# _Attitude and the Prediction of Behavior_
# _態度與行為的預測_
Martin FIShbein
馬丁·菲什賓
# Abstract
# 摘要
Throughout this book, relatively little has been said about the relations between attitude and behavior. To a large extent, this omission reflects the present state of the field. After more than seventy-five years of attitude research, there is still little, if any, consistent evidence supporting the hypothesis that knowledge of an individual's attitude toward some object will allow one to predict the way he will behave with respect to the object. Indeed, what little evidence there is to support any relationship between attitude and behavior comes from studies showing that a person tends to bring his attitude into line with his behavior rather than from studies demonstrating that behavior is a function of attitude (e.g., Cohen, 1960; Gerard, 1965; Landy, 1966).
在本書中,對於態度與行為之間的關係相對較少提及。在很大程度上,這一省略反映了該領域的現狀。在經過七十五年以上的態度研究後,仍然幾乎沒有一致的證據支持這一假設:對某個對象的個人態度的了解能夠預測他對該對象的行為。事實上,支持態度與行為之間任何關係的微薄證據來自於顯示一個人傾向於使其態度與行為一致的研究,而不是來自於證明行為是態度的函數的研究(例如,Cohen,1960;Gerard,1965;Landy,1966)。
It is my contention that we psychologists have been rather naive in our attempts to understand and to investigate the relationships between attitude and behavior. More often than not, we have attempted to predict some behavior from some measure of attitude and found little or no relationship between these variables. Yet, rather than questioning our basic assumption that there is a strong relationship between attitude and behavior, we have tended to blame our failures on our measuring instruments, on our definition of attitude, or on both. Thus, from its relatively simple beginning as a unidimensional concept that referred to the amount of affect for or against some psychological object, the concept of attitude has grown into a complex, multidimensional concept consisting of affective, cognitive, and conative components.
我認為我們心理學家在理解和研究態度與行為之間的關係時,顯得相當天真。往往,我們試圖根據某種態度的測量來預測某種行為,但卻發現這些變數之間幾乎沒有關係。然而,我們並沒有質疑我們的基本假設,即態度與行為之間存在強烈的關係,而是傾向於將失敗歸咎於我們的測量工具、對態度的定義,或兩者兼而有之。因此,態度這一概念從最初作為一個單維度的概念,指的是對某些心理對象的情感程度,發展成為一個複雜的多維度概念,包含情感、認知和意向成分。
While this view has done a great deal to stimulate new types of research and, more specifically, to improve our understanding of the relationship between belief and attitude (i.e., between cognition and affect), and while it has recently led to investigations of the relationships between attitude and behavioral intentions (i.e., between affect and conation), it has not, to the best of my knowledge, contributed very much to our understanding of the specific relationships between attitude and behavior.
雖然這種觀點在刺激新類型的研究方面做出了很大貢獻,並且更具體地改善了我們對信念與態度之間關係(即認知與情感之間的關係)的理解,並且最近促使了對態度與行為意圖之間關係(即情感與意志之間的關係)的研究,但據我所知,它對我們理解態度與行為之間的具體關係並沒有太大貢獻。
## The Concept of Attitude
## 態度的概念
Before turning to a consideration of these relationships, however, let us first briefly review the concept of attitude. Perhaps the best way to begin is to provide a brief (and admittedly selective) historical sketch. In 1935, Gordon Allport reviewed the general area of attitude theory and research. After considering more than one hundred different definitions of attitude, Allport concluded that most investigators basically agreed that an attitude is a learned predisposition to respond to an object or class of objects in a consistently favorable or unfavorable way. Furthermore, he pointed out that this bipolarity in the direction of an attitude (i.e., the favorable versus the unfavorable) is often regarded as the most distinctive feature of the concept. Thus attitude was conceptualized as a simple unidimensional concept.
在考慮這些關係之前,我們首先簡要回顧一下態度的概念。或許最好的開始方式是提供一個簡短(且無可否認地有選擇性)的歷史概述。在 1935 年,戈登·奧爾波特回顧了態度理論和研究的一般領域。在考慮了超過一百種不同的態度定義後,奧爾波特得出結論,大多數研究者基本上同意,態度是一種對某個對象或對象類別以一致的有利或不利方式作出反應的學習傾向。此外,他指出,態度方向上的這種雙極性(即有利與不利)通常被視為該概念最具特徵的特徵。因此,態度被概念化為一個簡單的單維概念。
However, as Allport noted, research based on this conception of attitude had not resulted in behavioral prediction, and it appeared to him that this unidimensional view was oversimplified. Indeed, from Allport's point of view, two people could be equally favorable toward an object, yet feel differently about components or characteristics of the object. For example, two people could feel equally favorable toward the church but feel quite differently about characteristics or practices of the church. Similarly, two people could be equally in favor of change but disagree on the modus operandi of a reform
然而,正如奧爾波特所指出的,基於這種態度概念的研究並未導致行為預測,而在他看來,這種單維度的觀點過於簡化。事實上,從奧爾波特的角度來看,兩個人對一個對象的看法可能同樣正面,但對該對象的組成部分或特徵卻感受不同。例如,兩個人可能對教會的看法同樣正面,但對教會的特徵或實踐卻有截然不同的看法。同樣,兩個人可能都支持變革,但對改革的運作方式卻存在分歧。
---
movement. Thus, according to Allport, although two people may have the same degree of affect toward an object, they may differ _qualitatively_ in their attitude toward it. Obviously, one reason why we cannot predict behavior from attitude, he continued, is that our measures of attitude are unidimensional and do not take this qualitative nature of attitude into account.
運動。因此,根據阿爾波特的說法,雖然兩個人對一個物體的情感程度可能相同,但他們對該物體的態度可能在質上有所不同。顯然,他接著說,我們無法從態度預測行為的原因之一是,我們對態度的測量是單維的,並未考慮到態度的這種質的特性。
Allport argued, therefore, for the consideration of the qualitative nature of attitude. Allport's plea was not accompanied by a technique for measuring these qualitative differences, however, and this, in part, may explain why the plea fell on deaf ears. Indeed, following Allport's review, investigators still continued to measure attitude by obtaining single scores that essentially placed the respondent somewhere along a single dimension of favorableness or un-favorableness toward the attitude object. In fact, it is worth noting that two of the major attitude measuring instruments that have been introduced since Allport's review have been designed specifically to obtain unidimensional scores (i.e., the Guttman Scale and the Semantic Differential). Nevertheless, the prediction of behavior remained a major concern, _and_ an unsolved problem.
奧爾波特因此主張應考慮態度的質性特徵。然而,奧爾波特的呼籲並未附帶測量這些質性差異的技術,這在某種程度上可能解釋了為何這一呼籲未能引起重視。事實上,在奧爾波特的回顧之後,研究者們仍然繼續通過獲取單一分數來測量態度,這些分數基本上將受訪者置於對態度對象的單一好感或不滿意度的某個位置。實際上,值得注意的是,自奧爾波特的回顧以來,推出的兩個主要態度測量工具專門設計用來獲取單維度分數(即古特曼量表和語義差異量表)。儘管如此,行為的預測仍然是一個主要關注點,並且是一個未解決的問題。
I think the second major step came when Leonard Doob (1947) suggested that there may not be any one-to-one relationship between attitude and behavior. Doob's argument was quite straightforward, and it seems surprising that so little attention has been paid to it.
我認為第二個重要的步驟出現在倫納德·杜布(1947 年)提出態度與行為之間可能不存在一對一關係的時候。杜布的論點相當直接,令人驚訝的是,對此關注如此之少。
Basically, Doob's argument was as follows: Attitude is a learned predisposition to respond; that is, it is a learned mediating response (_g_). Thus one has to learn the attitude--the appropriate predisposition toward any given object. But once one has learned the attitude, he also must learn what response to make to it--that is, there is no innate relationship between the attitude and behavior; one still has to learn a behavioral response. Two people may learn to hold the same attitude toward a given stimulus; clearly, however, they may also learn to make different responses given the same learned attitude.
基本上,杜布的論點如下:態度是一種學習到的傾向去反應;也就是說,它是一種學習到的中介反應(_g_)。因此,人必須學習這種態度——對任何特定對象的適當傾向。但一旦學會了態度,他還必須學習對其做出什麼反應——也就是說,態度和行為之間並不存在天生的關係;人仍然需要學習行為反應。兩個人可能會學會對某一特定刺激持有相同的態度;然而,顯然他們也可能在面對相同的學習態度時做出不同的反應。
For example, two students may learn to feel equally favorable toward a given instructor. Furthermore, this feeling may initially elicit the same overt response in both students (e.g., calling the instructor by his first name). From Doob's point of view, the probability that this behavior will persist is a function of the reinforcement the students get for making this response. For example, with respect to one student, the instructor might say, "Well, I'm glad you finally decided to drop that 'professor' nonsense," while he might tell the other student, "I'd prefer it if you wouldn't call me by my first name." If this were the case, Doob's theory would predict that the behavior would continue for the first student, but not for the second.
例如,兩位學生可能會對某位教師產生同樣的好感。此外,這種感覺最初可能會在兩位學生中引發相同的明顯反應(例如,稱呼教師的名字)。從杜布的角度來看,這種行為持續的概率取決於學生因做出這種反應而獲得的強化。例如,對於一位學生,教師可能會說:「好吧,我很高興你終於決定不再叫我‘教授’這種無聊的稱呼了」,而對於另一位學生,他可能會告訴他:「我希望你不要叫我名字。」如果是這樣,杜布的理論將預測第一位學生的行為會持續,但第二位學生則不會。
Obviously, there may be some problems with this approach. More specifically, it is possible that the students' attitudes toward the instructor may also change differentially over time as a function of this reinforcement. However, even if this is the case, the main point is that Doob provided a reasonable partial answer, or at least an approach to an answer, to the problem of the relation between attitude and behavior. Doob was essentially arguing that attitude and behavior could be unrelated; that it was perfectly reasonable for two people with the same attitude to behave differently. Unfortunately, Doob's major point was not only not accepted, but it actually produced something akin to a "boomerang effect." That is, in Chein's (1948) critique of Doob's article, he essentially thanked Doob for pointing out what was wrong with our measures and definitions of attitude. Indeed, said Chein, Doob has pointed out that two people may feel the same amount of affect toward an object but may behave differently with respect to that object, or they may hold different beliefs about what should be done with respect to that object, or both. Clearly then, because the "action" component of attitude is different, these people must have different attitudes. Similarly, Chein continued, two people may be equally favorable toward the object, but they may also have different cognitions about the object; they may believe different things about it. Here again, these people must have different attitudes. An attitude, then, has several components. Now, this was not the first time that a multicomponent definition was advanced, nor was it the last. Indeed, if you pick up almost any social psychology textbook, you will find that attitude is currently defined as a concept containing an affective, a cognitive, and a conative component.
顯然,這種方法可能存在一些問題。更具體地說,學生對教師的態度可能會隨著這種強化而隨時間不同而改變。然而,即使如此,主要的觀點是杜布提供了一個合理的部分答案,或者至少是一種解答的途徑,來解決態度與行為之間的關係問題。杜布基本上在主張,態度和行為可能是無關的;兩個擁有相同態度的人行為不同是完全合理的。不幸的是,杜布的主要觀點不僅未被接受,還實際上產生了類似於「回力鏢效應」的結果。也就是說,在蔡恩(1948)對杜布文章的批評中,他基本上感謝杜布指出我們的態度測量和定義中的問題。事實上,蔡恩說,杜布指出兩個人可能對一個對象有相同的情感,但在對該對象的行為上可能有所不同,或者他們可能對應該如何處理該對象持有不同的信念,或者兩者皆是。 顯然,由於態度的「行動」成分不同,這些人必須擁有不同的態度。同樣,切因繼續說,兩個人可能對某個對象持相同的好感,但他們對該對象的認知可能不同;他們可能對它有不同的看法。在這裡,這些人必須擁有不同的態度。因此,態度有幾個成分。這並不是第一次提出多成分的定義,也不是最後一次。事實上,如果你翻閱幾乎任何一本社會心理學教科書,你會發現態度目前被定義為一個包含情感、認知和意向成分的概念。
In contrast to this, I prefer, following Thurstone (1981), to view attitude as a relatively simple unidimensional concept, referring to _"the amount of affect for or against a psychological object"_ (p. 261). Rather than viewing beliefs and behavioral intentions as a part of attitude, I prefer to define them independently
與此相反,我更傾向於遵循瑟斯通(Thurstone, 1981)的觀點,將態度視為一個相對簡單的單維概念,指的是「對某一心理對象的情感量」(第 261 頁)。我更願意獨立定義信念和行為意圖,而不是將它們視為態度的一部分。
---
and to view them as phenomena that are related to attitudes. More specifically, I see beliefs and behavioral intentions as determinants or consequents of an individual's attitude. As Green (1954) has pointed out, the concept of attitude is a hypothetical variable, abstracted from the many statements and actions that an individual makes with respect to a given object. Thus, rather than viewing statements about the object (i.e., beliefs) and statements about actions that one would take with respect to the object (i.e., behavioral intentions) as part of attitude or as the attitude per se, I feel that these statements can best be viewed as indicators of an individual's attitude.
