Prior studies suggest that people often search insufficiently in sequential-search tasks compared with the predictions of benchmark optimal strategies that maximize expected payoff. However, those studies were mostly conducted in individualist Western cultures; Easterners from collectivist cultures, with their higher susceptibility to escalation of commitment induced by sunk search costs, could exhibit a reversal of this undersearch bias by searching more than optimally, but only when search costs are high. We tested our theory in four experiments. In our pilot experiment, participants generally undersearched when search cost was low, but only Eastern participants oversearched when search cost was high. In Experiments 1 and 2, we obtained evidence for our hypothesized effects via a cultural-priming manipulation on bicultural participants in which we manipulated the language used in the program interface. We obtained further process evidence for our theory in Experiment 3, in which we made sunk costs nonsalient in the search task-as expected, cross-cultural effects were largely mitigated.
Lind, Hall, Breidegard, Balkenius, and Johansson (2014a, 2014b) recently published articles tackling a core question concerning speech production: At which stage of processing are communicative intentions specified? Taking a position contrary to dominant models of speech production (e.g., Levelt, 2001), they suggested that utterances “often are semantically underspecified” (Lind, Hall, Breidegard, Balkenius, & Johansson, 2014a, p. 8) before articulation, and that “auditory feedback” (Lind et al., 2014a, 2014b) is an important mechanism for specifying the meaning of what one says. In their experiment (Lind et al. 2014b), they used real-time speech exchange, in which auditory information was occasionally manipulated while participants performed a Stroop task (i.e., naming the colors in which color names were printed). This methodology created trials in which participants produced the correct response (i.e., named the color of the text) while simultaneously hearing over headphones a recording of their own voice producing the incorrect response (i.e., reading the text). On these trials, when asked what they had said, participants sometimes reported the word that they had heard, rather than the word that they had produced. Failure to detect these intrusions was interpreted as evidence that auditory feedback is a “prime channel” (Lind, Hall, Breidegard, Belkenius, & Johansson, 2014a, p. 2) for monitoring the meaning of spoken words. We suggest that (a) the authors’ data constitute evidence against a prime role for auditory feedback in monitoring the meaning of self-produced speech, (b) the experimental manipulation is not necessarily a manipulation of auditory feedback, and (c) the findings may be explained by task demands rather than by mechanisms of speech production. In the experiment (Lind et al., 2014b) participants always heard their own responses, via normal bone conduction and also via headphones. In a subset of trials (4 out of 250 per participant), the word they heard over headphones was a recording, from earlier in the experiment, of the word that they currently needed to inhibit. Lind et al. focused their analysis and interpretation exclusively on those trials within this subset in which participants did not notice that the word had been exchanged. However, according to the criteria Lind et al. used, there was evidence that the majority of these exchanges were in fact detected (~73%), and only a minority went unnoticed (~27%). The frequent detection of the manipulations suggests that auditory feedback is unlikely to be a prime mechanism (Lind et al., 2014a) by which spoken intentions are specified, as participants must have had some awareness of what they actually said when they detected the exchanges. Although the data seemingly contradict the authors’ claim that speakers “always use [auditory feedback] as a source of evidence in interpreting their own utterances” (Lind et al., 2014b, p. 1199), it remains possible that speakers use auditory feedback some of the time, when self-monitoring. However, we argue that the experience of hearing the audio recordings differed so much from the usual experience of hearing one’s own voice that it cannot be assumed to necessarily reflect the same processes. People who listen to a recording of their own voice notice that their voice sounds different from how they hear it when they speak; they primarily hear their own voice via bone conduction but other people’s voices through air conduction. Lind et al. accounted for some of the acoustic differences between self- and other-produced speech as it is heard at the ear (e.g., by low-pass filtering the speech), but made no attempt to address others (e.g., by mimicking the spatial location of participants’ own voices). Furthermore, Lind et al. could not eliminate the perception of the concurrent somatosensory and bone-conducted auditory information that always results from speaking aloud. Within the dorsolateral temporal lobes, the neural response to self-produced speech differs from the response to other-produced speech (Agnew, McGettigan, Banks, & Scott, 2012; Wise, Greene, Buchel, & Scott, 1999), and peripherally, the loudness of one’s voice is dampened by the stapedius reflex, which occurs when one speaks. It is critical to determine, therefore, whether the auditory feedback in the experiment reflected the auditory, somatosensory, and neural consequences of perceiving one’s own voice while speaking, or whether it simply functioned more generally as an auditory distractor. This distinction matters, because