Search

Literature Review: Anhedonia, the loss of interest or pleasure in usually pleasurable activities, is a core symptom of depression and is associated with a reduction in positive affect (PA). Repetitive negative thought (RNT) is implicated in the development and maintenance of psychiatric disorders. It has been hypothesised that RNT causally contributes to anhedonia. The aim of this review was to explore this relationship to answer two questions: Is there a relationship between RNT and anhedonia? Does RNT causally contribute to anhedonia? Review inclusion criteria were: studies using standardised measures to report a relationship between RNT and anhedonia or reduced PA. Results suggest that cross-sectional and longitudinal studies identify a relationship between RNT and anhedonia. Preliminary evidence from experimental studies shows that RNT causally contributes to anhedonia. Limitations within the field are that anhedonia is rarely measured directly or behaviourally. Future research is warranted to explore the relationship between RNT and anhedonia with a particular focus on direct and behavioural measures of anhedonia. Empirical Paper: It is hypothesised that repetitive negative thought (RNT) causally contributes to anhedonia. There is cross-sectional and longitudinal evidence of this relationship, but it has not previously been investigated directly using experimental methods. In the present study, student participants were randomly assigned to an unresolved goal (RNT) manipulation (n = 43) or resolved goal (control) manipulation (n =41) prior to completing a reward sensitivity task. This task has been reliably found to train a response bias towards the stimuli that is differentially positively reinforced, with both depression and self-reported anhedonia associated with a reduced response bias. The unresolved goal versus resolved goal manipulation was effective, with the unresolved condition producing significantly higher levels of RNT during the reward sensitivity task relative to the resolved condition. Inconsistent with study predictions, there was no significant difference between the conditions on response bias, although there were trend findings, which tentatively suggest that RNT may influence anhedonia. Potential accounts for the null findings and future research are discussed.

The aim of this dissertation was to address a theoretical debate on belief bias. Belief bias is the tendency for people to be influenced by their prior beliefs when engaged in deductive reasoning. Deduction is the act of drawing necessary conclusions from premises which are meant to be assumed as true. Given that the logical validity of an argument is independent of its content, being influenced by your prior beliefs in such content is considered a bias. Traditional theories posit there are two belief bias components. Motivated reasoning is the tendency to reason better for arguments with unbelievable conclusions relative to arguments with believable conclusions. Response bias is the tendency to accept believable arguments and to reject unbelievable arguments. Dube et al. (2010) pointed out critical methodological problems that undermine evidence for traditional theories. Using signal detection theory (SDT), they found evidence for response bias only. We adopted the SDT method to compare the viability of the traditional and the response bias accounts. In Chapter 1 the relevant literature is reviewed. In Chapter 2 four experiments which employed a novel SDT-based forced choice reasoning method are presented, showing evidence compatible with motivated reasoning. In Chapter 3 four experiments which used the receiver operating characteristic (ROC) method are presented. Crucially, cognitive ability turned out to be linked to motivated reasoning. In Chapter 4 three experiments are presented in which we investigated the impact of cognitive ability and analytic cognitive style on belief bias, concluding that cognitive style mediated the effects of cognitive ability on motivated reasoning. In Chapter 5 we discuss our findings in light of a novel individual differences account of belief bias. We conclude that using the appropriate measurement method and taking individual differences into account are two key elements to furthering our understanding of belief bias, human reasoning, and cognitive psychology in general.

It is critical for researchers to be confident of the quality of survey data. Problems with data quality often relate to measurement method design, through choices made by researchers in their creation of standardised measurement instruments. This is known to affect the way respondents interpret and respond to these instruments, and can result in substantial measurement error. Current methods for removing measurement error are post-hoc and have been shown to be problematic. This research proposes that innovations can be made through the creation of measurement methods that take respondents' individual cognitions into consideration, to reduce measurement error in survey data. Specifically, the aim of the study was to develop and test a measurement instrument capable of having respondents individualise their own rating-scales. A mixed methodology was employed. The qualitative phase provided insights that led to the development of the Individualised Rating-Scale Procedure (IRSP). This electronic measurement method was then tested in a large multi-group experimental study, where its measurement properties were compared to those of Likert-Type Rating-Scales (LTRSs). The survey included pre-validated psychometric constructs which provided a baseline for comparing the methods, as well as to explore whether certain individual characteristics are linked to respondent choices. Structural equation modelling was used to analyse the survey data. Whilst no strong associations were found between individual characteristics and respondent choices, the results demonstrated that the IRSP is reliable and valid. This study has produced a dynamic measurement instrument that accommodates individual-level differences, not addressed by typical fixed rating-scales.