並將它們視為與態度相關的現象。更具體地說,我將信念和行為意圖視為個人態度的決定因素或結果。正如格林(1954)所指出的,態度的概念是一個假設變量,從個體對特定對象所做的許多陳述和行動中抽象出來。因此,我認為與對象有關的陳述(即信念)和個體對該對象所採取的行動的陳述(即行為意圖)並不是態度的一部分或態度本身,而是可以最好地視為個人態度的指標。
One argument for the unidimensional viewpoint is a pragmatic one. If one accepts a multidimensional view of attitude, this implies that the "attitude" of any one person toward an object may fall at three very different positions on three different dimensions. Nevertheless, the operations by which attitudes are measured almost invariably yield a single score that is unlikely to reflect these three different components in any very precise fashion. As a matter of fact, people who construct "attitude scales" rarely maintain that their instruments are measuring three components; instead, they usually contend that their scales indicate people's evaluations of, or affect toward, an object or concept. Thus, although attitudes are often said to include all three components, it is usually only evaluation or "the affective component" that is measured and treated by researchers as the essence of attitude.
對於單維觀點的一個論點是務實的。如果接受態度的多維觀點,這意味著任何一個人對某個對象的「態度」可能在三個不同的維度上落在三個非常不同的位置。然而,測量態度的操作幾乎總是產生一個單一的分數,這不太可能精確地反映這三個不同的組成部分。事實上,構建「態度量表」的人很少主張他們的工具在測量三個組成部分;相反,他們通常主張他們的量表顯示人們對某個對象或概念的評價或情感。因此,儘管態度通常被認為包括這三個組成部分,但研究人員通常只測量和處理評價或「情感組成部分」,並將其視為態度的本質。
Furthermore, there is considerable evidence showing that this single "affective" score is highly related to an individual's beliefs about the object. The research of Rosenberg (1956, 1960), Zajonc (1954), Fishbein (1963, 1965a, 1965b), and others has demonstrated that an individual's attitude (or affect) toward any object is a function of his beliefs about the object (i.e., the probability or improbability that the object is related to some other object, value, concept, or goal) and the evaluative aspects of those beliefs (i.e., the evaluation of--or attitude toward--the "related concept").
此外,有相當多的證據顯示,這個單一的「情感」分數與個體對該物體的信念高度相關。羅森伯格(1956, 1960)、扎永茨(1954)、費什賓(1963, 1965a, 1965b)等人的研究表明,個體對任何物體的態度(或情感)是其對該物體信念的函數(即該物體與其他物體、價值、概念或目標相關的可能性或不可能性)以及這些信念的評價方面(即對「相關概念」的評價或態度)。
In addition, as Fishbein (see pp. 257 to 266 in this book) has pointed out, a consideration of most standardized instruments for measuring attitudes will also demonstrate that the single "affective" score they obtain is in fact derived from a consideration of a subject's beliefs and the evaluative aspects of those beliefs. For example, in Thurstone Scaling and Likert Scaling the subject is confronted with a series of belief statements. In both cases, the attitude score is indexed from a consideration of the respondent's beliefs (i.e., his agreement or disagreement with each of the statements) or, as Green suggests, it is abstracted from several of his statements about the attitude object.
此外,正如 Fishbein(見本書第 257 至 266 頁)所指出的,考慮到大多數標準化的態度測量工具也會顯示出,他們所獲得的單一「情感」分數實際上是基於對受試者信念及其評價方面的考量。例如,在 Thurstone 量表和 Likert 量表中,受試者面對一系列信念陳述。在這兩種情況下,態度分數是根據受訪者的信念(即他對每個陳述的同意或不同意)來指標化的,或者如 Green 所建議的,則是從他對態度對象的幾個陳述中抽象出來的。
Similarly, the single "affective" score obtained from Bogardus' Social Distance Scales is based on a consideration of an individual's behavioral intentions and their evaluative aspects.
同樣地,從博加德斯的社會距離量表中獲得的單一「情感」分數是基於對個體行為意圖及其評價方面的考量。
Thus, from my point of view, this hypothetical variable that we call an "attitude" can be measured by considering either beliefs or behavioral intentions, or by attempting to get at evaluation per se (e.g., through the use of such instruments as the Semantic ! Differential). It seems that these different types of instruments, or approaches, are attempting to measure the same thing; each is attempting to arrive at a single score that will represent how favorable or unfavorable the individual is toward the attitude object in question.
因此,從我的角度來看,我們所稱的“態度”這個假設變數可以通過考慮信念或行為意圖來衡量,或者通過試圖直接獲得評價(例如,使用語義差異法等工具)。這些不同類型的工具或方法似乎都在試圖衡量同一件事;每一種都試圖得出一個單一的分數,以代表個體對於相關態度對象的好感或不滿。
## The relationships between attitude and other phenomena
## 態度與其他現象之間的關係
Taking a unidimensional view of attitude does not imply that one should ignore cognition and conation. Rather, it implies that beliefs and behavioral intentions must be studied in their own right, as independent phenomena that may be related to attitude and behavior. Thus the problem is not simply to investigate relationships between attitude and behavior; rather, one must be concerned with at least four things: attitudes, beliefs, behavioral intentions, and behavior. The problem, then, is to investigate the interrelations among all four of these concepts.
採取單一維度的態度觀並不意味著應該忽視認知和意向。相反,它意味著信念和行為意圖必須作為獨立的現象進行研究,這些現象可能與態度和行為相關。因此,問題不僅僅是調查態度與行為之間的關係;相反,必須關注至少四個方面:態度、信念、行為意圖和行為。因此,問題在於調查這四個概念之間的相互關係。
Although considerable progress has been made in understanding the relationships between beliefs and attitudes, little has been done in investigating the relationships between beliefs and behavioral intentions, attitudes and behavioral intentions, or any of these three and behavior per se. In the remainder of this paper, I should like to consider some of these interrelationships.
儘管在理解信念與態度之間的關係方面已取得相當大的進展,但在研究信念與行為意圖、態度與行為意圖,或這三者與行為本身之間的關係方面卻所做的工作不多。在本文的其餘部分,我想考慮這些相互關係中的一些。
### The Relations between Beliefs and Attitudes (That Is, Cognition and Affect)
### 信念與態度之間的關係(即認知與情感)
As already mentioned, it appears that we now have fairly conclusive evidence that a person's attitude toward any object can be seen as a
如前所述,我們現在似乎有相當確鑿的證據表明,一個人對任何物體的態度可以被視為一個
---
### Attitude Theory
### 態度理論
function of his beliefs about the object (i.e., the probability or improbability that the object has a specific relationship with some other object, value, concept, or goal) and the evaluative aspects of those beliefs (i.e., the subject's attitude toward, or evaluation of, the "related object").1
他對該物體的信念的功能(即該物體與其他物體、價值、概念或目標之間存在特定關係的概率或不可能性)以及這些信念的評價方面(即主體對“相關物體”的態度或評價)。1
Footnote 1: The _relative_ of the object is the relative of the object to the object, and the relative of the object to the object is the relative of the object to the object.
腳註 1:物件的_相對_是物件對物件的相對,而物件對物件的相對是物件對物件的相對。
It should be noted, however, that although this evidence provides strong support for the general hypothesis that an individual's attitude toward any object is a function of his beliefs about the object and their evaluative aspects, it does _not_ imply that any given belief will be correlated with the attitude. Indeed, this relationship between beliefs and attitudes indicates that it is quite probable that an individual will have some beliefs that appear "inconsistent" with his attitude. As Fishbein ipp. 257 to 266 in this book) has pointed out, "Although each belief suggests an attitude, the attitude per se can only be reliably abstracted by considering the many beliefs an individual holds" (p. 264). Thus, while an individual's attitude will be highly correlated with an estimate based on a consideration of many of his beliefs, it may be uncorrelated or even negatively correlated with any single belief considered in isolation. Although the implications of this for the various "consistency theories" are beyond the scope of the present paper, it should be noted that this view suggests that it is not necessarily "inconsistent" for an individual (\(a_{i}^{*}\) to have a favorable attitude toward some "'object" and (\(b\)) to believe that the "object" has some negative characteristics, qualities, or attributes.
然而,應該注意的是,儘管這些證據強烈支持一般假設,即個體對任何對象的態度是其對該對象及其評價方面的信念的函數,但這並不意味著任何特定的信念與態度之間會有相關性。事實上,信念與態度之間的這種關係表明,個體很可能會有一些信念與其態度看起來“矛盾”。正如 Fishbein(本書第 257 至 266 頁)所指出的,“雖然每個信念都暗示著一種態度,但態度本身只能通過考慮個體持有的多個信念來可靠地抽象出來”(第 264 頁)。因此,雖然個體的態度與基於考慮其多個信念的估計高度相關,但它可能與任何單一信念在孤立考慮時無相關性,甚至呈負相關。 儘管這對各種「一致性理論」的含義超出了本文的範疇,但應該注意到,這種觀點表明,個體(\(a_{i}^{*}\))對某個「對象」持有積極態度與(\(b\))認為該「對象」具有某些負面特徵、品質或屬性並不一定是「不一致的」。
In addition to its implications for "consistency theories," the strong relationship between beliefs and attitudes also suggests one major reason why many investigators have failed to find relationships between attitudes and behavior: often the attitude being measured may be inappropriate. For example, in many cases an individual's attitude toward a class of people or objects has been measured, and this attitude has been used in an attempt to predict the individual's behavior with respect to a particular member of that class.
除了對「一致性理論」的影響外,信念與態度之間的強關係還暗示了許多研究者未能發現態度與行為之間關係的一個主要原因:所測量的態度往往可能不合適。例如,在許多情況下,個體對某一類人或物的態度被測量,而這種態度被用來試圖預測個體對該類別中特定成員的行為。
Thus one frequently measures a subject's attitude toward "Negroes," and then attempts to predict his behavior with respect to a particular Negro individual. However, it is very unlikely that the subject's beliefs about this particular Negro individual are even similar to his beliefs about "Negroes in general." Clearly, if a subject were asked to describe a "Negro," he would give a very different set of responses than if he were asked to describe Martin Luther King or Cassius Clay. Because his beliefs about these stimuli would differ, his attitudes toward them would also differ. Indeed, probably the only reason why even minimal correlations between attitude and behavior have been obtained in some cases is the fact that one belief about the particular stimulus person is that "he is a Negro"; this belief, being quite strong, will contribute heavily to the attitude. However, even this may not occur with respect to other stimulus persons who cannot be readily identified as members of a given group (e.g., Jews, Catholics, etc.). It seems fairly obvious that the chances of predicting behavior from attitude are practically nil until we at least start measuring attitudes toward the appropriate stimulus.
