The classification of graphical elements.

In three experiments, participants classified stimuli depicting pie charts and stacked bar graphs on two criteria: a proportion shown in the graph, and the graph's overall size (scaling). Sorting times and errors were measured. For stacked bars, performance was impaired when participants sorted on the proportion and scaling varied. No such impairment occurred for pie charts. Experiment 1 showed that varying scaling produced Garner interference in classification of proportions with stacked bars, but not pies. Experiment 2 showed that this result held when the position of the pie slice was varied; Experiment 3 results showed facilitation for particular combinations of proportion and scaling levels. In general, the results showed that proportion and scaling had an asymmetric integral relation for stacked bar graphs, but were separable dimensions for pie charts.     

Displaying proportions and percentages

Pie and bar charts are commonly used to display percentage or proportional data, but professional data analysts have frowned on the use of the pie chart on the grounds that judgements of area are less accurate than judgements of lenth. Thus the bar chart has been favoured. When the amount of data to be communicated is small, some authorities have advocated the use of properly constructed tables, as another option. The series of experiments reported here suggests that there is little to choose between the pie and the bar chart, with the former enjoying a slight advantage if the required judgement is a complicated one, but that both forms of chart are superior to the table. Thus our results do not support the commonly expressed opinion that pie charts are inferior. An analysis of the nature of the task and a review of the psychophysical literature suggest that the traditional prejudice against the pie chart is misguided.     

Self‐framing of risky choice

Previous research on framing effects has largely focused on how choice information framed by external sources influences the response of a decision maker. This research examined how decision makers framed choice options and how the hedonic tone of self‐framing influenced their risk preference. By using pie charts and a complementary sentence‐completion task in Experiment 1, participants were able to interpret and frame the expected choice outcomes themselves before making a choice between a sure option and a gamble in either a life–death or a monetary problem. Each of these self‐frames (phrases) was then rated by a group of independent judges in terms of its hedonic tone. The hedonic tone of self‐frames was mostly positive and was more positive in the life–death than the monetary context, suggesting a motivational function of self‐framing. However, positive outcomes were still more likely to be framed positively than negative outcomes. In Experiment 2, choice outcomes were depicted with a whole‐pie chart instead of a pie slice in order to emphasize positive and negative outcomes equally. The results showed that the hedonic tone of self‐framing was still largely positive and more positive in the life domain than the monetary domain. However, compared to Experiment 1, the risk preference in the life–death domain was reversed, showing an outcome salience effect: when the pie‐slice chart emphasized only survival outcomes, participants were more risk taking under positive hedonic frames whereas when the whole‐pie chart depicted both survival and mortality outcomes, they became risk averse under positive frames. In sum, self‐framing reflected a positive bias in encoding risk information and affected the risk preference of the decision maker. Like the tone of voice used in communication, the hedonic tone of self‐framing, either positive or negative, can affect risk perception of a choice problem. Copyright © 2003 John Wiley & Sons, Ltd.     

Integration processes during frequency graph comprehension: Performance and eye movements while processing tree maps versus pie charts

Frequency graph types differ in the way how data are translated into visual representations. We compared 2 visualization methods, a traditional circular representation (pie chart) and a rectangular representation (constant column width tree map), which were hypothesized to differ regarding the cognitive ease of visual comparison processes. Performance was evaluated in tasks involving proportion and comparison judgments under both highly controlled and more realistic circumstances. The results showed performance benefits (in terms of reduced response times or error rates) for rectangular representations. Additional eye movement analyses revealed that this benefit was mainly due to a facilitation of scanning the graph for relevant information. The results suggest that facilitating comparison processes by representing the critical variable in less complex visual dimensions (i.e., straight length with constant orientation instead of surface area or curved length) eventually enhances the efficiency of integration processes during graph comprehension.     