In the field of neuropsychology, feigning symptoms, also known as malingering or response bias, is an important issue as a large number of assessment referrals are for individuals that may receive benefits if the assessment results suggest cognitive impairment. Therefore, it is crucial to have valid and reliable measures that detect the feigning of symptoms (Slick et al., 1999). Currently, the tests that are routinely used to detect response bias are vulnerable to coaching. The goal of this study was to validate a new response bias test aimed at being less susceptible to coaching than existing measures: The Tri-Choice Naming and Response Bias Measure (N-Tri). To this end, 400 participants were assigned to either the coached malingerers’ group, uncoached malingerers’ group, or control group. Participants completed an online survey consisting of the N-Tri, Reliable Digit Span (RDS), and Portland Digit Recognition Test (PDRT). ROC curves demonstrated that all three tests were able to detect coached malingerers from controls, but the N-Tri had higher sensitivity than the RDS and PDRT. Thus, the N-Tri was able to better detect coached malingerers than compared to the RDS and PDRT. Several additional hypotheses were examined.

The first paper argues that a contextualist strategy cannot save the popular thesis that two objects are perceptually indiscriminable (in terms of discriminatory response or introspection) given a certain subject and a certain context if and only if they look the same to that subject in that context from every version of phenomenal sorites paradox. It presents an explicitly valid phenomenal soritical argument, and argues that, given the limited sensitivity of human color vision, that popular thesis is committed to its apparent soundness with a false conclusion, and a contextualist strategy cannot undermine its apparent soundness. The second paper argues that the popular philosophical idealization that whether two objects look the same or different does not vary from occasion to occasion given the same subject and the same context as controlled in a real experiment leads to a dilemma concerning the limited sensitivity of human color vision. In particular, it argues that, assuming that idealization, both a rejection and an endorsement of the maximally plausible version (in terms of the standard fine color discrimination test in psychology) of the popular thesis featured in the first paper imply unlimited sensitivity of human color vision. The third paper proposes a statistical-physiological resolution to the phenomenal sorites paradox in two steps. In the first step, it argues that there is no solid justification for any account of the condition of two objects looking the same/different in terms of overt response or introspection, following the on-going statistical criticism in psychology of mapping overt response onto internal state. In the second step, it proposes an account of what makes two objects look the same/different in terms of the relevant physiological mechanisms, which is informed by a characteristic empirical project of approaching mechanical models of color appearances, and argues that that proposal would not make it appear empirically plausible for any phenomenal soritical argument to have all true premises with a false conclusion

The goal of many laboratory studies using the signal detection paradigm is to produce a bias-free estimate of listener sensitivity (Green and Swets, 1966). Forced-choice procedures with equal a priori probability of signal occurrence, a balanced payoff matrix, and trial-by-trial feedback are designed to assess sensitivity with response bias forced to zero. Attempts to approximate real-world listening situations in controlled laboratory settings do not lend themselves to traditional listening paradigms. Listeners in real-world listening tasks rarely encounter N-interval forced-choice opportunities. Further, many real-world listening situations do not lend themselves to trial-by-trial feedback. Davis (2015) reported the effects of incomplete feedback on response bias for a simple tone-in-noise listening experiment with a single interval yes-no procedure. Feedback provided ranged from no feedback on any trial, to eight conditions with feedback for some signal–response combinations, to feedback on every trial. Davis reported a descriptive data analysis for each subject. The current study conducted an inferential data analysis with Bayesian statistics to estimate the effect of incomplete feedback on response bias in each experimental condition. The main finding is that, as expected, complete feedback drives response criteria toward the optimum, and incomplete feedback conditions result in various degrees of deviation from the optimal criterion. The results can be interpreted in terms of how feedback conveys information about a priori probabilities of different states of the world and the values and costs of correct and incorrect responses.