因此,人們經常測量一個人對「黑人」的態度,然後試圖預測他對某個特定黑人個體的行為。然而,這個人對這個特定黑人個體的信念與他對「黑人一般」的信念幾乎不可能相似。顯然,如果問這個人如何描述一個「黑人」,他的回答將與問他如何描述馬丁·路德·金或卡西烏斯·克雷時的回答截然不同。因為他對這些刺激的信念會有所不同,所以他對它們的態度也會不同。事實上,甚至在某些情況下,態度與行為之間僅有的微弱相關性,可能僅僅是因為對特定刺激個體的信念是「他是一個黑人」;這種信念相當強烈,將對態度產生重大影響。然而,對於那些無法輕易識別為某一特定群體成員的其他刺激個體(例如猶太人、天主教徒等),這種情況可能並不會發生。 似乎相當明顯,從態度預測行為的機會幾乎為零,直到我們至少開始測量對適當刺激的態度。
The problem often becomes even more complex, however. For example, we are often concerned with predicting different types of sociometric choices. Thus we may try to predict which group members a particular member will choose as friends or as co-workers. Bales' (1958) research has provided strong evidence that the person we may like the best and the person we would most like to work with are usually two different people. This implies that our attitude toward a person as a co-worker may be quite different from our attitude toward the same person as a friend. This should not be surprising; if an individual were asked to describe a particular person as a card player, he would give a different set of beliefs about that person than if he were asked to describe the same person as a co-worker. Thus, as a minimal initial step in trying to understand the relationships between attitudes and behavior, we must first start measuring attitudes toward the appropriate stimuli, vis-a-vis the behaviors we are attempting to predict.
然而,問題往往變得更加複雜。例如,我們經常關心預測不同類型的社會選擇。因此,我們可能會試圖預測某個成員會選擇哪些小組成員作為朋友或同事。貝爾斯(Bales,1958)的研究提供了強有力的證據,表明我們最喜歡的人和我們最想與之合作的人通常是兩個不同的人。這意味著我們對某個人作為同事的態度可能與我們對同一個人作為朋友的態度截然不同。這並不令人驚訝;如果要求一個人描述某個人作為撲克玩家,他會給出一套與描述同一個人作為同事時不同的信念。因此,作為理解態度與行為之間關係的最初步驟,我們必須首先開始測量對適當刺激的態度,與我們試圖預測的行為相對應。
### The Relations between Attitude and Behavioral Intentions (That Is, Affect and Conation)
### 態度與行為意圖之間的關係(即情感與意志)
In a recent series of papers, Triandis (see pp. 208 to 219 in this book) and his associates (e.g., Triandis and Davis, 1964; Fishbein, 1964; Triandis, Fishbein, Hall, Tanaka, and Shanmugan, 1967) have been exploring the "behavioral" or "conative" component of attitude. In the first paper in this series, Triandis developed an instrument called the Behavioral Differential.
在最近的一系列論文中,Triandis(見本書第 208 至 219 頁)及其同事(例如,Triandis 和 Davis,1964;Fishbein,1964;Triandis、Fishbein、Hall、Tanaka 和 Shanmugan,1967)一直在探討態度的「行為」或「意向」成分。在這系列的第一篇論文中,Triandis 開發了一種名為行為差異的工具。
---
## Attitude and the prediction of behavior
## 態度與行為的預測
This instrument consists of a series of behavioral statements (e.g., "accept as an intimate friend," "go fishing with," "admire the ideas of," etc.) similar to those found on Bogardus' (1925) Social Distance Scales. The respondent is asked to indicate (usually on a nine-place scale) the extent to which he "would" or "would not" engage in these behaviors with respect to a given stimulus person. Starting with a large set of behaviors obtained from a content analysis of novels, Triandis initially selected behavioral items on the basis of a facet analysis, and further reduced the number of items through factor analyses. Generally speaking, factor analyses of a large number of item-by-item intercorrelation matrices (obtained in several different studies, utilizing different behavioral items and different stimuli) have indicated five basic clusters of items. That is, items always seem to cluster together into the five types shown in Table 1.
這個工具由一系列行為陳述組成(例如,「接受為親密朋友」、「一起釣魚」、「欣賞某人的想法」等),類似於博加杜斯(1925)社會距離量表中的內容。受訪者被要求指示(通常在九點量表上)他「會」或「不會」在與特定刺激對象相關的情況下參與這些行為。最初,Triandis 從對小說的內容分析中獲得了一大組行為,並根據面向分析選擇了行為項目,然後通過因素分析進一步減少項目數量。一般來說,對大量項目之間的相關矩陣(在幾個不同的研究中獲得,使用不同的行為項目和不同的刺激)進行的因素分析顯示出五個基本的項目集群。也就是說,這些項目似乎總是聚集成表 1 中顯示的五種類型。
It should be noted that these five types of intentions have not been referred to as different dimensions of behavioral intentions, but only as clusters or types. Depending largely on the types of stimulus persons that are rated, these different clusters tend to go together in different ways (i.e., they may load on the same or different dimensions in any given factor analysis). For example, in one study, admiration and friendship items may load on one factor, while social distance and subordination intentions may load on a second factor. In another study, admiration and subordination intentions may load on one factor, while friendship and social distance intentions may load on a second factor. Thus, although Triandis has essentially identified five different general types of behavioral intentions, it should be made clear that these five types are usually not independent and, furthermore, that their relations to one another depend largely on the type of stimulus person being rated.
應注意的是,這五種類型的意圖並未被稱為行為意圖的不同維度,而僅僅被視為群集或類型。根據被評估的刺激對象類型的不同,這些不同的群集往往以不同的方式相互關聯(即,在任何給定的因素分析中,它們可能在同一或不同的維度上加載)。例如,在一項研究中,欽佩和友誼項目可能加載在一個因素上,而社交距離和從屬意圖可能加載在第二個因素上。在另一項研究中,欽佩和從屬意圖可能加載在一個因素上,而友誼和社交距離意圖可能加載在第二個因素上。因此,儘管 Triandis 本質上已經確定了五種不同的一般行為意圖類型,但應明確指出這五種類型通常不是獨立的,此外,它們之間的關係在很大程度上取決於被評估的刺激對象類型。
Turning to the question of the relations between these five types of behavioral intentions and attitude (i.e., affect), it should first be noted that because these behavioral intentions intercorrelate differentially with one another, they also are differentially related to attitude. That is, in one study attitude may be more highly correlated with friendship intentions than with social distance intentions, while in another study the findings may be reversed. Generally speaking, however, the results of several studies seem to indicate that attitudes are most highly correlated with admiration and friendship intentions (\(\overset{\rightarrow}{r}=.55\)), are moderately correlated with subordination and social distance intentions (\(\overset{\rightarrow}{r}=.35\)), and are least correlated with marital intentions (\(\overset{\rightarrow}{r}=.15\)).
轉向這五種類型的行為意圖與態度(即情感)之間的關係問題,首先應注意到,由於這些行為意圖之間的相互關聯性不同,它們與態度的關係也有所不同。也就是說,在一項研究中,態度可能與友誼意圖的相關性高於與社交距離意圖的相關性,而在另一項研究中,結果可能會相反。然而,總體而言,幾項研究的結果似乎表明,態度與欽佩和友誼意圖的相關性最高(\(\overset{\rightarrow}{r}=.55\)),與隸屬和社交距離意圖的相關性中等(\(\overset{\rightarrow}{r}=.35\)),而與婚姻意圖的相關性最低(\(\overset{\rightarrow}{r}=.15\))。
While the correlations between attitude and the different types of behavioral intentions vary considerably, the correlation between attitude and the sum of the behavioral intentions tends
雖然態度與不同類型的行為意圖之間的相關性差異相當大,但態度與行為意圖總和之間的相關性傾向於
\begin{table}
\begin{tabular}{c c c} Cluster & Typical Items \\
1. & Marital & I would: & go on a date with \\ & & fall in love with \\ & & marry \\
2. & Admiration & I would: & admire character of \\ & & believe \\ & & admire idea of \\ & & praise suggestions \\
3. & Social distance & I would: & invite to my club \\ & & exclude from my neighborhood \\ & & accept as a close kin by marriage \\
4. & Friendship & I would: & accept as an intimate friend \\ & & treat as equal \\
5. & Subordination-superordination & I would: & be commanded by \\ & & elect to political office \\ & & treat as a subordinate \\ & & work for \\ \end{tabular}
\end{table}
Table 1: **Item Clusters in Behavioral Differential**
\begin{table} \begin{tabular}{c c c} 群集 & 典型項目 \\ 1. & 婚姻 & 我會: & 約會 \\ & & 陷入愛河 \\ & & 結婚 \\ 2. & 欽佩 & 我會: & 欽佩的性格 \\ & & 相信 \\ & & 欽佩的想法 \\ & & 讚美建議 \\ 3. & 社交距離 & 我會: & 邀請加入我的俱樂部 \\ & & 排除在我的社區之外 \\ & & 接受作為親密的姻親 \\ 4. & 友誼 & 我會: & 接受為親密朋友 \\ & & 平等對待 \\ 5. & 從屬-超屬 & 我會: & 被命令 \\ & & 競選公職 \\ & & 將其視為下屬 \\ & & 為其工作 \\ \end{tabular} \end{table} 表 1: **行為差異中的項目群集**
---
## 482 Attitude Theory
## 482 態度理論
to be quite stable and high (\(\overset{\sim}{r}=.70\)). Thus, just as any given belief may be uncorrelated with an individual's attitude, any given behavioral intention (e.g., "I would elect this person to political office") or any one type of behavioral intention (e.g., subordination-superordination intentions) may also be uncorrelated with an individual's attitude. However, when one considers the totality (or at least a large set) of an individual's behavioral intentions, good estimates of attitude can be obtained. We shall return to this point shortly; first, however, it is worth noting that it is with respect to behavioral intentions (and behavior per se) that Doob's argument seems most appropriate.
要相當穩定且高(\(\overset{\sim}{r}=.70\))。因此,就像任何特定的信念可能與個體的態度無關,任何特定的行為意圖(例如,「我會選這個人擔任政治職位」)或任何一種類型的行為意圖(例如,隸屬-超屬意圖)也可能與個體的態度無關。然而,當考慮到個體行為意圖的整體(或至少是一個大集合)時,可以獲得對態度的良好估計。我們將很快回到這一點;不過,首先值得注意的是,正是關於行為意圖(以及行為本身),杜布的論點似乎最為恰當。
More specifically, if we consider what has been happening in the South, we can get a clear picture of what Doob was referring to. That is, the Civil Rights Act of 1964 and various judicial decisions have had a profound effect on some behavioral intentions and behaviors of many Southerners. Many individuals in the South, if asked today, would say that they would admit Negroes to their hotels, motels, or restaurants. (And, indeed, Negroes have been admitted to various places that previously had refused to serve them.) I doubt very much, however, if any of us would really want to say that these individuals have thereby changed their attitudes toward Negroes.
更具體地說,如果我們考慮南方所發生的事情,我們可以清楚地了解杜布所指的內容。也就是說,1964 年的民權法案和各種司法裁決對許多南方人的某些行為意圖和行為產生了深遠的影響。如果今天問南方的許多人,他們會說他們會接納黑人進入他們的酒店、汽車旅館或餐廳。(事實上,黑人已被接納進入以前拒絕服務他們的各種場所。)然而,我非常懷疑我們中的任何人是否真的想說這些人因此改變了他們對黑人的態度。
As Doob pointed out, however, behavioral intentions (as well as overt behaviors) can be either positively or negatively reinforced. If our initial behavioral intentions or behaviors were always positively reinforced, it is likely that there would always be a clear relationship between attitude and behavioral intentions. Behavioral intentions are not always positively reinforced; indeed, they are often negatively reinforced. Furthermore, we are often positively reinforced for having behavioral intentions that are inconsistent with our attitudes. Once again, I think we can all see many examples of this in the South today.
然而,正如杜布所指出的,行為意圖(以及明顯的行為)可以受到正向或負向的強化。如果我們最初的行為意圖或行為總是受到正向強化,那麼態度與行為意圖之間很可能會始終存在明確的關係。行為意圖並不總是受到正向強化;事實上,它們經常受到負向強化。此外,我們經常因擁有與我們的態度不一致的行為意圖而受到正向強化。我再次認為,我們都能在今天的南方看到許多這方面的例子。
Before going any further, one point should be made clear. Namely, although everything that has been said about the relationship between attitude and behavioral intentions is also true of the relationship between attitude and behavior per se, this _does not_ imply that there is any given relationship between behavioral intentions and behavior. As we shall see below, the size of the correlation between a behavioral intention and behavior is determined largely by the specificity of the behavioral intention that is being considered. Indeed, everything that has been said about the relationships between attitudes and behavioral intentions may also apply to the relationships between Triandis' five types of behavioral intentions and overt behaviors.
