOK predictions were intermediate between 2 kinds of normative predictions: Ss' FOK predictions were more accurate than predictions derived from normative FOK ratings but were less accurate than predictions derived from base-rate item difficulty (normative probabilities of correct recall). Subsidiary analyses showed that factors other than unreliability were responsible for the partial inaccuracy of Ss' FOK. Possible ways to improve the accuracy of an individual's FOK predictions are discussed. (32 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The feeling of knowing experience in educable retarded children.

Required 50 educable retarded children to predict their recognition accuracy when recall failed. Ss were divided into 3 ability levels: (a) old Ss (mean CA 13.25 yrs, mean MA 10.5 yrs), (b) medium Ss (mean CA 11.25 yrs, mean MA 8.75 yrs), and (c) young Ss (mean CA 9 yrs 5 mo, mean MA 6 yrs 9 mo). Results indicate that old and medium Ss could reliably predict their recognition accuracy, suggesting sensitivity to their own feeling of knowing experience, but that the young Ss showed no evidence of this sensitivity. Even though the young Ss had difficulty predicting their recognition accuracy in advance they were able to estimate the success or failure of their responses after they had occurred. Results are discussed in terms of the complexity of the metamemory judgment required, and the advisability of evaluating metamemorial knowledge across several situations is emphasized. (21 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Confidence level and feeling of knowing in question answering: The weight of inferential processes.

What are the sources of metamemory judgments in question answering? Exp 1 showed that the level of confidence in the correctness of a recalled answer (CL) is a reliable measure (test–retest G?=?.87) and a valid predictor of recognition (G?=?.55/.58). In contrast, the feeling of knowing the correct answer (FOK), although reliable (G?=?.82), is a poor predictor of recognition (G?=?.11). In Exp 2 it was discovered that FOK levels are more highly correlated with beliefs about what should be known than are CL judgments. Controlling for the relationship between FOK and those beliefs eliminates any correlation between FOK and recognition (G drops from .27 to .08), which is not the case with CL (G remains as high as .52). These data suggest that CL is based on the retrieval of at least some elements of the right answer, whereas FOK may be derived from recency or familiarity effects elicited by the question itself. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The combined contributions of the cue-familiarity and accessibility heuristics to feelings of knowing.

A model for the basis of feeling of knowing (FOK) is proposed, which combines 2 apparently competing accounts, cue familiarity (L. M. Reder, see record 1987-12470-001) and accessibility (A. Koriat, see record 1994-04361-001). Both cue familiarity and accessibility are assumed to contribute asynchronously to FOK, but whereas the effects of familiarity occur early, those of accessibility occur later and only when cue familiarity is sufficiently high to drive the interrogation of memory for potential answers. General information questions were used to orthogonally manipulate cue familiarity and accessibility. As expected, both familiarity and accessibility enhanced FOK judgments, but the effects of accessibility were found mostly when familiarity was high. This interactive pattern was replicated when FOK judgments were delayed but not when they were immediate. The results support the proposed cascaded model of FOK but also imply a differentiation between 2 variants of the accessibility heuristic. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The enhancement effect in letter detection: Further evidence for the structural model of reading.

According to the structural model of reading (A. Koriat and S. N. Greenberg; 1994), the extraction of structure leads the way to the analysis of meaning. Consistent with this model, previous letter-detection studies have documented an inordinately high rate of letter omissions in function morphemes, suggesting that the cognitive representation of function morphemes is diminished once they have been utilized to set phrase structure. The present study revealed a new and complementary enhancement effect: Letter detection in content morphemes that immediately followed functors was superior to that of content morphemes positioned elsewhere in the text. Together these effects suggest an on-line figure–ground representation of text in which structural elements recede as semantic elements are pushed to the foreground. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Foundations for a model of knowing: I. Constructing reality.

Contends that traditional views of knowledge are being challenged. An emerging "constructivist" perspective, as proposed by G. Kelly (1955, 1967), an engineer turned clinician, suggests that to a large degree individuals construct reality. In his "constructive alternativism" Kelly assumes that people validate their hypotheses and beliefs through subjectively construed goodness-of-fit criteria applied to perceived differences between anticipations and feedback. His model of construing is compatible with those emerging in the history and philosophy of science and in cognitive psychology. Nevertheless, constructivists must answer a perplexing question: How can fallible knowledge, constructed as it is from abstracted and incomplete representations of objects and events, capture and maintain one's confidence, as it does, and furthermore prove highly functional, as it does? (French abstract) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Further evidence for mixed emotions.