The goal of the current study was to examine differences in neurocognitive processes across groups marked by binge drinking and depression. Participants were recruited from undergraduate courses based on symptoms of depression and alcohol use and EEG data was obtained during an emotional Go/No-Go task. There were significant Trial Type X Emotion X Depression X Binge-Drinking interactions for N2 (F(3, 80) = 6.62, p < .01) and P3 (F(3, 80) = 4.65, p < .01) components, and a significant Emotion X Depression X Binge Drinking interaction for response bias (F (3, 65) = 3.11, p < .05). Differences in early inhibitory control were observed across emotions based on trial type among depressed non-binge drinkers, and these differences were attenuated in the presence of binge drinking. Further, the effects of depression on later inhibitory control were specific to non-binge drinkers. Lastly, depressed individuals exhibited response biases to Happy versus Angry stimuli.

In the field of neuropsychology, there is a need for reliable measures that assess for both memory and effort (response bias). A sample of college students were instructed to feign memory deficits. They were administered two well established measures of response bias, the Test of Memory Malingering (TOMM) and the Reliable Digits Span (RDS), as well as the Poreh Nonverbal Memory Test (PNMT). The study shows that all of the three measures were able to identify students who were coached to demonstrate memory deficits. A more detailed analysis showed that the TOMM and the PNMT produced higher sensitivity and specificity then the RDS. Process analysis of the PNMT showed that the ability of this measure to detect response bias improved when one analyzed the distance between the target on geometric(simple) cards of the PNMT. Namely, during the delayed recall trial of the PNMT subject who feigned memory deficits clicked on more distant stimuli (from the target) then the control group.

Due to the dearth of research examining the impact of underreporting on the validity of test scores, and the inconsistent findings from previous research regarding the utility of protocol validity scales (e.g., McGrath et al., 2010; Rohling et al., 2011), it is necessary to further investigate the utility of underreporting validity scales. The aims of the current study were three-fold: first, to investigate the impact of underreporting on MMPI-2-RF substantive scale score validity; second, to investigate the discriminant validity of the underreporting scales in differentiating various forms of simulated underreporting; and third, to examine the moderating effects of the MMPI-2-RF underreporting Validity Scale scores on substantive scale score criterion validity. This study incorporated a simulation design with a sample of 275 undergraduate college students who were all administered criterion mesures under standard instructions that assess various domains relevant to the MMPI-2-RF substantive scales. Next, participants were randomly assigned to receive either standard instructions, instructions to underreport appearing very well-adjusted, or instructions to appear very virtuous when completing the MMPI-2-RF. The results indicated that underreporting impacts substantive scale score validity on the MMPI-2-RF, and there were moderating effects of the underreporting Validity Scales (notably K-r) on associations between MMPI-2-RF Restructured Clinical Scales and external criterion measures. These findings provide evidence for the utility of underreporting validity indicators when interpreting substantive scale scores, as well as demonstrate the utility of MMPI-2-RF L-r and K-r in identifying protocols marked by underreporting.

Much of the research about college student engagement is based upon self-reported surveys, but little is known about how students formulate responses to these instruments. The purpose of this study was to specifically address this dearth of knowledge by deepening our understanding of how students’ perceptions of their environments and demographic characteristics influenced their response patterns on self-reported surveys. Bronfenbrenner’s (2005) human ecology model of development, Bourdieu and Passeron’s (1990) theory of social reproduction, and Tourangeau, Rips, and Rasinski’s (2000) four phase survey response process were used, as the theoretical framework to better understand this phenomenon. This was an explanatory sequential mixed methods study, and the participants were first-year undergraduate students at a four-year, private institution in New England. Students completed the College Student Report (CSR) as well as a series of time-use diaries, and the results of the instruments were compared using descriptive and multivariate analyses. Finally, semi-structured individual interviews were conducted, which included aspects of retrospective cognitive interviewing, with twenty-seven (27) students to understand how their experiences and response processes were shaped by their individual campus experiences and identities. Findings from this study suggest the construct validity of self-reported survey data measuring behavioral frequency patterns is questionable, as students statistically significantly under reported time spent preparing for class, engaging in co-curricular activities, commuting to campus, and relaxing and socializing. Furthermore, student characteristics such as racial/ethnic identity and satisfaction with college choice statistically significantly explained some of the variance in the reporting behaviors of students after controlling for other factors. This information coupled with the data gleaned from the semi-structured individual interviews indicate factors related to how students differentially experience the campus environment based upon their unique ecological niches affects how they respond on self-reported surveys, which means the data provided by such instrumentation is likely providing substantively different information than how it is most commonly interpreted and applied.

