A "change-of-standard" perspective on the relations among context, judgment, and memory

We conducted three studies that tested a "change-of-standard" perspective on the relations among context, judgment, and recall. Each study consisted of two or three sessions held a few days apart. All subjects read about the sentencing decisions of one or two target trial judges and of six nontarget trial judges who consistently gave either higher or lower sentences than the target judge(s). Each study varied both the standard that was available when subjects initially judged the sentencing decisions of a target judge and the standard available when subjects subsequently recalled those decisions. To accomplish this, we varied the context of judgment, the timing of judgment, and the overall category norm for trial judges' sentencing decisions that was available at recall. We found that although subjects had been exposed to the same target information and had initially judged it in the same way, their recall of the information was different depending on whether and how a change-of-standard had occurred between judgment and recall. Unique predictions of the change-of-standard perspective were confirmed that could not be accounted for in terms of other types of context effects on judgment and memory.     

Guidelines are not enough: The need for written sentencing opinions

Guideline sentencing systems, including the new federal guidelines, have not settled on a clear conception of when and how a trial judge should explain sentences. Indeterminate sentencing systems did not have a tradition of written sentencing decisions and recent sentencing reforms do not focus on the trial judge's role. This article suggests the many advantages of written sentencing opinions. Initial experience under the federal sentencing guidelines bolsters the conclusion that written sentencing opinions in appropriate cases—including both sentences “within” guidelines and guideline “departures”—are the next step in the evolving law of sentencing and the best way to recognize trial judges as an essential engine of principled change.     

Memory Errors From A Change of Standard: A Lack of Awareness or of Understanding?

Memory errors occur when the context standard that is used when judging target behaviors is different from the category norm standard that is available later when the behaviors are recalled. Is insufficient awareness the reason for this change-of-standard effect? Two kinds of awareness were maximized in each of two studies: (a) awareness of the relation between the judgment and the context - telling subjects to be sure to use the non-target persons for comparison when judging the target person (salience); and (b) awareness at recall of the earlier judgmental context - asking subjects to recall the non-target persons before recalling the target person (reinstatement). Context reinstatement reduced memory errors. But when context reinstatement and salience were combined, the memory errors reappeared. In Study 2, an attempt at debiasing failed. The change-of-standard effect is explained in terms of a "natural" tendency to use the current categorical meaning of a judgment to reconstruct the referent of its past contextualized meaning.     

Aging, expertise, and narrative processing.

In a study of how aviation expertise influences age differences in narrative processing, young and older pilots and nonpilots read and recalled aviation and general narratives. They chose referents for sentences referring to a protagonist or minor character mentioned 1 sentence (recent character) or 3 sentences (distant character) before this target sentence. All groups chose referents less accurately for sentences about distant and minor characters than about recent and protagonist characters, perhaps because these referents were less likely to be in working memory. Young readers and pilots were more accurate for distant and minor character target sentences in aviation narratives, and recalled aviation narratives more accurately. Expertise did not reduce age differences. Expertise differences may reflect decreased demands on working memory capacity, and age declines may reflect reduced capacity.     

Indirect social influence: Change in judgmental process or anticipatory conformity?

In three experiments, subjects imposed sentence in a criminal case before and after reading the harsh sentence imposed by the trial judge. They then read and imposed sentence in another case. In one experimental condition the subjects anticipated learning the judge's sentence in Case 2; in the other condition they were told not to anticipate learning the judge's sentence. Control subjects made their judgments without ever seeing the judge's sentence. Direct influence was the change in sentences within cases; indirect influence was the difference between initial sentences in Cases 1 and 2. Though direct influence was minimal and not significant, indirect influence was substantial, and was greater in the anticipation than no-anticipation than control conditions, indicating that both modification of the judgmental process and anticipatory conformity were operating. Indirect influence may then reflect a change of either type, depending on a number of methodological and other factors.     

A rule-based categorization deficit in Alzheimer's disease?