Judging proportion with graphs: the summation model

People take longer to judge part-to-whole relationships with bar graphs than with pie charts or divided bar graphs. Subjects may perform summation operations to establish the whole with bar graphs, which would be unnecessary for other graph types depicting the whole with a single object. To test this summation model, the number of components forming the whole was varied with bars, divided bars, reference bars, and pies in three experiments. Response time increased with the number of components for bar graphs but there was little increase for other graph types in Experiment 1. An accuracy emphasis in Experiment 2 produced generally longer response times, but had little effect on the time per summation. The summation operation was not used when graphs were displayed briefly in Experiment 3, although subjects still took longer with bars. The estimated time for a summation operation is consistent with estimates derived from other research. In general, the bar graph is not effective for proportion judgments, and its disadvantage becomes potentially greater as the number of components increases. © 1998 John Wiley & Sons, Ltd.     

When a 12.86% mortality is more dangerous than 24.14%: implications for risk communication

Participants assessed the riskiness of 11 well-known causes of death. Each participant was presented with an estimation of the number of deaths in the population due to that particular cause. The estimates were obtained from a previous study of naive participants' intuitive estimations. For instance, based on the result of the previous study, the number of deaths due to cancer was presented as: ‘2,414 out of 10,000’, ‘1,286 out of 10,000’, ‘24.14 out of 100’ and ‘12.86 out of 100’. The estimates of deaths were presented in analogous ways for the remaining ten causes of death. It was hypothesized that the judged degree of riskiness is affected by the number of deaths, irrespective of the total possible number (such as 10,000 or 100). Results from Experiment 1 were consistent with this prediction. Participants rated cancer as riskier when it was described as ‘kills 1,286 out of 10,000 people’ than as ‘kills 24.14 out of 100 people’, and similar results were observed regarding the remaining 10 causes of death. Experiment 2 replicated this trend. Implications for risk communications are discussed. © 1997 John Wiley & Sons, Ltd.     

Judgement of two causal candidates from contingency information: Effects of relative prevalence of the two causes

In four experiments participants made judgements about two possible causes of an effect. The prevalence of the causes was manipulated independently of their degree of contingency with the effect. Significant effects of the prevalence manipulation were obtained: In particular, ratings of the unmanipulated candidate tended to decline as the prevalence of the other candidate increased, and there was also a significant but smaller effect on judgements of the latter. These tendencies were independent of the order in which the two candidates were judged. These results were replicated under two stimulus presentation procedures, the instance list procedure (Experiments 1 and 2) and the trial-by-trial procedure (Experiment 3). It was hypothesized that people judge, to some degree, the proportionate strength of the candidates, in other words the proportion of occurrences of the effect in the presence of each, and that the cause prevalence effect is a consequence of this tendency. This hypothesis was supported by the results of Experiment 4: Those participants whose judgements of one candidate were negatively correlated with the frequency of occurrence of the effect in the presence of the other candidate showed a significantly stronger cause prevalence effect than the remainder.     

Visual perception of dynamic properties: cue heuristics versus direct-perceptual competence

Constructivist and Gibsonian approaches disagree over the possibility of direct perceptual use of advanced information. A trenchant instance concerns visual perception of underlying dynamic properties as specified by kinematic patterns of events. For the paradigmatic task of discrimination of relative mass in observed collisions, 2 mathematical models are developed, 1 model representing a direct, invariant-based approach, and 1 representing a cue-heuristic approach. The models enable a critical experimental design with distinct predictions concerning performance data and confidence ratings. Although pretraining results were mixed, the invariant-based model was empirically confirmed after a minimal amount of training: Competence entails the use of advanced kinematic information in a direct-perceptual ("sensory") mode of apprehension, in contrast to beginners' use of simpler cues in an inferential ("cognitive") mode.     