The effects of response bias in signal detection tasks are often intentionally minimized in order to focus on unbiased estimates of sensitivity. While reducing bias can be useful in many laboratory tasks, such diminution does little to describe real-world detection behavior. One interesting manipulator of response bias is the concept known as Knowledge of Results (KR), or more commonly known as Feedback. Feedback allows observers to not only improve their ability to detect signals but can also give information necessary for optimizing their responses to meet any one of the many decision strategies available (such as maximizing the percentage of correct responses (Green & Swets, 1966)). What is relatively unknown, however, is how a reduction in KR for individual response types (Hits, Misses, False Alarms, and Correct Rejections) can influence response bias. The proposed terminology for this type of KR is "Incomplete Knowledge of Results (IKR)". This study examined four cases of shifts in response bias due to IKR symmetry, organization, implicitness, and amount. The question of IKR symmetry examined how conditions with a single feedback type might influence listener response bias equally or unequally. The question of IKR organization was designed to uncover any relationship between 2-feedback conditions yielding KR for "yes" responses, signal trials, and correct responses. The question of implicit IKR examined whether listeners could utilize missing feedback information to optimize their response bias. Finally, the question of IKR amount examined whether an increasing quantity of feedback elements could decrease the quantity of non-optimal responses.A group of ten human listeners were given a series of single-interval yes-no detection tasks designed to measure response bias due to the previously discussed research questions. The single-interval adjustment matrix (SIAM) procedure (Kaernbach, 1990) was used to determine unbiased detection thresholds at 75% correct for 1 kHz sinusoidal tones in a broad-band white noise masker. These thresholds were used to present a fixed-level stimulus during the IKR detection task. The resulting analyses reveal that IKR does have significant effects on response bias according to the decision strategy which maximizes proportion correct bias (Chapter 1 of Green & Swets, 1966), though these effects are manifested differently in each of the four research questions. The effects due to IKR symmetry were found to yield significant differences in 100% of subjects for an average of 67% of all possible comparisons. The effects due to IKR amount yielded significant differences between conditions for 90% of the subjects for an average of 67% of all possible comparisons. Significant effects between conditions with implicit feedback were found for an average of 67% of all possible comparisons. The slopes of the data designed to test IKR amount were significant in only 50% of subjects. An analysis of the results from the four research questions suggests that the effects due to asymmetry might influence the results of complex combinations of feedback, especially with regards to the amount of IKR.

Patients with dementia due to Alzheimer's disease (AD) tend to exhibit impairments in in their episodic memory. In yes-no tests of recognition memory, patients with AD often display liberal response bias, a stronger tendency to recognize unstudied items as already-studied "old" items. Such tendency is believed to be related to false memory, which can decrease the quality of life in many AD patients. In this study, we analyzed the effect of different instructional manipulations within yes-no recognition memory task on response bias. Younger healthy adults, older healthy adults and one AD patient were evaluated for recognition memory performance and response bias in three different conditions of instructional manipulation. In each session separated by a week-long interval, participants were shown 120 words to study and 240 words, half of which were studied items, to be tested for recognition memory. Instructional manipulation was added in the testing phase of each condition. In one session, the participants were asked if the words were old, studied items; in another session, they were asked if the words were new, unstudied items; finally in the third session, participants were asked to identify if the words were either old or new. Our findings corroborated previous studies by observing liberal response bias in AD and moderately conservative response bias in health adults. We found that the instructional manipulations did not have a significant effect on response bias in either control group while the effect in the AD patient was inconclusive.