This study examined the categorization processes that Alzheimer's disease (AD) patients use during assessments of semantic memory. Rule-based categorization involves the careful, analytic processing of strict criteria to determine category membership, particularly for items from graded categories with ambiguous category membership; similarity-based categorization requires an overall comparison of a test stimulus with a prototype or remembered exemplar of the category and is relatively effective for the rapid categorization of items with unambiguous category membership. To assess these processes in AD, patients were asked to decide the category membership of test stimuli for categories with poorly defined or fuzzy boundaries (e.g., VEGETABLE) and for categories with well-defined boundaries (e.g., FEMALE) and then to judge the representativeness of the test stimulus for its chosen category. A subgroup of AD patients demonstrated a typical pattern of impaired semantic memory compared to healthy control subjects; that is, difficulty deciding the category membership of test items from fuzzy categories. Among these patients, we found no deficit in category membership decisions about items taken from well-defined categories. We also found that AD patients and healthy controls do not differ in their representativeness judgments of items within a correctly judged category. These findings are most consistent with the hypothesis that rule-based categorization difficulty limits semantic memory in AD.     

Comprehension,selective imitation,and the CIP hypothesis

The role of comprehension training in the selective imitation of indirect-direct object sentences was assessed for six preschool children. A modeling condition resulted in normal usage of indirect-direct object sentences for five of six subjects, but reversed usage was not obtained when modeling was reversed. Subsequent receptive comprehension training on normal and reversed forms of indirect-direct object sentences resulted in generalization to the productive mode. These results were related to an hypothesis about language acquisition in which selective imitation is a function of comprehension.     

The effects of thematic context and presentation mode on memory for sentence voice

A sentence in discourse may appear in the passive voice to emphasize the logical object rather than the logical subject when it is thematically more important. Two experiments are reported that explore the impact of this textual function of voice on sentence memory. The first experiment required subjects to listen to prose passages and then recall them. Sentences were recalled predominantly in the active voice regardless of voice or thematic focus in the prose passage, showing that the English-language bias for the active voice was a more important determinant of sentence reconstruction than was the experimental manipulation of thematic context. The second experiment required subjects to listen to or read either prose passages or lists of unrelated sentences and then to try to recognize key sentences that were either unchanged or changed lexically, semantically, or in voice. Recognition, both overall and specifically for voice, was better for sentences that were read than for those that were heard, and recognition for semantic change was consistently higher than for any other. Only when passages were read was there evidence in support of a thematic textual influence on memory for sentence voice.This research was supported in part by NIMH Grant 32063.     

Through the looking glass: Sex differences in memory for self-facial poses

This study investigated recognition memory of photographs of the subject's own face. Male and female subjects were photographed as they projected sociable faces, trustworthy faces, and intelligent faces. After deciding which face of 10 best represented each characteristic, and judging which photograph best represented their “real self,” a recognition memory test of poses was given. Half of each sex were tested under intentional learning conditions and the remainder were tested under incidental learning conditions. Females demonstrated superior recognition memory of their own facial projections and, in particular, recalled photographs of their “real self” and “most sociable” self most easily. No differences were found between the two learning conditions. Subjects' recognition performance was not related to their confidence of judgments. The results were discussed in terms of sex differences and the role of self in memory.     

Situational information contributes to object categorization and inference

Three experiments demonstrated that situational information contributes to the categorization of functional object categories, as well as to inferences about these categories. When an object was presented in the context of setting and event information, categorization was more accurate than when the object was presented in isolation. Inferences about the object similarly became more accurate as the amount of situational information present during categorization increased. The benefits of situational information were higher when both setting and event information were available than when only setting information was available. These findings indicate that situational information about settings and events is stored with functional object categories in memory. Categorization and inference become increasingly accurate as the information available during categorization matches situational information stored with the category.     

Individual differences in field articulation as a factor in language comprehension

Previous research has suggested that performance differences associated with field independence/field articulation (FA) reflect differences in the efficiency of a limited capacity “working memory” (WM) system. An information-processing analysis of language comprehension suggests the involvement of WM. This experiment tested the hypothesis that performance differences in a language comprehension task could be predicted on the basis of level of FA and the mediating role of WM. Three levels of memory load were established by varying the logical structure of sets of semantically related sentences which subjects were required to integrate. In the low memory load condition, there was no difference between high and low FA subjects on a test of integration (inference) and specific memory (recognition). In the high memory load condition, there was a decline in performance for low FA subjects on both inference and recognition. Other aspects of the results suggested that the breakdown in information processing, for low FA subjects in the high memory load condition, occurred during or shortly after initial encoding of the sentences.     

