《一种新的衡量情绪识别能力的测验：松本和艾克曼的日本人与高加索人的短暂表情识别测验（JACBART） 毕业论文外文翻译.doc》由会员分享，可在线阅读，更多相关《一种新的衡量情绪识别能力的测验：松本和艾克曼的日本人与高加索人的短暂表情识别测验（JACBART） 毕业论文外文翻译.doc（18页珍藏版）》请在三一文库上搜索。
1、A NEW TEST TO MEASURE EMOTION RECOGNITION ABILITY: MATSUMOTO AND EKMANS JAPANESE AND CAUCASIAN BRIEF AFFECT RECOGNITION TEST (JACBART)ABSTRACT: In this article, we report the development of a new test designed to measure individual differences in emotion recognition ability(ERA), five studies examin
2、ing the reliability and validity of the scores produced using this test, and the first evidence for a correlation between ERA measured by a standardized test and personality. Utilizing Matsumoto and Ekmans (1988) Japanese and Caucasian Facial Expressions of Emotion (JACFEE) and Neutral Faces (JACNeu
3、F), we call this measure the Japanese and Caucasian Brief Affect Recognition Test (JACBART). The JACBART improves on previous measures of ERA by (1) using expressions that have substantial validity and reliability data associated with them, (2) including posers of two visibly different races (3) bal
4、anced across seven universal emotions (4) with equal distribution of poser race and sex across emotions (5) in a format that eliminates afterimages associated with fast exposures. Scores derived using the JACBART are reliable, and three studies demonstrated a correlation between ERA and the personal
5、ity constructs of Openness and Conscientiousness, while one study reports a correlation with Extraversion and Neuroticism.Research on judgments of emotion from facial expressions has a long and important history in psychology, and has contributed greatly to the literature concerning the universality
6、 of emotion, and to knowledge concerning differences between gender, ethnicity, culture, and psychiatric status. Studies examining the relationship between individual differences in judgments of emotion (hereon referred to as Emotion Recognition AbilityERA) and personality also have a considerable h
7、istory, but is checkered with inconsistent findings. On one hand, ERA has been correlated with emotional expression (e.g., Lanzetta & Kleck, 1970; Levy, 1964; Zuckerman, Hall, DeFrank, & Rosenthal, 1976; Zuckerman, Larrance, Hall, DeFrank, & Rosenthal, 1979), self-monitoring (Mill,1984; Mufson&Nowic
8、ki,1991;Riggio & Friedman, 1982); social desirability (Cunningham, 1977); depression, control, aggression, and gregariousness (Toner & Gates, 1985); and social style, mental ability , achievement , and psychological mindedness (LeRoux, 1987) . On the other hand, Cunningham (1977) failed to replicate
9、 a relationship between self-monitoring and ERA, and Zuckerman et al. (1979) found a relationship for women but not men. Buck, Savin, Miller, and Caul (1972) found a relationship between ERA and extraversion, as did Zuckerman et al. (1979). Cunningham (1977), however, did not replicate these finding
10、s, and instead found a relationship with neuroticism.Theoretically, it is not unreasonable to consider that ERA should be related to stable personality traits. Individuals who are better at judging emotions in others should have greater degrees of interpersonal consciousness or concern; they should
11、be more in tune with their environment, and with others. As an important component of our nonverbal communication system, such skills would be necessary for successful adaptation and manipulation of the environment, ensuring the stability and integrity of the self.Because ERA is an important part of
12、 our daily lives, it is easy to consider how it should be related to various personality constructs, such as those specified in the five factor model. Extraversion, for example, is associated with stimulation seeking from others and the environment. As such, extraverts should be more willing to take
13、 in data concerning the emotions of others, being more interpersonally conscious of others in the environment. Individuals who score high on neuroticism, however, tend to be emotionally avoidant; because they are prone to experience negative emotions, they should have a tendency to avoid the recogni
14、tion and awareness of othersemotions. The personality construct of openness is similar to extraversion in the sense that open individuals tend to be curious and interested in stimulation; they should be more attendant to the emotions of others. Conscientiousness is related to cooperation with and at
15、tending to others; conscientious individuals are more thorough, reliable, and efficient. They should be better at recognizing emotions because they are more attentive to details, and are better able to participate in such emotion judgment tasks. Why have previous attempts to establish a relationship
16、 between personality and ERA been awash with contradictory findings? One possible reason is the stimuli used in previous studies, which were different in each study and thus not equivalent across the studies nor, as Bruner and Tagiuri (1954) suggested, did they cover a representative spectrum of emo