Emotion theorists have long debated whether valence, which ranges from pleasant to unpleasant states, is an irreducible aspect of the experience of emotion or whether positivity and negativity are separable in experience. If valence is irreducible, it follows that people cannot feel happy and sad at the same time. Conversely, if positivity and negativity are separable, people may be able to experience such mixed emotions. The authors tested several alternative interpretations for prior evidence that happiness and sadness can co-occur in bittersweet situations (i.e., those containing both pleasant and unpleasant aspects). One possibility is that subjects who reported mixed emotions merely vacillated between happiness and sadness. The authors tested this hypothesis in Studies 1–3 by asking subjects to complete online continuous measures of happiness and sadness. Subjects reported more simultaneously mixed emotions during a bittersweet film clip than during a control clip. Another possibility is that subjects in earlier studies reported mixed emotions only because they were explicitly asked whether they felt happy and sad. The authors tested this hypothesis in Studies 4–6 with open-ended measures of emotion. Subjects were more likely to report mixed emotions after the bittersweet clip than the control clip. Both patterns occurred even when subjects were told that they were not expected to report mixed emotions (Studies 2 and 5) and among subjects who did not previously believe that people could simultaneously feel happy and sad (Studies 3 and 6). These results provide further evidence that positivity and negativity are separable in experience. (PsycINFO Database Record (c) 2011 APA, all rights reserved)     

Confidence level and feeling of knowing for episodic and semantic memory: an investigation of lorazepam effects on metamemory

The effects of lorazepam (0.026 or 0.038 mg/kg), a benzodiazepine, and of a placebo on metamemory, i.e. knowledge about one's own memory capabilities, were investigated in 36 healthy volunteers. Accuracy of confidence levels (CL) in the correctness of recalled answers and accuracy of feeling of knowing (FOK) the answers when recall fails were measured using a sentence memory task assessing episodic memory and a task consisting of general information questions and assessing semantic memory. Lorazepam impaired episodic memory. Unexpectedly, it also impaired performance in both the recall and recognition phases of the task assessing semantic memory, suggesting that it decreased the ability to distinguish between correct and incorrect information. In episodic memory, lorazepam 0.038 mg/kg-treated subjects exhibited an impaired CL accuracy, compared to placebo-treated subjects, and their FOK accuracy was at chance. In semantic memory, their overall CL and FOK accuracy was apparently spared. However, these subjects selectively overestimated their CL judgements for incorrect answers; moreover, secondary analyses showed that FOK accuracy for a subset of low-accuracy items was virtually nil. These results suggest that lorazepam impairs metamemory for both episodic and semantic memory.     

Manipulating the shape of distance effects in visual curve tracing: Further evidence for the zoom lens model.

Reexamined the predictive power of the zoom lens curve tracing model in the research of P. A. McCormick and P. Jolicoeur (see record 1992-07800-001) by selectively manipulating distractor proximity in 4 conditions and regressing the predictions of the model on human response time. 12 university students participated. The model was compared with 2 other curve tracing models: pixel by pixel tracing and the bipartite receptive field operator of Jolicoeur et al (see record 1992-15217-001). Results provide strong support for McCormick and Jolicoeur's visual attention based zoom lens operator. (French abstract) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Foundations for a model of knowing: II. Fallible but functional knowledge.

Argues that an evolving theory known as "constructivism" challenges the traditional view of how individuals generate and revise knowledge. Constructivism helps address a major issue raised by modern scholars of the history and philosophy of science and decision theory. The question relates to how to reduce the search and solution space of complex and changing environments to "mind size" (i.e., to fit person's limited memory and computational capacity). One emerging answer is that individuals rely heavily on robust presuppositions and simplified representations of environmental structure. Two issues are addressed: Under what conditions knowledge can be highly fallible and at the same time be highly functional and whether one can make a plausible case, within this constructivist frame of reference, for realism, for knowledge that approximates "reality." (French abstract) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

What determines initial feeling of knowing? Familiarity with question terms, not with the answer.