Health and Care Related Risk Factors for Suicide Among Nursing Home Residents: A Data Linkage Study

This study aimed to examine associations between health status and care needs of nursing home residents and risk of death from suicide compared to other causes through a retrospective data linkage cohort study examining nursing home resident deaths in Australia between 2000 and 2013. Data linkage was performed between aged care assessment tools—Resident Classification System and Aged Care Funding Instrument—and the National Coronial Information System. A competing risks survival analysis was performed to determine the association between care assessment variables (activities of daily living (ADL), behavior, and complex health care) and the risk of death from suicide and any other cause. Of the 146 nursing home residents who died from suicide, 130 (89%) were matched to their assessment data, with comparable information available for 95 residents (65%). Residents who required high levels of care with ADL, physical health care, and cognitive and behavioral issues had a higher risk of dying from all other causes, yet lower risk of dying from suicide. The study findings demonstrate the feasibility and value of linking these two data sets; highlight a need for improved data collection processes; and support a person‐centered care approach for prevention of suicide among nursing home residents.     

Judgments about Felony‐Murder in Hindsight

The purpose of this research was to investigate the role of hindsight bias in application of the felony‐murder law, a controversial rule stating that felons can be held responsible for any foreseeable deaths that occur as a result of their felony. Some versions of the rule involve notions of proximate cause requiring legal decision makers to determine foreseeability and assess case evidence. Those judgments may be biased by outcome information and the process of cognitive sense‐making. Jury eligible participants read a crime vignette modeled on an actual felony‐murder case manipulated with regard to outcome information and relevant case facts. They made likelihood judgments and rated felons on intent and culpability. Results indicated that outcome information biased death likelihood ratings in hindsight but did not significantly affect assessments of case evidence. Implications for further applied cognitive research regarding the felony‐murder rule are discussed. Copyright © 2013 John Wiley & Sons, Ltd.     

Youth Suicide Deaths: Investigation of Clinical Predictors in a Statewide Sample

Death by suicide is a significant cause of mortality among youth. However, there is limited information on the demographic and clinical factors associated with youth suicide deaths. The objective of this study was to link large statewide databases to describe demographic, clinical, and cause of death characteristics among youth who died by suicide. We examined 1,218 decedents under age 26 who died by suicie between 2000 and 2014. Eighteen died before age 12, 53 died between ages 12 and 14, 292 died between ages 15 and 18, and 855 died between ages 19 and 25. Most were male (83%), and firearm was most common cause of death; 28% previously attempted suicide, 31% had a mental health diagnosis, and 17% were prescribed psychotropic medication. Younger children died by hanging/smothering (89% of all 7‐ to 11‐year olds), and overdose/poisoning increased progressively with age. Adolescents had a higher proportion of females than young adults (23% vs. 14%, p  = .002). Combining data from the medical examiner and large hospital systems allows examination of youth suicide from a developmental perspective. Differences between age groups included gender, method, diagnosed mental illness, and diagnosis of attention deficit hyperactivity disorder. These data point to missed opportunities for effective interventions for specific developmental stages.     

Are absolute frequencies,relative frequencies,or both effective in reducing cognitive biases?

Biases in probabilistic reasoning are affected by alterations in the presentation of judgment tasks. In our experiments, students made likelihood judgments that an event was produced by various causes. These judgments were made in terms of probability, relative frequency or absolute frequency on a full or a pruned list of causes. When they had little personal experience of the event (causes of death), the pruning bias was smaller with relative frequencies than with absolute frequencies or probabilities. When they had more personal experience of the event (missing a lecture), the bias was less with both types of frequency than with probability but still lowest with relative frequency. We suggest that likelihood information is usually stored as relative frequencies when it has been obtained from public sources but that it is based on event counts when it is derived from personal experience. Copyright © 2000 John Wiley & Sons, Ltd.     

"Fixed Pie" a la Mode: Information Availability,Information Processing,and the Negotiation of Suboptimal Agreements

Negotiators often make errors in judgment that lead to suboptimal negotiated agreements. Although researchers agree that these errors result from "fixed pie" expectations, there is disagreement regarding how "fixed pie" expectations lead to poor outcomes. One view suggests that suboptimal agreements result when negotiators accurately process faulty and incomplete information (information availability errors); a second view holds that poorly negotiated agreements result when negotiators inaccurately process valid or complete information (information processing errors). The two experiments described in this paper confirmed that negotiators′ "fixed pie" expectations lead to suboptimal agreements via both information availability and information processing errors.     