在進一步討論之前,有一點需要澄清。也就是說,雖然關於態度與行為意圖之間的關係所說的一切,同樣適用於態度與行為本身之間的關係,但這並不意味著行為意圖與行為之間存在任何特定的關係。正如我們將在下面看到的,行為意圖與行為之間的相關性大小在很大程度上取決於所考慮的行為意圖的具體性。事實上,關於態度與行為意圖之間關係的所有論述,也可能適用於 Triandis 的五種類型的行為意圖與明顯行為之間的關係。
This is important, for I do not think a change in a behavioral intention leads necessarily to a change in attitude unless the change in behavioral intention is accompanied by a change in behavior per se. That is, a change in behavior implies a new set of relationships between the individual and the attitude object. This may then lead to the learning of new beliefs about the attitude object, and thus to attitude change. These beliefs will not be learned overnight, however, and the change in attitude may lag well behind the change in behavior.
這是重要的,因為我認為行為意圖的改變不一定會導致態度的改變,除非行為意圖的改變伴隨著行為本身的改變。也就是說,行為的改變意味著個體與態度對象之間關係的新組合。這可能會導致對態度對象的新信念的學習,從而導致態度的改變。然而,這些信念不會在一夜之間學會,態度的改變可能會明顯滯後於行為的改變。
Thus, although an individual's attitude might initially influence and be related to specific behavioral intentions (or to specific behaviors), this relationship may or may not persist, depending on the nature and schedule of reinforcement associated with the behavioral intentions. Furthermore, unlike a change in belief that immediately produces an attitude change, changes in a behavioral intention may not produce a change in attitude unless the change in behavioral intention eventually leads to changes in behavior and thus changes in belief about the attitude object.
因此,儘管個體的態度最初可能影響並與特定的行為意圖(或特定行為)相關,但這種關係可能會持續或不持續,這取決於與行為意圖相關的增強的性質和時間表。此外,與立即產生態度變化的信念變化不同,行為意圖的變化可能不會產生態度的變化,除非行為意圖的變化最終導致行為的變化,從而改變對態度對象的信念。
Therefore, rather than being viewed as parts of attitude, behavioral intentions should be viewed as independent phenomena. Just as a consideration of beliefs (or the cognitive component) yields a single score that indicates how favorable or unfavorable an individual is with respect to the attitude object, so, too, does a consideration of behavioral intentions (or the conative component). That is, as was mentioned earlier, an individual's attitude toward any object may be assessed by measuring affect or evaluation per se, by considering his beliefs or his behavioral intentions, or some combination of all of these.
因此,行為意圖應被視為獨立的現象,而不是態度的一部分。正如對信念(或認知成分)的考量產生一個單一的分數,顯示個體對態度對象的好惡程度,對行為意圖(或意向成分)的考量也是如此。也就是說,如前所述,個體對任何對象的態度可以通過測量情感或評價本身、考量其信念或行為意圖,或這些的某種組合來評估。
It appears, then, that the multicomponent view of attitude is misleading and inappropriate. Indeed, in contrast to the major criticism leveled at the unidimensional view of attitude, this single "affective" score does, in a sense, take into account and reflect an individual's beliefs and his behavioral intentions. Although two people may have the same attitudes, how
看來,多元組件的態度觀點是誤導且不恰當的。事實上,與對單維度態度觀點的主要批評相對比,這個單一的「情感」分數在某種程度上確實考慮並反映了一個人的信念和行為意圖。儘管兩個人可能擁有相同的態度,但如何
---
ever, these attitudes may be based on different beliefs or may be related to different behavioral intentions. As the multicomponent theorists have pointed out, this may well be one of the major reasons why attitudes are uncorrelated with behavior. But here I would argue that it is not because the attitude measure is inappropriate or incomplete. Rather, it is because attitude is a hypothetical variable abstracted from the _totality_ of an individual's beliefs, behavioral intentions, and actions toward a given object. Any given belief, behavioral intention, or behavior, therefore, may be uncorrelated or even negatively correlated with his attitude. Thus, rather than viewing specific beliefs or classes of beliefs and specific behavioral intentions or types of behavioral intentions as part of attitude, these phenomena must be studied as variables in their own right, which, 'like attitudes, may or may not function as determinants of a specific behavior.
這些態度可能基於不同的信念,或與不同的行為意圖有關。正如多元組件理論家所指出的,這可能是態度與行為不相關的主要原因之一。但在這裡我想辯稱,這並不是因為態度測量不恰當或不完整。相反,這是因為態度是一個假設變數,抽象自個體對特定對象的信念、行為意圖和行動的「整體」。因此,任何特定的信念、行為意圖或行為,可能與他的態度不相關,甚至呈負相關。因此,與其將特定的信念或信念類別以及特定的行為意圖或行為意圖類型視為態度的一部分,這些現象必須作為獨立的變數進行研究,這些變數「像態度一樣,可能或可能不會作為特定行為的決定因素」。
## The Relationships between Attitude and Behavior (That Is, Affect and Behavior)
## 態度與行為之間的關係(即情感與行為)
On this note we may finally turn our full attention to the presumed topic of this paper. First, however, I want to make it clear that when I use the term "attitude," I am simply referring to a learned predisposition to respond to any object in a consistently favorable or unfavorable way. Operationally, I do not care whether we have obtained our estimate of attitude through a consideration of affect per se, through a consideration of beliefs and their evaluative aspects, or through a consideration of behavioral intentions and their evaluative aspects. That is, I am talking about a single score that places an individual along a continuum ranging from favorableness toward the attitude object to unfavorableness. Let us further assume that this measure is reliable and valid (at least insofar as tapping a true underlying dimension of favorableness or unfavorableness toward the attitude object is concerned).
在這一點上,我們終於可以將全部注意力轉向本文的假定主題。然而,首先我想澄清的是,當我使用“態度”這個術語時,我只是指對任何對象以一致的正面或負面方式作出反應的學習性傾向。在操作上,我不在乎我們是通過對情感本身的考量、對信念及其評價方面的考量,還是通過對行為意圖及其評價方面的考量來獲得我們的態度估計。也就是說,我所談論的是一個單一的分數,將個體置於一個從對態度對象的正面到負面的連續體上。讓我們進一步假設這個測量是可靠和有效的(至少在觸及對態度對象的真正正面或負面基礎維度方面是如此)。
Thus far, I have indicated two possible reasons for our failure to predict behavior from attitude:
到目前為止,我已經指出了我們無法從態度預測行為的兩個可能原因:
1. We have often measured attitude toward an inappropriate stimulus object--for example, we have often measured attitude toward a class of people or objects when we should have been measuring attitude toward a particular member of the class.
2. The particular behavior being studied may be completely or partially unrelated to attitude. This point must be emphasized because most investigators of attitude have been unwilling to accept it. To a large extent, however, their unwillingness to accept this statement is surprising since most investigators firmly believe that any behavior is determined by a large number of variables. Yet time and again, a behavior is investigated because the experimenter assumes that it should be a function of attitude and then he is surprised and disappointed to find that his measure of attitude failed to predict the behavior.
1. 我們經常測量對不適當刺激對象的態度——例如,我們經常測量對一類人或物體的態度,而應該測量對該類中特定成員的態度。2. 正在研究的特定行為可能與態度完全或部分無關。這一點必須強調,因為大多數態度研究者不願意接受它。然而,在很大程度上,他們不願意接受這一說法是令人驚訝的,因為大多數研究者堅信任何行為都是由大量變數決定的。然而,一次又一次地,行為被研究是因為實驗者假設它應該是態度的函數,然後他驚訝和失望地發現他的態度測量未能預測該行為。
Unfortunately, one of the most common responses to this failure has been to question the attitude measure; it is criticized for measuring only affect and not taking into account the so-called cognitive or conative dimensions of attitude. What one should be questioning however, is the initial assumption that the behavior being considered is a function of the attitude being measured. Furthermore, even if measures of specific sets of beliefs or types of behavioral intentions were available, the chances would still be high that the behavior could not be predicted. That is, it must be realized that the appeal to beliefs and behavioral intentions may be based on the same false assumption that led to the expectation of an attitude-behavior relationship --namely, the assumption that the behavior being investigated is a function of some type of belief or behavioral intention.
不幸的是,對於這一失敗,最常見的反應之一是質疑態度測量;它被批評僅測量情感,而未考慮所謂的認知或意向維度。然而,應該質疑的是最初的假設,即所考慮的行為是所測量的態度的函數。此外,即使有特定信念或行為意圖類型的測量可用,行為仍然無法預測的可能性依然很高。也就是說,必須意識到對信念和行為意圖的訴求可能基於同樣的錯誤假設,這一假設導致了對態度-行為關係的期望——即假設所調查的行為是某種信念或行為意圖的函數。
Granted, this assumption may be true in some cases, in most cases however, it is quite likely that the particular beliefs or behavioral intentions we would choose to measure would themselves be highly correlated with attitude. Thus they would add little, if anything, to our predictive ability. The only times when it may be useful to consider beliefs and behavioral intentions (in addition to attitude) are those situations in which the particular beliefs or behavioral intentions that we wish to consider are not themselves highly correlated with the individual's attitude. However, even in those cases we must be willing to recognize that the beliefs or behavioral intentions that we do select may also be unrelated to the behavior itself. That is, just as we must be willing to accept the fact that two people with the same attitudes may behave differently, so, too, must we be willing to accept the fact that two people with the same beliefs or two people with the same behavioral intentions may also behave quite differently.
雖然這個假設在某些情況下可能是正確的,但在大多數情況下,我們所選擇測量的特定信念或行為意圖很可能與態度高度相關。因此,它們對我們的預測能力幾乎沒有任何增益。唯一可能有用的情況是考慮信念和行為意圖(除了態度之外),當我們希望考慮的特定信念或行為意圖與個體的態度並不高度相關。然而,即使在這些情況下,我們也必須願意承認我們所選擇的信念或行為意圖可能與行為本身無關。也就是說,就像我們必須願意接受兩個擁有相同態度的人可能會有不同的行為一樣,我們也必須願意接受兩個擁有相同信念或相同行為意圖的人也可能會有相當不同的行為。
---
## 44 Attitude Theory
## 44 態度理論
### Behavioral Prediction and Measures of Behavior
### 行為預測與行為測量
If you will notice, however, I have stopped discussing the relations between attitudes and behavior and started discussing the general problem of behavioral prediction. Essentially what I have been pointing out is how we might predict behavior in those situations where attitude is not a relevant variable. More specifically, I have indicated that when most attitude researchers fail to find a relation between attitude and behavior, they usually suggest that if we were to take one or two other variables into account, we might be able to predict the behavior. Although this is reasonable if our primary concern is to predict _behavior_--and this is what I believe our primary concern really is --it is unfortunate that two of the variables most commonly suggested (i.e., belief and behavioral intention) are arrived at through a consideration of the predictor, rather than through a consideration of the behavior per se. Here, I think, is where our major problem lies: we psychologists have never really studied behavior per se. By this I mean that we have usually taken behavior as a _given_; to the best of my knowledge, we--at least in the attitude area--have seldom, if ever, subjected our behavioral criteria to the same rigorous analyses to which we subject our paper-and-pencil tests. Yet this is what we must do if we are to thoroughly understand the relations between attitudes and behavior (i.e., if we are ever going to predict the conditions under which, and the extent to which, behavior is determined by, or related to, attitudes--or to any other variable, for that matter). In the remainder of this section, I shall try to clarify this by discussing a hypothetical example.