How do people know whether they have an answer to a question before they actually find it in their memory? Two experiments explored this question. Ss were trained on relatively novel 2-digit?×?2-digit arithmetic problems (e.g., 23?×?27). Before answering each problem, Ss made a quick feeling of knowing judgment as to whether they could directly retrieve the answer from memory or had to compute it. Knowing the answer initially appeared to be linearly related to having a feeling of knowing the answer; however, when the frequency of exposure to complete problems and the frequency of exposure to parts of the problems were separately varied, feeling of knowing was better predicted by the frequency of presentation of the problem parts, not by knowledge of the answer. This suggests that the processes involved in knowing the answer are different from those involved in having a feeling of knowing. Specifically, an early feeling of knowing is not just based on an early read of the answer. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

On the relation between feeling of knowing and lexical decision: Persistent subthreshold activation or topic familiarity?

Exp 1 replicated I. Yaniv and D. E. Meyer's (1987) finding that lexical decision and episodic recognition performance was better for words previously yielding high-accessibility levels (a combination of feeling-of-knowing and tip-of-the-tongue ratings) in comparison with those yielding low-accessibility levels in a rare word definition task. Exp 2 yielded the same pattern even though lexical decisions preceded accessibility estimates by a full week. Exp 3 dismissed the possibility that the Exp 2 results may have been due to a long-term influence from the lexical decision task to the rare word judgment task. These results support a model in which Ss (1) retrieve topic familiarity information in making accessibility estimates in the rare word definition task and (2) use this information to modulate lexical decision performance. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Individual differences in the development of sensation seeking and impulsivity during adolescence: Further evidence for a dual systems model.

Consistent with social neuroscience perspectives on adolescent development, previous cross-sectional research has found diverging mean age-related trends for sensation seeking and impulsivity during adolescence. The present study uses longitudinal data on 7,640 youth from the National Longitudinal Study of Youth Children and Young Adults, a nationally representative sample assessed biennially from 1994 to 2006. Latent growth curve models were used to investigate mean age-related changes in self-reports of impulsivity and sensation seeking from ages 12 to 24 years, as well individual differences in these changes. Three novel findings are reported. First, impulsivity and sensation seeking showed diverging patterns of longitudinal change at the population level. Second, there was substantial person-to-person variation in the magnitudes of developmental change in both impulsivity and sensation seeking, with some teenagers showing rapid changes as they matured and others maintaining relatively constant levels with age. Finally, the correlation between age-related changes in impulsivity and sensation seeking was modest and not significant. Together, these results constitute the first support for the dual systems model of adolescent development to derive from longitudinal behavioral data. (PsycINFO Database Record (c) 2011 APA, all rights reserved)     

Internship hours revisited: Further evidence for a national standard.

Inconsistent training requirements across predoctoral internships have problematic implications for psychologists seeking licensure and for the profession as a whole. This investigation is an 11-year follow-up to a previous study (Bartle & Rodolfa, 1999) that explored the variation in hours accrued across predoctoral internship programs. The current study examined the total number of supervised training hours provided by internships and the methods utilized by training directors to calculate these hours. The total number of supervised hours reported by training directors ranged from 1,840 to 2,080, with 58.2% calculating hours by approximation, rounding off, or estimation. As training directors reported more total hours, they were increasingly likely to include holiday, vacation, and sick leave. Recommendations for state licensing boards and internship training programs are provided, including a national licensure requirement of 1,800 hours accrued during internship. (PsycINFO Database Record (c) 2011 APA, all rights reserved)     

Further evidence for a Jamesian model of self-worth: Reply to Marsh (1995).

A critical analysis is offered of H. W. Marsh's (1993, 1995) and B. W. Pelham's (see record 1996-14327-001) recent studies of global self-esteem. It is concluded that a strong empirical case has now been made for the validity of W. James's (1890) hypothesis about self-investment and self-esteem. The personal importance people place on their specific strengths and weaknesses is an important predictor of global self-esteem. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Supervised relaxation training: A model for greater accessibility of behavioral interventions.

Developed a method of relaxation training—supervised relaxation training—to address (1) problems of cost and availability, which limit the applicability of such training to relatively few individuals, and (2) the inherent limitations of self-help programs. Although this training method required less professional involvement than other methods, it was not intended as a self-help approach to relaxation. Using 55 volunteers (aged 19–67 yrs), 2 treatment groups and 1 control group were formed: One treatment group consisted of self-selected Ss from the community at large; the other represented Ss from a specific work site in the community. The procedure consisted of the use of a self-study manual in conjunction with professional training. Outcome was determined on the basis of changes in blood pressure, scores on the State-Trait Anxiety Inventory, and emotional indicators on human figure drawings. MANOVA indicated a positive effect for both treatment groups. Potential applications of the program are discussed, with recommendations for the use of this procedure as a model for other skill-developing interventions. (29 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)