there are many instances in which hearing other-produced speech influences perception and production. Properties of the speech sounds in the environment can be assimilated into a speaker’s output (Delvaux & Soquet, 2007; Pickering & Garrod, 2004), and speech that a person is trying to ignore can be confused with attended speech, particularly if the utterances are semantically related (Brungart, 2001; Gray & Wedderburn, 1960). Thus, had participants heard any voice (not necessarily their own), they might have made the same misattributions. To ensure that the observed effect was dependent on participants hearing their own voice, the authors would have needed an additional control condition in which participants heard a voice other than their own overlaid on their vocal responses. Finally, we question the extent to which performance on the Stroop task generalizes to natural speech interactions. Stroop interference results from competition between the color of the text and the distractor (the incongruent written word; MacLeod, 1991; van Veen & Carter, 2005). Both are automatically processed and prepared for response production, and executive-control systems are required for the final response selection. Although manipulation of stimulus onset asynchrony shows that Stroop interference dissipates over relatively short latencies (~400–500 ms; for reviews, see MacLeod, 1991; Roelofs, 2003), congruency sequence effects suggest that cognitive control is sustained and adjusted over longer time scales (Egner, 2007; Kerns et al., 2004). Cognitive-control demands likely reduce capacity for error monitoring of one’s own speech. Indeed, individuals make fewer self-repairs of their speech in dual-task paradigms than in single-task paradigms (Oomen & Postma, 2002) and show poorer pre-articulatory self-monitoring when naming under time pressure than when naming under no time pressure (Dhooge & Hartsuiker, 2012)—effects that are concordant with the notion that error monitoring is under central control and is capacity limited (Levelt, 1983). The data reported by Lind et al. (2014b) seem best explained by breakdowns in error monitoring consequent to the executive processing demands of the Stroop task, rather than by processes underlying speech production more generally. To conclude, we are sympathetic to the wider concept that the act of speaking helps people construe their intended meanings in interactions. However, the articles by Lind et al. (2014a, 2014b) do not make a convincing case for placing auditory feedback at the heart of this mechanism.
We examined the potential cost of practicing suppression of negative thoughts on subsequent performance in an unrelated task. Cues for previously suppressed and unsuppressed (baseline) responses in a think/no-think procedure were displayed as irrelevant flankers for neutral words to be judged for emotional valence. These critical flankers were homographs with one negative meaning denoted by their paired response during learning. Responses to the targets were delayed when suppression cues (compared with baseline cues and new negative homographs) were used as flankers, but only following direct-suppression instructions and not when benign substitutes had been provided to aid suppression. On a final recall test, suppression-induced forgetting following direct suppression and the flanker task was positively correlated with the flanker effect. Experiment 2 replicated these findings. Finally, valence ratings of neutral targets were influenced by the valence of the flankers but not by the prior role of the negative flankers.
Upright static faces are widely thought to recruit holistic representations, whereby individual features are integrated into nondecomposable wholes for recognition and interpretation. In contrast, little is known about the perceptual integration of dynamic features when viewing moving faces. People are frequently exposed to correlated eye and mouth movements, such as the characteristic changes that accompany facial emotion, yawning, sneezing, and laughter. However, it is unclear whether the visual system is sensitive to these dynamic regularities, encoding facial behavior relative to a set of dynamic global prototypes, or whether it simply forms piecemeal descriptions of feature states over time. To address this question, we sought evidence of perceptual interactions between dynamic facial features. Crucially, we found illusory slowing of feature motion in the presence of another moving feature, but it was limited to upright faces and particular relative-phase relationships. Perceptual interactions between dynamic features suggest that local changes are integrated into models of global facial change.
Reanalysis of the survey data sets of Lewandowsky, Oberauer, and Gignac (2013) and Lewandowsky, Gignac, and Oberauer (2013) indicates that the conclusions of those articles—that conspiracist ideation predicts skepticism regarding the reality of anthropogenic climate change—are not supported by the data. Nonlinear relationships were overlooked in both analyses, and this resulted in model misspecification. The authors used structural equation modeling (SEM) assuming linear relationships between the variables, and it is essential to test this assumption (Bentler and Chou, 1987, p. 86; Ullman, 2007, p. 683). In this Commentary, we show, using nonparametric local regression, that this assumption does not hold for the relationship between conspiracist ideation and views on climate science, the relationship that produced one of the central claims of both articles and the majority of the press interest (e.g., Corner, 2012; Pearlman, 2012).