然而,如果你注意到,我已經停止討論態度與行為之間的關係,並開始討論行為預測的一般問題。基本上,我所指出的是,在那些態度不是相關變數的情況下,我們如何預測行為。更具體地說,我已經指出,當大多數態度研究者未能找到態度與行為之間的關係時,他們通常會建議如果我們考慮一兩個其他變數,我們可能能夠預測行為。雖然如果我們的主要關注是預測_行為_,這是合理的——而這正是我相信我們的主要關注所在——但不幸的是,兩個最常被建議的變數(即信念和行為意圖)是通過考慮預測變數而得出的,而不是通過考慮行為本身。在這裡,我認為我們的主要問題就在於:我們心理學家從未真正研究過行為本身。 我所指的是,我們通常將行為視為一種_理所當然_;據我所知,至少在態度領域,我們很少,甚至從未,對我們的行為標準進行過與我們對紙筆測試所進行的同樣嚴格的分析。然而,如果我們要徹底理解態度與行為之間的關係(即,如果我們要預測行為在何種條件下以及在多大程度上受到態度或任何其他變數的影響),這正是我們必須做的。在本節的其餘部分,我將通過討論一個假設的例子來澄清這一點。
First, let me describe a very simple behavioral situation. A naive subject shows up for an experiment and is asked to wait in an adjoining room while the experimenter sets up the experiment. The adjoining room is fairly small, containing only four chairs, arranged so that they form the four corners of a square. Seated in one of the chairs is a confederate of the experimenter. When the naive subject enters the room, the confederate looks up and says, "Hello, my name is Robert Springer." The two people then remain in the room for a five-minute period.
首先,讓我描述一個非常簡單的行為情境。一位天真的受試者來到實驗室,並被要求在一個相鄰的房間等候,期間實驗者正在設置實驗。這個相鄰的房間相當小,裡面只有四把椅子,排列成一個正方形的四個角落。坐在其中一把椅子上的是實驗者的共謀者。當天真的受試者進入房間時,共謀者抬起頭來說:“你好,我的名字是羅伯特·斯普林格。”然後這兩個人就留在房間裡五分鐘。
Now let us suppose that, as the experimenter, I want to make an hypothesis about the subject's communicative behavior. Furthermore, let us assume that I have a highly valid and reliable instrument for measuring the number, duration, or type of communicative acts that the subject makes. Finally, let us also assume that I have a well-trained confederate who behaves in a highly standardized manner.
現在讓我們假設,作為實驗者,我想對受試者的交際行為提出一個假設。此外,讓我們假設我擁有一個高度有效且可靠的工具來測量受試者所進行的交際行為的數量、持續時間或類型。最後,讓我們也假設我有一位訓練有素的同夥,他的行為非常標準化。
Now, to return to my hypothesis, let us suppose I have a strong bunch that the more the subject likes the confederate, the more he will talk to him. Thus I have essentially made the assumption that the amount of the subject's communicative behavior is a function of his attitude. To put it another way, I have implicitly made the assumption that this behavior is a good indicator of the subject's attitude toward the confederate. That is, rather than viewing the behavior as a criterion, I can also view it as a measure of the variable I assume it reflects. In this case, then, I would essentially be estimating the subject's attitude toward the confederate through a consideration of his communicative behavior.
現在,回到我的假設,讓我們假設我有一個強烈的觀點,即受試者越喜歡同盟者,他就越會與他交談。因此,我基本上假設受試者的交際行為量是其態度的函數。換句話說,我隱含地假設這種行為是受試者對同盟者態度的良好指標。也就是說,我不僅可以將這種行為視為標準,還可以將其視為我假設它所反映的變量的衡量。在這種情況下,我基本上將通過考慮受試者的交際行為來估計他對同盟者的態度。
Once we start considering the behavioral measure as an estimate of attitude (or any other variable), we can start subjecting the behavior to the same types of critical analyses to which we subject our paper-and-pencil measures. For example, one of the most obvious questions that should be asked concerns the reliability of the behavior. That is, even in the simple situation I have described, we could bring the subject back at the same time on another day and obtain a measure of test-retest reliability. Even this simple test provides us with a considerable amount of new information; if the behavior is unreliable, that finding almost immediately suggests that this behavior is not a direct function of any variable that _is_ reliable or has not changed from Time 1 to Time 2. "Unreliability" provides a clear indication that the behavior is not determined directly by any of the controlled variables that are part of our experimental situation, whether they can be specified or not. More specifically, "unreliability" allows us to conclude that this behavior is neither situationally determined, nor solely a function of any factor that is consistently related to or based upon a consideration of the stimulus person, that is, the confederate.2
一旦我們開始將行為測量視為態度(或任何其他變數)的估計,我們就可以開始對行為進行與我們對紙筆測量所進行的相同類型的批判性分析。例如,應該提出的最明顯的問題之一是行為的可靠性。也就是說,即使在我所描述的簡單情況下,我們也可以在另一個日子的同一時間將受試者帶回來,並獲得測試-重測可靠性的測量。即使這個簡單的測試也為我們提供了相當多的新信息;如果行為不可靠,這一發現幾乎立即表明這種行為不是任何可靠變數的直接函數,或者從時間 1 到時間 2 並未改變。“不可靠性”清楚地表明,該行為並不是由我們實驗情境中任何受控變數直接決定的,無論這些變數是否可以被具體化。 更具體地說,「不可靠性」使我們得出結論,這種行為既不是由情境決定的,也不僅僅是任何與刺激對象(即共謀者)相關或基於考慮的因素的功能。2
Footnote 2: Obviously, a change in behavior could be a function of a change in attitude. This example, however, is only concerned with the case where attitude does not change.
腳註 2:顯然,行為的改變可能是態度改變的結果。然而,這個例子僅關注態度不變的情況。
If the behavior is shown to be reliable, this
如果行為被證明是可靠的,這
---
knowledge permits us to go on to the more important question of validity. Obviously, I am not talking about the validity of the behavior per se (for, by definition, behavior is "real"), but rather about the validity of the behavior as an indicant or measure of some other variable.
知識使我們能夠進一步探討更重要的有效性問題。顯然,我並不是在談論行為本身的有效性(因為根據定義,行為是「真實的」),而是關於行為作為某個其他變數的指標或衡量標準的有效性。
Here, rather than starting with the usual question of convergent validity (i.e, the extent to which two different estimates or measures of a given variable are correlated), I should like first to discuss briefly what Campbell and Fiske (1959) have called discriminant validity. According to these authors, a valid test should not only measure what it is supposed to measure, but also it should not measure anything else. To quote them, "tests can be invalidated by too high correlations with other tests from which they were intended to differ" (p. 81)."
在這裡,我想先簡要討論一下坎貝爾和費斯克(1959)所稱的區別效度,而不是從通常的收斂效度問題開始(即兩個不同的估計或測量在多大程度上與給定變量相關)。根據這些作者的說法,一個有效的測試不僅應該測量它應該測量的內容,還應該不測量其他任何東西。引用他們的話,“測試可能因與其他測試的相關性過高而失效,而這些測試本來是希望有所區別的”(第 81 頁)。
Furthermore, they point out that every test or task employed for measurement purposes should be viewed as a trait-method unit, that is, as "a union of a particular trait [or variable] content with measurement procedures not specific to that content." Thus, they continue, "The systematic variance among test scores can be due to responses to the measurement features as well as responses to the trait content" (p. 81)."
此外,他們指出,為了測量目的而使用的每個測試或任務都應被視為一個特徵-方法單位,即“特定特徵[或變量]內容與不特定於該內容的測量程序的結合。”因此,他們繼續說,“測試分數之間的系統變異可能是由於對測量特徵的反應以及對特徵內容的反應所致”(第 81 頁)。
To return to our simple situation, it can be seen that what I have described is a general, content-independent measurement procedure. Essentially, I have described a behavioral method for measuring a subject's attitude toward any given stimulus person. Variations in the characteristics or behavior of the stimulus person (or the situation per se) are analogous, then, to what Campbell and Fiske have called variations in trait content.
回到我們簡單的情況,可以看出我所描述的是一種一般性的、與內容無關的測量程序。基本上,我描述了一種行為方法,用於測量受試者對任何給定刺激對象的態度。因此,刺激對象的特徵或行為(或情境本身)的變化,類似於坎貝爾和費斯克所稱的特質內容的變化。
Continuing, Campbell and Fiske point out that "In order to examine discriminant validity, and in order to estimate the relative contributions of trait and method variance, _more than one trait as_ well as _more than one method_ must be employed in the validation process" (p. 81). Thus, in contrast to our usual validation technique of correlating some pencil-and-paper measure of attitude with one or more behavioral criteria (that we assume or hypothesize to be correlated with attitude), we should, _at a minimum_, be obtaining measures of attitude toward at least two stimulus objects and measuring the subject's behavior with respect to each of them. It is only in this way that we can start to obtain some indication of a more general pattern of relationships between attitude and behavior. That is, we must start to pay more attention to the question of discriminant validity if we ever hope to distinguish between those behaviors that are functions of the situation and those that are related to characteristics associated with the attitude object.
坎貝爾和費斯克指出:「為了檢驗區別效度,並估計特質和方法變異的相對貢獻,必須在驗證過程中使用不止一個特質以及不止一種方法」(第 81 頁)。因此,與我們通常的驗證技術,即將某些紙筆測量的態度與一個或多個行為標準進行相關(我們假設或假設這些標準與態度相關)相比,我們至少應該獲得對至少兩個刺激對象的態度測量,並測量受試者對每個對象的行為。只有這樣,我們才能開始獲得態度與行為之間更一般的關係模式的某些指示。也就是說,如果我們希望區分那些受情境影響的行為和那些與態度對象相關的特徵的行為,我們必須開始更加關注區別效度的問題。
Although a complete discussion of Campbell and Fiske's multitrait, multimethod matrix technique is beyond the scope of the present paper, it appears that this technique (with some modifications) provides a simple and straightforward procedure for beginning to analyze the relationship between attitude and behavior. Because Campbell and Fiske's original concerns were with the measurement of personality traits (which are usually assumed to be independent), at least one modification in interpretation should be mentioned. Specifically, the Campbell and Fiske analysis suggests that when the same method is used to measure two or more traits and the monomethod,' multitrait correlations are high, a large (and irrelevant) "methods" factor may be influencing the responses. This is not necessarily the case in attitude research, because subjects' attitudes toward two or more stimuli may often be highly intercorrelated.
儘管對坎貝爾和費斯克的多特質、多方法矩陣技術的完整討論超出了本文的範疇,但這一技術(經過一些修改)似乎提供了一個簡單明瞭的程序,用於開始分析態度與行為之間的關係。由於坎貝爾和費斯克最初關注的是人格特質的測量(通常假設這些特質是獨立的),因此至少應提及一項解釋上的修改。具體而言,坎貝爾和費斯克的分析表明,當使用相同的方法來測量兩個或更多特質時,如果單一方法的多特質相關性很高,則可能有一個大型(且無關的)“方法”因素在影響反應。在態度研究中情況不一定如此,因為受試者對兩個或更多刺激的態度往往可能高度相關。
While this relationship does not require change in any of the criteria that Campbell and Fiske propose for analyzing discriminant validity, it does suggest that one alternative criterion may be useful in considering discriminant validity in attitude research. Specifically, variations in the attitude object should produce reliable variation in the attitude scores. For example, although individuals' attitudes toward two stimuli may be highly correlated (e.g., authoritarian subjects' attitudes toward Negroes and Jews), the method used to measure these attitudes would have discriminant validity with respect to the two stimulus persons if, in addition to Campbell and Fiske's criteria, one found a reliable mean difference between the two attitude scores, that is, if the subjects' attitudes toward Negroes were significantly different from their attitudes toward Jews. A similar argument applies to the behavioral measures. This may be seen more clearly below.
雖然這種關係不需要改變 Campbell 和 Fiske 提出的任何用於分析區別效度的標準,但它確實暗示了一個替代標準在考慮態度研究中的區別效度時可能是有用的。具體而言,態度對象的變化應該能夠產生可靠的態度分數變化。例如,儘管個體對兩個刺激的態度可能高度相關(例如,威權主義者對黑人和猶太人的態度),但如果除了 Campbell 和 Fiske 的標準外,還發現兩個態度分數之間存在可靠的平均差異,即受試者對黑人的態度與對猶太人的態度顯著不同,那麼用於測量這些態度的方法將對這兩個刺激對象具有區別效度。類似的論點也適用於行為測量。這一點在下面會更清楚地顯示出來。
## The relationship of attitude and behavior as seen in a multimettitude object, multimethod matrix
## 在多元態度物件、多方法矩陣中觀察到的態度與行為的關係
To return to the main point, it appears that in those cases where the attitudinal and behavioral measures are uncorrelated (i.e., where there is no convergent validity), the multitrait
回到主要點,似乎在那些態度和行為測量不相關的情況下(即,沒有收斂效度),多特質
---
(or multiatitude object), multimethod approach provides a considerable amount of information that would not be obtained in other ways. This is not to say that one will not obtain convergent validity between some attitudes and some behaviors, nor that these findings are unimportant or uninformative. However, the Campbell and Fiske approach is probably most useful in respect to the further leads it provides in those cases where convergent validity is not obtained. This may perhaps be seen most clearly if we discuss some of the major cases that are possible.