The discrimination of graphical elements

A model is proposed to account for how people discriminate quantities shown in pie charts and divided bar graphs (i.e. which proportion is larger, A or B?). The incremental estimation model assumes that an observer sequentially samples from the available perceptual features in a graph. The relative effectiveness of sampled perceptual features is represented by the spread of probability distributions, in the manner of signal detection theory. The model's predictions were tested in two experiments. Participants took longer with pies than divided bars and longer with non‐aligned than aligned proportions in Experiment 1. In Experiment 2, participants took longer with divided bars than pies when graphs were of unequal size. Generally, graphical formats producing longer response times incurred a greater time penalty when the difference between proportions was reduced. These results were in accordance with the model's predictions. Implications for graphical display design are discussed. Copyright © 2001 John Wiley & Sons, Ltd.     

Relative versus absolute measures of explicit attitudes: Implications for predicting diverse attitude-relevant criteria

The authors report 4 studies exploring a self-report strategy for measuring explicit attitudes that uses "relative" ratings, in which respondents indicate how favorable or unfavorable they are compared with other people. Results consistently showed that attitudes measured with relative scales predicted relevant criterion variables (self-report of behavior, measures of knowledge, peer ratings of attitudes, peer ratings of behavior) better than did attitudes measured with more traditional "absolute" scales. The obtained pattern of differences in prediction by relative versus absolute measures of attitudes did not appear to be attributable to differential variability, social desirability effects, the clarity of scale-point meanings, the number of scale points, or overlap with subjective norms. The final study indicated that relative measures induce respondents to consider social comparison information and behavioral information when making their responses more than do absolute measures, which may explain the higher correlations between relative measures of attitudes and relevant criteria.     

Completion of binary statements by children at three academic levels

This study was designed to examine the strategy employed by school-aged subjects when completing binary statements. Thirty-two incomplete binary statements representing four statement conditions (true affirmative, true negative, false affirmative, and false negative) were administered to 120 children randomly selected from three academic levels. A response-index change model is postulated as representing the strategy employed by these school-aged subjects when completing binary statements. According to this model, the factor that causes response time to differ across the four statement conditions is the number of times the child is required to change the set of his response index. The child begins the task of completing a binary statement by reading and encoding the incomplete statement. He then selects the numeral that corresponds to the adjective “odd” or “even” contained in the predicate of the incomplete statement. The child’s response index is now set to respond with this numeral. What the child does at this point will depend upon the type of statement condition with which he is working. If the child is completing a TA statement, no response-index change is required. He simply gives as his answer the numeral to which his response index is set. If the child is completing an FA statement or a TN statement, one response-index change is required. If the child is completing an FN statement, two response-index changes are required.     

其它可能原因对于因果共变信息作用的影响

研究使用相继呈现信息的方法控制了被试获得信息的顺序,将其它可能原因在因果判断过程中所起到的作用分离了出来,并按照其对不同共变信息的影响进行了分析。研究结果表明：(1)其它可能原因会影响到被试利用因果共变信息而进行的因果推断。(2)其它可能原因与待判断原因共同存在的程度在很大程度上影响着其它可能原因所起到的作用。(3)其它可能原因对于不同共变信息的影响有着明显的差别。     

Using statistical risk information to assess oral contraceptive safety

The effect of variations in the presentation of oral contraceptive risk information on perceived safety was investigated. Four tests of the hypothesis that consumers focus on the absolute value of the numbers presented when assessing risk probabilities were performed with the following results: (1) when base rates are given, risks are perceived as less likely than when base rates are not given; (2) risks seem smaller when the information provided focuses on the number that will not die than when the number of expected deaths is presented; (3) contrary to our expectation, respondents do not focus on foreground numbers when risk was represented in an ‘X out of Y occurrences’ format; (4) respondents judged a risk of ‘4.15 times greater’ as less likely than when the information was presented as ‘415% greater risk’. We concluded that consumers convert statistical risk information into judgements about the magnitude of the number presented (small or large number) and utilize this more meaningful information when making risk judgements.