17、tional expressions (LeRoux, 1987). Another possible reason is the fact that, with only one exception (LeRoux, 1987, but these data are not published), many previous studies used measures specifically gene- rated in each study rather than accepted, standardized tests. This distinction is important (O
18、Sullivan,1982), because there is no guarantee that accuracy judgments were made against a valid standard.If a standardized test were available, data could be generated using a valid standard, and the same test can be used across studies. At the very least, inconsistencies in the nature of the stimul
19、i could be ruled out as a possible moderator of the contradictions.Previous Tests of ERA There has been a number of such tests developed in the past, each assessing some aspect of ERA (and its close relative, nonverbal decoding skills,)1 but each with its own limitations (see review by OSullivan, 19
20、82). Some focus on nonverbal behaviors, such as the Profile of Nonverbal Sensitivity (PONS: Rosenthal, Hall, DiMatteo, Rogers, & Archer, 1979), the Social Skills Inventory (SSI: Riggio, 1986), the Social Interpretations Test (Archer & Akert, 1977), and the Diagnostic Analysis of Nonverbal Accuracy S
21、cale (DANVA: Nowicki & Duke, 1994). But, these do not focus on the recognition of discrete emotional states. Other tests focus more closely on emotion, such as the Communication and Reception of Affect Test (CARAT: Buck, 1976), the Test of Emotion Styles (TES: Allen & Hamsher, 1974), the Understandi
22、ng our Feelings test (Elmore,1985), the Feldstein Affect Judgment Test (Wolitzky,1973), the Affective Communication Test (Friedman, Prince, Riggio, & DiMatteo, 1980), and the Contextual and Affective Sensitivity test (CAST: Trimboli & Walker, 1993). But, these are also questionable because of the la
23、ck of validity of the expressions used to portray emotion, the ability to produce specific scores on discrete emotions, or the lack of balance within the test to portray encoder characteristics (e.g., sex, race)equally.The use of facial expressions of emotion that are universally recognized would ad
24、dress one concern.The data associated with expressions of anger, contempt, disgust, fear, happiness, sadness, and surprise provide sufficient evidence of their external validity to portray accurately and reliably these discrete emotional states.In fact, some studies have used these expressions as me
25、asures (e.g., Matsumoto, 1989 , 1992) . But , while they address some concerns, one artifact of their use is the high agreement level in judgments, which precludes the measurement of individual differences.There are at least three ways to address the issue of high agreement levels: (1) reduce image
26、size, (2) distort temporal and/or spatial resolution, or (3) increase presentation speed. Ekman, Brattesani, OSullivan, and Friesen (1979) explored the first method, using two cameras to videotape nurses during“honest”and “dishonest”interviews. One camera provided the“small face”condition in which t
27、he image size was one-fifth the area of a typical human face. The other camera provided the“large face”condition in which the image size was twice the area. Image size did not affect judgments about the nursesaffective states. Ekman et al. (1979) concluded“facial actions provide consistent informati
28、on despite considerable size reduction”(p. 61).Wallbott (1992) examined the second method, using a series of videotaped sequences developed by Scherer (1986) that depict 14 emotional states, and distorted either spatial resolution (pixel resolutionthe number of points or squares constituting a video
29、 frame) or temporal resolution (refreshment ratethe number of frames transmitted per second).Although recognition rates decreased as distortion increased, most recognition rates still remained above chance levels. The stimuli used, however, did not meet independent criteria for validity.The third me
30、thod is to present the stimuli at such fast speeds that judgment accuracy is compromised. Ekman and Friesens (1974) Brief Affect Recognition Task (BART) was created in this manner. It involves very brief (under 1/5s) presentations of facial stimuli, based on Ekman and Friesens (1969) observation of
31、micro-momentary expressions that occur almost outside of conscious awareness, and has been used to assess individual ERA (Ekman & Friesen, 1974; Mufson & Nowicki, 1991). One problem, however, is that facial physiognomy and poser sex are not balanced across emotions; another is the production of afte
32、rimages that affect judgments.Matsumoto and Ekmans (1988) Japanese and Caucasian Facial Expressions of Emotion(JACFEE) addresses the limitation of the expressions used in Ekman and Friesens (1974) BART, and improves on them in several ways. First, it includes equal numbers of posers of two visibly d
33、ifferent ethnic groups , and of males and females within each group, for each of the seven universal emotions. Second, the faces were scored using Ekman and Friesens (1978) Facial Action Coding System (FACS; reliability .91) to verify that the same expressions were shown across posers within each em