(或多態物件),多方法方法提供了大量無法以其他方式獲得的信息。這並不是說某些態度和某些行為之間不會獲得收斂效度,也不是說這些發現不重要或不具信息性。然而,坎貝爾和費斯克的方法在收斂效度未獲得的情況下,可能最有用,因為它提供了進一步的線索。如果我們討論一些可能的主要案例,這一點或許能更清楚地表現出來。
### Case A: Discriminant Validity in the Attitudinal, but not in the Behavioral Measures
### 案例 A:態度測量中的區別效度,但行為測量中則不然
Suppose that in our hypothetical situation we find that, irrespective of variations in the characteristics or behaviors of our confederate, the behavior in this situation does not vary. That is, suppose the results indicate that the number of communicative acts do not vary whether the confederate is Negro or White, male or female, Catholic or Protestant, dressed neatly or sloppily; or whether the confederate is instructed to agree or disagree with the subject, to respond with long sentences or short phrases, etc. In other words, suppose we find that, at least with respect to the variables we are manipulating, our behavioral measure has no discriminant validity. Let us further suppose that, in contrast, our paper-and-pencil measure does reflect these differences, that is, that as we manipulate certain characteristics of the confederate, the subject's attitudes toward the confederate change. Thus our attitude measure does have discriminant validity. Finally, let us assume that this particular paper-and-pencil measure has a high degree of convergent validity with respect to other measures of attitude, that is, that it has construct validity.
假設在我們的假設情境中,我們發現無論我們的同盟者的特徵或行為如何變化,這種情況下的行為並不變化。也就是說,假設結果顯示,無論同盟者是黑人還是白人、男性還是女性、天主教徒還是新教徒、穿著整齊還是邋遢,溝通行為的數量都不變;或者無論同盟者被指示同意還是不同意受試者的觀點,回應長句還是短語等等。換句話說,假設我們發現,至少就我們正在操控的變數而言,我們的行為測量沒有區別效度。進一步假設,相對而言,我們的紙筆測量確實反映了這些差異,也就是說,當我們操控同盟者的某些特徵時,受試者對同盟者的態度會改變。因此,我們的態度測量確實具有區別效度。最後,假設這個特定的紙筆測量在態度的其他測量中具有高度的收斂效度,也就是說,它具有建構效度。
Clearly, then, our failure to obtain convergent validity with the behavioral measures in this situation does not imply that our measure of attitude is invalid; rather, it strongly indicates that this behavior is not sensitive to or not related to the attitude we have measured--nor to any variable that is likely to correlate with either the attitude or any of the characteristics or actions that were manipulated. Indeed, rather than concern ourselves with attitudes, we can be fairly certain that this behavior is either a function of some stable individual difference variable or some nonstimulus (i.e., nonconfederate) associated variable or variables in this situation.
顯然,我們在這種情況下未能獲得行為測量的收斂效度並不意味著我們的態度測量是無效的;相反,這強烈表明這種行為對我們所測量的態度不敏感或無關聯——也與任何可能與態度或任何被操控的特徵或行為相關的變數無關。事實上,我們不必過於關注態度,我們可以相當確定這種行為要麼是某種穩定的個體差異變數的函數,要麼是與此情況相關的某些非刺激(即非同盟者)變數。
### Case B: Discriminant Validity in the Behavioral, but not in the Attitudinal Measures
### 案例 B:行為測量中的區別效度,但在態度測量中則不然
Alternatively, of course, the lack of convergent validity could derive from the finding that variations in the stimulus object produce variations in the behavior but not in the paper-and-pencil measure. In th's case our behavioral measure would have discriminant validity, but, at least with respect to the particular variables we have manipulated, our paper and pencil measure would not. Furthermore, let us again assume that the particular paper-and-pencil test does have a high degree of construct validity. This is an interesting case, for here we have a change in behavior that is related to the manipulation of a specified variable but is not correlated with the attitude measured. Here again, I would argue that our failure to obtain convergent validity does not imply that our measure of attitude is invalid; it would only indicate that this behavior is not related to the attitude measured, nor to any variable that is likely to correlate with the attitude measured. Indeed, what we have found is that certain variables that function as determinants of behavior with respect to a given stimulus person are unrelated to attitude toward that person. Thus, if our main interest is in predicting the behavior, we should concentrate on measuring those particular variables rather than on measuring attitudes toward the person.
或者,當然,缺乏收斂效度可能源於刺激物的變化產生行為的變化,但在紙筆測量中卻沒有變化。在這種情況下,我們的行為測量將具有區別效度,但至少就我們操控的特定變數而言,我們的紙筆測量則不具此效度。此外,讓我們再次假設特定的紙筆測試確實具有高程度的建構效度。這是一個有趣的案例,因為在這裡,我們有一種與特定變數的操控相關的行為變化,但與所測量的態度並不相關。在這裡,我再次主張,我們未能獲得收斂效度並不意味著我們的態度測量是無效的;這僅僅表明這種行為與所測量的態度無關,也與任何可能與所測量的態度相關的變數無關。事實上,我們發現某些變數在對特定刺激對象的行為決定中起作用,但與對該對象的態度無關。 因此,如果我們的主要興趣在於預測行為,我們應該專注於測量那些特定的變數,而不是測量對個人的態度。
Before continuing, I should mention one other possible reason for not obtaining convergent validity in this case, namely, it could be that we have measured attitude toward an inappropriate stimulus object. For example, \(E\) measured attitudes toward "Catholics" and "Protestants" rather than toward the specific Catholic individual and the specific Protestant individual who served as confederates. The failure to find discriminant validity in our attitude measure argues against this, however. That is, inappropriate stimuli are likely to be fairly general or stereotypic, and because our instrument has construct validity, this type of error would increase, rather than decrease, the degree of discriminant validity. Although this type of error is not indicated in the present case, it is indicated in the final set of alternatives I shall discuss.
在繼續之前,我應該提到另一個可能導致在這種情況下未獲得收斂效度的原因,即我們可能測量了對不適當刺激對象的態度。例如,\(E\) 測量了對「天主教徒」和「新教徒」的態度,而不是對作為同夥的特定天主教個體和特定新教個體的態度。然而,我們的態度測量未能找到區別效度,這與此相反。也就是說,不適當的刺激往往是相當一般或刻板印象的,而因為我們的工具具有建構效度,這類錯誤會增加而不是減少區別效度的程度。雖然在目前的情況下並未顯示出這類錯誤,但在我將討論的最後一組選擇中則顯示出來。
### Case C: Discriminant Validity in Both the Attitudinal and the Behavioral Measures
### 案例 C:態度和行為測量中的區別效度
Consider the case where both our behavioral and our pencil-and-paper measures have discrim
考慮這種情況:我們的行為測量和紙筆測量都有區分能力
---
inant validity but are not correlated with each other. Again, this lack of convergent validity does not imply that our measure of attitude is invalid. Here, however, the alternatives are two-fold: (1) as mentioned above, this may well be a situation in which we have mcasfired attitudes toward an inappropriate attitude object; (2) this is also the kind of situation where we may want to look for some other variable that is somewhat related both to our attitude measure and to our manipulations. That is, this is the type of situation in which variables such as beliefs or behavioral intentions are truly likely to increase our ability to predict behavior.
在這種情況下,雖然我們的測量工具具有顯著的效度,但彼此之間並無相關性。再者,這種缺乏收斂效度並不意味著我們的態度測量是無效的。然而,這裡的替代方案有兩種: (1) 如上所述,這可能是一種我們對不適當的態度對象有錯誤態度的情況; (2) 這也是一種我們可能想要尋找某些與我們的態度測量和操控之間有一定關聯的其他變數的情況。也就是說,這是變數如信念或行為意圖真正可能提高我們預測行為能力的情況。
Obviously, there are many other alternatives that may be differentiated when one uses a multiattitude object, multimethod approach (e.g., both attitude and behavioral measures may have low discriminant validity; convergent validity may be obtained with only one set of the attitude and behavioral measures). Although a complete discussion of these alternatives is beyond the scope of the present paper, it is worth noting that each alternative carries different implications for an understanding of the attitude-behavior relationship. In contrast to this, the usual procedure of obtaining a single-attitude measure and attempting to predict one or more behaviors provides us with little or no information in those cases where convergent validity is not obtained. That is, the failure to find convergent validity in such cases can be interpreted in many ways. As was pointed out above, the most common interpretation has been that there was something wrong with the measure of attitude (e.g., it was inappropriate; it was only a measure of affect and did not take the cognitive and the action component into account). However, as was also pointed out above, this lack of convergent validity can also be interpreted as indicating that the behavior under consideration is unrelated to the attitude that was measured. More specifically, the multiattitude object, multimethod approach clearly indicates that many of an individual's behaviors with respect to a given object are likely to be primarily under the control of variables other than the individual's attitude toward that object. In particular, it emphasizes the importance of the situation as a factor determining behavior.
顯然,當使用多態度物件和多方法的方式時,還有許多其他的替代方案可以區分(例如,態度和行為測量可能具有低的區分效度;收斂效度可能僅通過一組態度和行為測量來獲得)。雖然對這些替代方案的完整討論超出了本文的範疇,但值得注意的是,每個替代方案對於理解態度與行為之間的關係都有不同的含義。與此相對,獲得單一態度測量並試圖預測一個或多個行為的常規程序,在收斂效度未獲得的情況下,幾乎無法提供任何信息。也就是說,在這種情況下未能找到收斂效度可以有多種解釋。如上所述,最常見的解釋是態度測量存在問題(例如,它不合適;它僅僅是情感的測量,並未考慮到認知和行動成分)。 然而,如上所述,這種缺乏收斂效度也可以解釋為所考慮的行為與所測量的態度無關。更具體地說,多態度對象、多方法的方式清楚地表明,個體對於特定對象的許多行為可能主要受其他變數的控制,而非個體對該對象的態度。特別是,它強調情境作為決定行為的一個因素的重要性。
Although many investigators have often recognized the importance of situational and "other" variables as determinants of behavior, most treatments of attitude have not dealt with
儘管許多研究者經常認識到情境和「其他」變數作為行為決定因素的重要性,但大多數對態度的處理並未涉及
## Attitude and the prediction of behavior
## 態度與行為的預測
these factors explicitly. Rather, they have usually been viewed as sources of "crror" variance. The conception of attitude as a "predisposition to respond" has generally led to the assumption that an individual's behavior with respect to some object is a direct function of his attitude toward the object, and "other" factors are viewed as "noise" in the system. In the final section of this paper, a theory will be presented that attempts to deal with this larger constellation of behavioral determinants within a single formulation of the attitude problem. Rather than starting with the assumption that there is a strong underlying relationship between attitude and behavior, the theory to be presented emphasizes the importance of situational, motivational, and normative variables as factors influencing overt behavior.
這些因素是明確的。相反,它們通常被視為“誤差”變異的來源。將態度概念化為“反應的傾向”通常導致假設個體對某個對象的行為是其對該對象態度的直接函數,而“其他”因素則被視為系統中的“噪音”。在本文的最後一部分,將提出一個理論,試圖在單一的態度問題框架內處理這一更大範圍的行為決定因素。該理論不以態度和行為之間存在強烈的潛在關係為假設,而是強調情境、動機和規範變數作為影響顯性行為的因素的重要性。
## The attitude-behavior relationship: a reanalysis
## 態度與行為的關係:重新分析
The theory to be presented below can best be seen as an adaptation of Dulany's (1961, 1964). theory of propositional control Although Dulany's theory has largely been developed within the context of studies of verbal conditioning and concept attainment, it is essentially a theory that leads to the prediction of overt behavior. It is hoped that the present approach will provide a more complete understanding of the relationships between attitudes and behavior, by identifying some of the ways in which other variables interact with attitudes as determinants of overt behavior. More specifically, the present approach may be viewed as an attempt to bridge the gap between traditional measures of attitude toward a stimulus object and behavior with respect to that object in a given situation.