The mechanism by which nonspatial features, such as color and shape, are bound in visual working memory, and the role of those features' location in their binding, remains unknown. In the current study, I modified a redundancy-gain paradigm to investigate these issues. A set of features was presented in a two-object memory display, followed by a single object probe. Participants judged whether the probe contained any features of the memory display, regardless of its location. Response time distributions revealed feature coactivation only when both features of a single object in the memory display appeared together in the probe, regardless of the response time benefit from the probe and memory objects sharing the same location. This finding suggests that a shared location is necessary in the formation of bound representations but unnecessary in their maintenance. Electroencephalography data showed that amplitude modulations reflecting location-unbound feature coactivation were different from those reflecting the location-sharing benefit, consistent with the behavioral finding that feature-location binding is unnecessary in the maintenance of color-shape binding.
Memory is critical for learning, cognition, and cognitive development. Recent work has suggested that preschool-age children are vulnerable to catastrophic levels of memory interference, in which new learning dramatically attenuates memory for previously acquired knowledge. In the work reported here, we investigated the effects of consolidation on children's memory by introducing a 48-hr delay between learning and testing. In Experiment 1, the delay improved children's memory and eliminated interference. Results of Experiment 2 suggest that the benefit of this delay is limited to situations in which children are given enough information to form complex memory structures. These findings have important implications for understanding consolidation processes and memory development.
Decisions about allocation of scarce resources, such as transplant organs, often entail a trade-off between efficiency (i.e., maximizing the total benefit) and fairness (i.e., dividing resources equally). In three studies, we used a hypothetical scenario for transplant-organ allocation to examine allocation to groups versus individuals. Study 1 demonstrated that allocation to individuals is more efficient than allocation to groups. Study 2 identified a factor that triggers the use of fairness over efficiency: presenting the beneficiaries as one arbitrary group rather than two. Specifically, when beneficiaries were presented as one group, policymakers tended to allocate resources efficiently, maximizing total benefit. However, when beneficiaries were divided into two arbitrary groups (by hospital name), policymakers divided resources more equally across the groups, sacrificing efficiency. Study 3 replicated this effect using a redundant attribute (prognosis) to create groups and found evidence for a mediator of the grouping effect--the use of individualizing information to rationalize a more equitable allocation decision.
Spouses influence each other's psychological functioning and quality of life. To explore whether this interdependence continues after a person becomes widowed, we tested whether deceased spouses' characteristics were associated with their widowed partners' later quality of life using couples drawn from a multinational sample of aging adults. Independent subsamples (ns = 221 and 325) were assessed before and after a spouse's death. Regressions revealed that deceased partners' quality of life prior to their death positively predicted their spouses' quality of life after the partners' death, even when we controlled for spouses' prior quality of life to account for environmental factors shared within couples. Further, widowed participants' quality of life was lower than nonwidowed couples' 2 years before and after their partners' death, but was equivalent 4 years prior. Finally, the strength of the association between partners' earlier quality of life and participants' later quality of life did not differ between widowed and nonwidowed participants. These findings suggest that interdependence in quality of life continues after one's partner has passed away.
In the study reported here, we examined the effects of fetal exposure to a synthetic stress hormone (synthetic glucocorticoids) on children's susceptibility to postnatal sociodemographic adversity. We recruited children who were born healthy and at term. Twenty-six had been treated with steroid hormones (glucocorticoids) during the prenatal period, and 85 had not. Only children exposed to both prenatal stress hormones and postnatal sociodemographic adversity showed impaired performance on standardized tests of memory function. The association was specific to long-term memory. General intellectual functioning and expressive language were not affected by fetal glucocorticoid exposure. Results were independent of maternal intelligence and maternal depression at the time of the study. These findings are consistent with a vulnerability-stress model: Prenatal exposure to synthetic stress hormones is associated with increased susceptibility to subsequent adversity, with consequences for cognitive functioning that persist 6 to 10 years after birth.