Stereotype threat is the threat of confirming a negative stereotype about a group with which a person identifies. Researchers have found that stereotype threat can result in underperformance in multiple domains, shifts in social behavior, and shifts in assessed implicit attitudes, the likelihood of which increases as an individual's concern about the domain of interest increases. According to theory, this threat can be "alleviated",thereby diminishing or eliminating its impact. In this project, over the course of two experiments, the impact of stereotype threat and stereotype threat-alleviation on explicit self-report measures are examined. In experiment one, white college student participants were exposed (or not) to an on-line task intended to elicit race-based stereotype threat. Differences in reporting style (i.e., bias) between the two groups on self-reported measures of race-related attitudes were examined. It was hypothesized that the group exposed to stereotype threat would endorse lower racism and lower stereotypicality (i.e., stereotypic "White" behaviors, attitudes, adjectives, and beliefs). The data provided only partial support for the hypothesis - the threat group reported significantly less stereotypicality than the non threat group. However, the groups were not statistically different on measures of racism or race and social policy. In experiment two, again examining white college students who participated on-line, a stereotype threat-alleviation task was added, and whether this diminished or removed bias was examined. It was hypothesized the threat group would endorse lower stereotypicality and racism than the non threat group and the group receiving the threat alleviation task. The findings from study one did not replicate in study two. Instead, contrary to predictions, across measures of racism and stereotypicality, it was the non threat group that consistently showed the lowest scores. Potential explanations for these findings are offered, including the possibility of having eliciting stereotype threat, cognitive dissonance, or both for the threat and non threat groups via their filler task. Finally, implications for assessing, broaching, and reducing stereotype threat in clinical and research applications are also discussed.

Abstract: Anchoring is judgmental bias in which quantitative estimates assimilate to seemingly irrelevant numerical reference values (Tversky & Kahneman, 1974). While Tversky and Kahneman (1974) originally proposed that anchoring results from the application of a deliberate anchoring-and-adjustment heuristic, other researchers have stressed the role of automatic processes with the introduction of priming-based accounts (Mussweiler & Strack, 1999). In this paper, we present a new perspective on anchoring called consistency theory. On this view, people first determine whether the true target value is above or below the anchor value, and then they provide an estimate that is consistent with the “Greater” or “Less” judgment. Differing from the selective accessibility account, consistency theory assumes that people can be affected by factors such as the response format of the initial comparative judgment. As predicted, we obtained context effects—participants’ judgments of target items were influenced by their judgments of filler items. That is, participants responded “Greater” more often for the target items when they had made fewer “Greater” judgments for the fillers items, and vice versa. Overall, these findings suggest that people can interact with numerical information in a number of different ways, which challenges the view that anchoring is driven by automatic, activation-based processes.

Discrimination of psychogenic nonepileptic (PNES) and epileptic seizures (ES) is a difficult task that is often aided by neuropsychological evaluation. The present study examined differences in neuropsychological performance between these groups using signal detection theory (SDT) to investigate differences in response styles on the Wechsler Memory Scale Word List Test. Raw scores on this task failed to discriminate the two groups; however, patients with PNES were found to have a negative response bias and increased memory sensitivity as compared to patients with ES. Thus, memory impairment in PNES patients may be related to faulty decision making strategies, rather than true memory impairment. When comparing patients with left (LTLE) vs. right temporal lobe epilepsy (RTLE), patients with LTLE demonstrated decreased memory sensitivity but a similar response bias as compared to RTLE patients. This indicates that differences in memory performance in these groups are related to actual differences in memory abilities.

Eating disorders have become prevalent in Western society, and it is estimated that approximately 19% of college females have disordered eating. Cognitive processing differences based on attentional biases of words pertaining to eating disorders are believed to play a large role in assessing people's pathological thoughts. Participants were 165 undergraduate female psychology students (mean age = 19.2) at a large Midwestern university. The present study used self-report questionnaires and a computer-facilitated Stroop task to examine the moderating effect of mood on the relation between pre-existing disordered eating symptoms and Stroop reaction time. This Stroop task included color identification of three word groups (food-related words, neutral words, and color words) that was designed to measure differential speed in cognitive processing of salient versus non-salient words in individuals with and without troubled eating patterns. As predicted by the moderator hypothesis, a statistically significant interaction effect between bulimic symptoms and negative mood was found on food-related reaction time. Post-hoc analysis of the interaction showed that women endorsing more bulimic symptoms responded slower to food-related cues than women with low bulimic symptoms among individuals who reported negative mood. Among those who reported positive mood, women who endorsed more characteristics associated with bulimia showed a faster response rate than those with fewer symptoms. The study results indicate that women who have problematic eating patterns and experience negative mood are hyperattentive to food-related cues.