以下將要呈現的理論可以最佳地視為對杜蘭尼(Dulany,1961,1964)命題控制理論的適應。儘管杜蘭尼的理論主要是在語言條件反射和概念獲得的研究背景下發展起來的,但它本質上是一種預測明顯行為的理論。希望目前的方法能夠通過識別其他變數如何與態度互動,作為明顯行為的決定因素,提供對態度與行為之間關係的更全面理解。更具體地說,目前的方法可以被視為試圖彌合對刺激對象的傳統態度測量與在特定情境下對該對象的行為之間的差距。
As mentioned above, Dulany's theory has largely been developed within the context of studies of verbal conditioning and concept attainment. More specifically, Dulany has been concerned with predicting the probability with which an individual will make a particular verbal response or class of verbal responses. The central equation of the theory can be expressed as follows:
如上所述,杜蘭尼的理論主要是在語言條件反射和概念獲得的研究背景下發展起來的。更具體地說,杜蘭尼關注的是預測個體做出特定語言反應或語言反應類別的概率。該理論的核心方程可以表達如下:
\[BI = [(RHd)(A)]w_{0} + [(BH)(Mc)]w_{1}\]
where
哪裡
\[BI = \text{the subject's intention to make a particular response or class of responses}\] \[RHd = \text{a ``hypothesis of the distribution of reinforcement,'' that is, the subject's
\[BI = \text{受試者意圖做出特定反應或反應類別}\] \[RHd = \text{“增強分佈的假設”,即受試者的}
---
Atttitude theory
態度理論
hypothesis that the occurrence of the particular response will lead to a certain event or class of events \(A = \frac{}{}\) the affective value of the reinforcement, that is, the subject's evaluations of (or attitudes toward) those events \(BH = \frac{}{}\) the subject's "behavioral hypothesis," that is, his belief as to what he is _expected_ to do, or what he _should_ do in the situation \(Mc = \frac{}{}\) the subject's "motivation to comply," that is, how much the subject _wants_ to do what he believes is expected of him \(w_{0}\) and \(w_{1} = \frac{}{}\)beta weights, which may take any value
假設特定反應的發生將導致某個事件或事件類別 \(A = \frac{}{}\) 強化的情感價值,即主體對這些事件的評價(或態度) \(BH = \frac{}{}\) 主體的「行為假設」,即他對在該情境中應該做什麼或預期會做什麼的信念 \(Mc = \frac{}{}\) 主體的「遵從動機」,即主體有多想做他認為被期望的事情 \(w_{0}\) 和 \(w_{1} = \frac{}{}\) beta 權重,可以取任何值
Two additional points about the theory should be mentioned:
關於該理論還應提及兩個額外的要點:
1. According to the theory, behavioral intention is the immediate antecedent of overt behavior. Unlike the general types of behavioral intentions that most attitude researchers (e.g., Triandis, 1964; Triandis, Fishbein, and Hall, 1964) have been concerned with (e.g., intentions to marry, to show social distance, to subordinate, to accept as a friend), Dulany has been concerned with a more precise and specific type of behavioral intention--namely, an individual's intention to perform a given action in a given situation.
根據理論,行為意圖是顯性行為的直接前因。與大多數態度研究者(例如,Triandis,1964;Triandis、Fishbein 和 Hall,1964)所關注的一般類型的行為意圖(例如,結婚意圖、表現社交距離的意圖、屈從的意圖、接受為朋友的意圖)不同,Dulany 關注的是一種更精確和具體的行為意圖——即個體在特定情境下執行特定行動的意圖。
That is, it will be recalled that in order to obtain a measure of an individual's behavioral intention (to show social distance, for example), Triandis would ask the subject to indicate whether he (i.e., the subject) "would" or "would not" engage in a specific set of behaviors (e.g., "invite to my club," "exclude from my neighborhood," "accept as a close kin by marriage," etc.) with a given stimulus person. Triandis would then sum the responses to these items, and this sum would be taken as the measure of the respondent's behavioral intention (to show social distance). In contrast to this, Dulany has been concerned with measuring the subject's intention to perform the specific behavior he is interested in predicting. Furthermore, unlike Triandis, Dulany would not ask the subject to indicate his intention to engage in the act "in general," (e.g., measure his intention to "obey" the stimulus person), but rather he would measure the subject's intentions to engage in the act in a particular situation. Because of this close correspondence between the measure of the behavioral intention and the actual behavior that Dulany wants to predict, the correlation between the measure of behavioral intention and the actual overt behavior is almost perfect (i.e., the correlations are always in the mid .90's). Thus if one can predict the specific behavioral intention, one may, with only slightly attenuated accuracy, predict the overt behavior. It must be emphasized, however, that these near-perfect correlations between behavioral intentions and behavior are only obtained, and are only expected, when one considers an individual's intention to perform a specific act in a specific situation. The more abstract or generalized the intention becomes, the lower will be its correlation with a specific behavior.
也就是說,值得回顧的是,為了獲得個體行為意圖的衡量(例如,表現社交距離),Triandis 會要求受試者指出他(即受試者)「會」或「不會」在特定情境下進行一系列行為(例如,「邀請到我的俱樂部」、「排除在我的社區之外」、「接受為婚姻中的親密親屬」等)與特定的刺激對象。然後,Triandis 會將這些項目的回應進行總和,這個總和將被視為受訪者的行為意圖(表現社交距離)的衡量。與此相比,Dulany 更關注於測量受試者執行他感興趣的特定行為的意圖。此外,與 Triandis 不同,Dulany 不會要求受試者指出他「一般」的行為意圖(例如,測量他對刺激對象的「服從」意圖),而是會測量受試者在特定情境下進行該行為的意圖。 由於杜蘭尼希望預測的行為意圖的衡量與實際行為之間存在密切的對應關係,因此行為意圖的衡量與實際明顯行為之間的相關性幾乎是完美的(即,相關性總是在中間的 90 幾)。因此,如果能夠預測特定的行為意圖,則可能以僅稍微降低的準確性預測明顯行為。然而,必須強調的是,這些行為意圖與行為之間的近乎完美的相關性僅在考慮個體在特定情境中執行特定行為的意圖時才能獲得,並且僅在這種情況下才會被預期。意圖越抽象或一般化,與特定行為的相關性就越低。
2. As can be seen above, the algebraic expression of the theory takes the form of a linear multiple regression equation. That is, [(\(RHd\))(\(A\))] is viewed as one component influencing behavioral intentions and [(\(BH\))(\(Mc\))] is seen as a second component. The precise weights to be given these two components as determinants of behavioral intentions within a given situation may be determined by standard multiple-regression procedures.
如上所示,該理論的代數表達式呈現為線性多重回歸方程式。也就是說,[(\(RHd\))(\(A\))] 被視為影響行為意圖的一個組成部分,而[(\(BH\))(\(Mc\))] 則被視為第二個組成部分。在特定情境中,這兩個組成部分作為行為意圖的決定因素所需的精確權重可以通過標準的多重回歸程序來確定。
A liberal interpretation of this theory as applied to social behavior may be given as follows. An individual's intention to perform a specific act, with respect to a given stimulus object, in a given situation, is a function of the following:
對於社會行為應用此理論的自由解釋可以如下給出:個體在特定情境下對特定刺激物執行特定行為的意圖,是以下因素的函數:
1a. His beliefs about the consequences of performing a particular behavior (in a given situation), that is, the probability or improbability that the performance of behavior \(X\) will lead to some consequence \(y_{x}\) (\(B_{U}\)).
1b. The evaluative aspect of \(B_{U}\) that is, the _S_'s evaluation of \(y_{x}\) (\(a_{U}\)).
2a. A normative belief, that is, the _S_'s belief about what he _should_ do in this situation (_NB_).
2b. His motivation to comply with the norm, that is, his desire, or lack of desire, to do what he thinks he should do (\(Mc\)).
1a. 他對於在特定情境下執行某種行為的後果的信念,即行為 \(X\) 導致某種後果 \(y_{x}\) 的可能性或不可能性 (\(B_{U}\))。1b. \(B_{U}\) 的評價方面,即 _S_ 對 \(y_{x}\) 的評價 (\(a_{U}\))。2a. 一種規範信念,即 _S_ 對於在這種情況下他 _應該_ 做什麼的信念 (_NB_)。2b. 他遵循該規範的動機,即他想要或不想要做他認為應該做的事情 (\(Mc\))。
Thus in the present adaptation of Dulany's theory, \(RHd\) (the hypothesis of the distribution of a reinforcer) is conceptualized as being analogous to an individual's beliefs about the consequences of performing a specific behavior, and \(A\) (the affective value of the reinforcer)
因此,在當前對杜蘭尼理論的改編中,\(RHd\)(增強物分佈的假設)被概念化為類似於個體對執行特定行為後果的信念,而\(A\)(增強物的情感價值)
---
is conceptualized as being equivalent to the evaluative aspects of those beliefs. It is interesting to note that this conceptualization ref defines the first component of Dulany's theory [(_RHd_)(_A_)] as a measure of attitude. That is, as was mentioned above, the work of Rosenberg (1956, 1965), Zajonc (1954), Fishbein (1963, 19655\(b\), 1967a, 1967b), and others has provided strong evidence that an individual's attitude toward any object can be predicted from a knowledge of the individual's beliefs about the object and the evaluative aspects of those beliefs. Algebraically, this may be expressed as follows:
被概念化為與這些信念的評價方面等同。值得注意的是,這一概念化重新定義了杜蘭尼理論的第一個組成部分 [(_RHd_)(_A_)] 為態度的衡量。也就是說,如上所述,羅森伯格(1956, 1965)、扎永茨(1954)、費什賓(1963, 1965\(b\), 1967a, 1967b)等人的研究提供了強有力的證據,表明個體對任何對象的態度可以從對該對象的信念及其評價方面的知識中預測出來。在代數上,可以表達如下:
$A_{o}=\sum_{i=1}^{n} B_{i} a_{i}$
where
哪裡
\[A_{\theta} = \text{the attitude toward some object ``o''}\]
\[A_{\theta} = \text{對某個物體 ``o'' 的態度}\]
\[B_{i} = \text{belief }i\text{ about }o,\text{ that is, the probability that }o\text{ is related to some other object }x_{i}\]
\[B_{i} = \text{信念 }i\text{ 關於 }o,\text{ 即 }o\text{ 與某個其他物件 }x_{i}\text{ 相關的概率} \]
\[a_{i} = \text{the evaluative aspect of }B_{i}\text{ that is, the respondent's attitude toward }x_{i}\]
\[a_{i} = \text{評估方面的 }B_{i}\text{,即受訪者對 }x_{i}\text{ 的態度} \]
\[n = \text{the number of beliefs}\]
\[n = \text{信念的數量}\]
It must be emphasized, however, that the attitude under consideration is an attitude toward performing a given behavioral act, and is _not_ an attitude toward a given object, person, or situation. That is, the algebraic formula presented above leads to the hypothesis that an individual's attitude toward any object is a function of the individual's beliefs about _that_ object. In the present analysis, we are concerned with an individual's beliefs about the performance of a given behavioral act, and thus the attitude being assessed is the individual's attitude toward the performance of that act. More specifically, from the point of view of Dulany's theory, we should be assessing the individual's beliefs about what will happen if he performs behavior \(X\) with respect to stimulus \(Y\) in situation \(Z\), and the evaluative aspects of those beliefs, that is, we should be assessing the individual's attitude toward the performance of a given act, with respect to a given stimulus object, in a given situation. Although these beliefs _may_ vary considerably as a function of the stimulus object (e.g., the person) toward which the act is directed and the situation in which the act is to occur (e.g., in public or in private), these beliefs are still beliefs about the performance of the act, and not beliefs about the stimulus object or the situation. We shall return to this point below.