34、otion, and that these are associated with universal emotions (Ekman & Friesen,1975,1986). Third, observers in multiple countries and cultures agree in their emotion judgments of the JACFEE expressions (Biehl et al.,1997). There is, therefore, ample support for the validity and reliability of these e
35、xpressions. The next issue is how to alter their presentation to produce reliable individual difference scores.The Development of the JACBARTWe used Matsumoto and Ekmans(1988) JACFEE and Neutral Faces(JACNeuF) (consisting of neutral poses by the JACFEE posers) to develop a new test of ERA which we c
36、all the Japanese and Caucasian Brief Affect Recognition Test(JACBART). Items were created by embedding a JACFEE expression in the middle of a 1s presentation of that posers JACNeuF expression on videotape. This format eliminated afterimages of the target JACFEE expression. Items were placed in a ran
37、dom order, with the condition that the same emotion was not presented consecutively. There was a 3s interstimulus interval, with an orienting tone accompanied by a presenation number shown 1s prior to the item. This format was repeated for all 56 items.Overview of the Studies Reported HereWe report
38、five studies that explore the reliability and validity of the JACBART. Within the realm of reliability, two issues need to be consideredinternal and temporal reliability. With 56 items measuring different aspects of expressionsemotion, poser race, and poser sexand multiple items representing each as
39、pect, items measuring each characteristic of the expressions must be internally consistent with each other, and consistent across time.There are multiple concerns about validity. Face validity is assessed by the overall appearance of the test, and the use of the JACFEE and JACNeuF amply addresses th
40、is concern. Content validity concerns the lexical, logical, and methodological definitions of the construct (OSullivan, 1982), and is addressed by the structure of the JACBART, its rating scales and instructions. Two forms of cons- truct validityconvergent and divergentare demonstrated empirically.
41、Convergent validity refers to the ability of the JACBART to correlate with measures of the same or similar construct, or by intercorrelations among the different JACBART scale scores. Divergent validity requires that the test demonstrate that it is not perfectly correlated with an already existing t
42、est, and that it assesses an aspect of the construct that already existing tests do not assess.Predictive validity refers to the ability of a test to accurately predict other constructs, and there are two types. Concurrent validity refers to the ability of the JACBART to predict scores on a differen
43、t construct when the scores are gathered at the same time. Future predictive validity refers to its ability to predict scores on a different construct measured at a future time. A different version of predictive validity is incremental validity, which refers to its ability to predict a different con
44、struct above and beyond what is already predicted by other similar tests. In the studies reported below, predictive validity was assessed by examining correlations between the JACBART and widely used personality measures.A final consideration is the need to demonstrate that the ERA scores generated
45、by the JACBART are specific to the nature of emotion judgment, and not to general abilities related to taking such tests, including visual acuity, motivation, and the like.General Discussion All studies provided strong evidence for the internal reliability of the JACBART, as well as its convergent v
46、alidity through its intercorrelations.Study 3 provided strong evidence for the temporal reliability of the scores, while Studies 1, 2, and 5 provided strong evidence for its concurrent validity with Openness, and moderate evidence for its concurrent validity with Conscientiousness. Study 4 also prov
47、ided preliminary support for its concurrent validity with extraversion and neuroticism, but specific to one scale. Study 5 demonstrated that the validity coefficients between JACBART scores and the personality measures were not confounded by individual differences in visual acuity.These are the firs
48、t findings in support of a valid and reliable measure of ERA. These data, combined with the external validity associated with the JACFEE expressions used in the JACBART from previous judgment studies and FACS coding, and with the balanced poser race and sex design of the JACFEE, allows the JACBA- RT
49、 to improve on limitations of other existing tests . We recommend the use of JACBART Version 3 with multiscalar ratings, as these produce the best internal re- liability statistics. But, these ratings are cumbersome, and should they be unwieldy, the forced-choice judgment task is much more user-friendly. And, there is no sacrifice in terms of reliability or validity associated with its use.To our knowledge, the findings reported are also the first published