Sentencing Disparity: Perforce or Perchance?1

Important sources of variability in sentencing judgments are identified by decomposing these judgments in a Crime × Judge analysis of variance into a Crime main effect (representing consensus) and several forms of disparity. Two systematic forms of disparity are (1) the Judge main effect, reflecting differences in overall harshness or leniency of different judges; and (2) the Crime × Judge interaction, reflecting idiosyncracies—perhaps principled—in the way particular judges view particular crimes. A third unsystematic form of disparity is reflected in variations in how the same judge views the same crime/offender on different occasions, that is, unreliability. Data from an experiment in which state district judges made sentencing decisions about the same crime/offenders on two occasions are presented to illustrate application of the model and obtain disparity estimates for 13 crimes. Results reveal that for most of the crimes disparity can be attributed primarily to simple unreliability over occasions rather than principled disagreements among judges. Implications of these results and the importance of distinguishing different types of disparity are discussed.     

Sentencing Under Uncertainty: Anchoring Effects in the Courtroom1

Research on juridical decision making has demonstrated that largely disparate sentences are often given for identical crimes. This may be the case because judges' sentencing decisions are influenced by a recommended or demanded sentence. Building on research on judgmental anchoring (Tversky & Kahneman, 1974), the present investigation examines whether a sentencing demand has a direct influence on a given sentence. Using criminal trial judges as participants, Study 1 demonstrates that such a direct influence does, in fact, exist. Sentencing decisions are assimilated to the sentence demanded by the prosecutor. Study 2 further reveals that this influence is independent of the perceived relevance of the sentencing demand. Study 3 demonstrates that this influence is also independent of judges' experience.     

Apolitical courts? Rape sentencing in Montana

This analysis examines dockets of the state courts of Montana for the years 1987-1989 in rape cases to determine whether sentencing decisions are significantly related to three political variables: the percentage of women in county executive office as a measure of a relevant judicial constituency, the extent to which the county in which a court is located is Democratic, and whether the judge who enters a decision was appointed or elected. The analysis suggests that two legally relevant facts—the crime rate and multiple charges against a defendant—are significantly related to outcome. Political factors do not appear to influence state court sentencing decisions in rape cases since: (1) sentencing decisions in rape cases are not ‘major’, publicized issues; (2) the selection of state court judges (appointed and elected) is nonpartisan; and (3) interest groups were not involved in rape sentencing cases.     

Attention, similarity, and the identification-categorization relationship

A unified quantitative approach to modeling subjects' identification and categorization of multidimensional perceptual stimuli is proposed and tested. Two subjects identified and categorized the same set of perceptually confusable stimuli varying on separable dimensions. The identification data were modeled using Shepard's (1957) multidimensional scaling-choice framework. This framework was then extended to model the subjects' categorization performance. The categorization model, which generalizes the context theory of classification developed by Medin and Schaffer (1978), assumes that subjects store category exemplars in memory. Classification decisions are based on the similarity of stimuli to the stored exemplars. It is assumed that the same multidimensional perceptual representation underlies performance in both the identification and categorization paradigms. However, because of the influence of selective attention, similarity relationships change systematically across the two paradigms. Some support was gained for the hypothesis that subjects distribute attention among component dimensions so as to optimize categorization performance. Evidence was also obtained that subjects may have augmented their category representations with inferred exemplars. Implications of the results for theories of multidimensional scaling and categorization are discussed.     

The effect of pictorial context on sentence memory in Broca's and Wernicke's aphasia

Eighteen aphasic patients (8 Broca's and 10 Wernicke's aphasics), 11 right hemisphere damaged patients and 12 normal subjects were tested to assess the effect of pictorial context on verbal memory with a sentence recognition task. The subjects were read aloud a stimulus sentence describing a simple event and simultaneously shown a picture congruent or incongruent with the sentence. Immediately following or after an interval of 30 sec, the subjects were read aloud a second sentence and asked to judge whether this sentence was the same or different from the stimulus sentence. The results indicated that verbal memory was better retained in supportive situations than in distracting situations, and that this contextual effect was greater in aphasic patients than in any of the other groups. Verbal memory declined rapidly after an interval in aphasic patients, but not in normal subjects. The Broca's and the Wernicke's aphasics demonstrated different performance patterns when semantic analysis of sentences was critical.