然而,必須強調的是,所考慮的態度是對執行特定行為的態度,而不是對特定物體、個人或情境的態度。也就是說,上述的代數公式導致了這樣的假設:個體對任何物體的態度是個體對該物體的信念的函數。在本分析中,我們關心的是個體對執行特定行為的信念,因此所評估的態度是個體對執行該行為的態度。更具體地說,從杜蘭尼理論的角度來看,我們應該評估個體對於在情境 \(Z\) 中,針對刺激 \(Y\) 執行行為 \(X\) 時會發生什麼的信念,以及這些信念的評價方面,也就是說,我們應該評估個體對於在特定情境中,針對特定刺激物體執行特定行為的態度。儘管這些信念可能會根據刺激物體而有相當大的變化(例如。該行為所指向的對象(即人)以及該行為將發生的情境(例如,在公共場合或私下),這些信念仍然是關於行為執行的信念,而不是關於刺激對象或情境的信念。我們將在下面回到這一點。
Turning to the second component of Dulany's theory [(_BH_)(_Mc_)], it can be seen that _BH_ (the behavioral hypothesis) has been conceptualized as a normative belief, that is, a belief about what _should_ be done in the situation. More specifically, this is a belief about whether the particular act _should_ or _should not_ be performed. Here, however, it may be necessary to distinguish between two types of normative beliefs: (1) the individual's belief about what he _personally_ feels he should do (i.e., a personal norm or rule of behavior); and (2) the individual's belief about what "society" (i.e., most other people, his "significant others," etc.) "says" he should do (i.e., a social or group norm). Although an individual's personal norms will often coincide with,'or reflect, the social norms, this is not necessarily the case. For example, while social norms concerning interracial relations are markedly different in some Northern and Southern communities, a given individual in a Southern community may hold the same personal normative belief as an individual in a Northern community. Thus it may be necessary to take both of those types of normative beliefs into account. This can easily be done, for as was mentioned above, the central equation of Dulany's theory takes the form of a multiple regression equation. Thus, as Dulany has pointed out, this theory is an open theory, and additional components can be taken into account. We shall discuss this more fully below.
轉向杜蘭尼理論的第二個組成部分[_BH_][_Mc_],可以看出_BH_(行為假設)被概念化為一種規範信念,即對於在特定情況下應該做什麼的信念。更具體地說,這是一種關於特定行為是否應該或不應該被執行的信念。然而,在這裡,可能有必要區分兩種類型的規範信念:(1)個體對於他自己認為應該做什麼的信念(即個人規範或行為準則);以及(2)個體對於“社會”(即大多數其他人、他的“重要他人”等)“說”他應該做什麼的信念(即社會或群體規範)。儘管個體的個人規範通常會與社會規範相符或反映社會規範,但這並不一定是情況。例如,儘管某些北方和南方社區對於種族關係的社會規範有顯著不同,但在南方社區中的某個個體可能會持有與北方社區中的個體相同的個人規範信念。因此,可能有必要考慮這兩種類型的規範信念。 這可以很容易地做到,正如上面提到的,杜蘭尼理論的中心方程式呈現為多重回歸方程式。因此,正如杜蘭尼所指出的,這個理論是一個開放的理論,並且可以考慮額外的組件。我們將在下面更詳細地討論這一點。
The final element in the equation (_Mc_--an individual's motivation, to comply) is self-explanatory. Here we are concerned with the degree to which the individual "wants" to comply with the norm. Clearly, if two types of norms are considered, it will be necessary to measure the individual's motivation to comply with each of them.
方程式中的最後一個元素(_Mc_--個體的遵從動機)是不言自明的。在這裡,我們關心的是個體「想要」遵從規範的程度。顯然,如果考慮到兩種類型的規範,就有必要測量個體對每一種規範的遵從動機。
Returning to the central hypothesis, it can be seen that in its adapted form Dulany's theory essentially leads to the prediction that an individual's intention to perform any behavior (and his actual performance of the behavior) is a function of (1) his attitude toward performing the behavior in a given situation, and (2) the norms governing that behavior in that situation and his motivation to comply with those norms. Algebraically, this can be expressed as follows:
回到中心假設,可以看出在其適應形式中,杜蘭尼的理論本質上預測個體執行任何行為的意圖(以及他實際執行該行為)是以下因素的函數:(1)他對在特定情境中執行該行為的態度,以及(2)該情境中規範該行為的規範及他遵循這些規範的動機。代數上,可以表達如下:
\[B \approx \mathit{BI} = [A_{\mathit{act}}]w_{0} + [(\mathit{NB})(\mathit{Mc})]w_{1}\]
However, as was discussed above, it is not clear
然而,如上所述,這並不清楚
---
### Attitude Theory
### 態度理論
whether the second component [(_NB_)(_Mc_)] can best be viewed as referring to a personal norm, a social norm, or both. The simplest way to handle this problem within the framework of the theory is to expand the algebraic formulation as follows:
第二個組件 [(_NB_)(_Mc_)] 是否可以最好地視為指涉個人規範、社會規範,或兩者皆是。在理論框架內處理這個問題的最簡單方法是擴展代數公式如下:
\[\begin{array}{l}B\simeq BI=[A_{act}]w_{0}+[(NB_{p})(Mc_{p})]w_{1}\\ +[(NB_{p})(Mc_{p})]w_{2}\end{array}\]
where the subscripts \(p\) and \(s\) refer to personal and social norms respectively.3
其中下標 \(p\) 和 \(s\) 分別指個人規範和社會規範。3
Thus, as mentioned previously, the present approach suggests a radical change in investigating and understanding the attitude-behavior relationship. Instead of assuming some underlying relationship between an individual's attitude toward a given object and his behavior with respect to that object, the proposed theory recognizes the importance of situational variables, norms, and motivation as factors influencing behavior. Rather than viewing attitude toward a stimulus object as a major determinant of behavior with respect to that object, the theory identifies three kinds of variables that function as the basic determinants of behavior: (1) attitudes toward the behavior; (2) normative beliefs (both personal and social); and (3) motivation to comply with the norms. Furthermore, although the theory suggests that other variables can also influence behavior, it indicates that these other variables operate _indirectly,_ that is, by influencing one or more of the three basic determinants.
因此,如前所述,目前的方法建議在調查和理解態度與行為之間的關係時進行根本性的變革。該理論並不假設個體對特定對象的態度與其對該對象的行為之間存在某種潛在關係,而是認識到情境變數、規範和動機作為影響行為的因素的重要性。該理論並不將對刺激對象的態度視為影響與該對象行為的主要決定因素,而是確定了三種作為行為基本決定因素的變數:(1)對行為的態度;(2)規範信念(包括個人和社會);以及(3)遵循規範的動機。此外,儘管該理論表明其他變數也可以影響行為,但它指出這些其他變數是_間接_運作的,即通過影響一個或多個基本決定因素。
Thus under certain conditions an individual's attitude toward a stimulus object may be related to his behavior with respect to that object. For example, it is quite possible that an individual's beliefs about the consequences of performing some behavior (X) will be quite different if the behavior is to be performed with respect to someone he likes (p1) than with respect to someone he dislikes (p2). If this were the case, we would expect the individual to have a different attitude toward performing X with respect to p1, than toward performing X with
因此,在某些條件下,個體對刺激物的態度可能與他對該物體的行為有關。例如,個體對於執行某種行為(X)後果的信念,若該行為是針對他喜歡的人(p1)而執行,與針對他不喜歡的人(p2)而執行時,可能會有很大的不同。如果是這樣,我們會預期個體對於針對 p1 執行 X 的態度,會與針對 p2 執行 X 的態度有所不同。
Footnote 3: An alternative formulation can be seen as follows:
腳註 3:另一種表述可以如下所示:
\[BI=[A_{act}]w_{0}+\left[\sum_{i=1}^{n}NB_{p}Mc_{i}\right]w_{1}\]
This formulation suggests that it may be necessary to consider many different types of normative beliefs; for example, beliefs about what one's (_a_) parents, (_b_) friends, (_c_) co-workers, (_d_) religious group, etc., "says" the individual _should_ do.
這種表述暗示可能需要考慮許多不同類型的規範信念;例如,關於一個人的 (_a_) 父母、(_b_) 朋友、(_c_) 同事、(_d_) 宗教團體等,所“說”的個體 _應該_ 做什麼。
respect to \(\hat{p}_{2}\). In other cases, however, the consequences of performing a given act may not vary with variation in \(\hat{p}\), and here attitude toward performing \(X\) may not change. In the former case, it is possible that we would obtain a correlation beween attitude toward \(\hat{p}\) and behavior with respect to \(\hat{p}\), while in the latter case this is unlikely.
對於 \(\hat{p}_{2}\) 的尊重。然而,在其他情況下,執行某個行為的後果可能不會隨著 \(\hat{p}\) 的變化而變化,在這裡對執行 \(X\) 的態度可能不會改變。在前者的情況下,我們可能會獲得對 \(\hat{p}\) 的態度與對 \(\hat{p}\) 的行為之間的相關性,而在後者的情況下,這是不太可能的。
Similarly, we might also expect an individual's attitude toward a given stimulus person to influence his motivation to comply with a norm. That is, if we assume that an individual believes that he _should_ perform behavior \(X\) with respect to individual \(\hat{p}\) in situation \(X\) and, further, if we assume that the individual views behavior \(X\) as a behavior that indicates friendship (e.g., "invite to my club"), then it seems likely that the more positive the individual's attitude is to \(\hat{p}\), the more he will want to comply with the norm. Similarly, the more negative his attitude is to \(\hat{p}\), the less he may want to comply with the norm, that is, the less he may want to behave in "a friendly manner" even though he believes he should. Here again, however, it must be recognized that there will be cases where the individual's motivation to comply may be completely unrelated to his attitude toward \(\hat{p}\).
類似地,我們也可以預期個體對特定刺激對象的態度會影響他遵循規範的動機。也就是說,如果我們假設一個個體認為他在情境 \(X\) 中應該對個體 \(\hat{p}\) 執行行為 \(X\),而且進一步假設該個體將行為 \(X\) 視為表示友誼的行為(例如,「邀請加入我的俱樂部」),那麼似乎越是對 \(\hat{p}\) 持正面態度,越會想要遵循這一規範。相似地,對 \(\hat{p}\) 持負面態度的程度越高,他可能越不想遵循這一規範,也就是說,即使他相信自己應該這樣做,他也可能不想以「友好的方式」行事。然而,在這裡必須再次認識到,會有一些情況下,個體遵循規範的動機可能與他對 \(\hat{p}\) 的態度完全無關。
Along these same lines, it can be seen that variations in the situation may also influence one or more of the primary determinants of behavior. Clearly, an individual is likely to have quite different beliefs about the consequences of performing a given behavior in a public situation than in a private situation. Similarly, we would also expect that the individual has different normative beliefs (both personal and social) for different situations. Here too, however, it must be noted that if the situational variables that are being considered do _not_ influence the individual's attitude toward the act, or his personal or social normative beliefs, or his motivation to comply with these beliefs, then, according to the theory, they will not influence his behavior. Indeed, one of the advantages of the theory is that it provides an explanation for a large number of results that may initially appear inconsistent. That is, from the point of view presented here, one would expect a considerable amount of variation in the relationship between any given variable and behavior.
沿著這些相同的思路,可以看出情境的變化也可能影響一個或多個行為的主要決定因素。顯然,個體在公共場合執行某一行為的後果信念,與在私人場合的信念會有很大不同。同樣,我們也會預期個體在不同情境下擁有不同的規範信念(包括個人和社會的)。然而,在這裡也必須指出,如果所考慮的情境變數並不影響個體對該行為的態度,或其個人或社會的規範信念,或其遵循這些信念的動機,那麼根據理論,它們將不會影響其行為。事實上,這一理論的一個優勢在於,它為大量初看似不一致的結果提供了解釋。也就是說,從這裡提出的觀點來看,預期任何給定變數與行為之間的關係會有相當大的變化。
In addition, it should be recalled that the specific weights given to the three basic components [i.e., \((A_{\text{act}})\), \((NB_{p}\times M\epsilon_{p})\), and \((NB_{q}\times M\epsilon_{q})\)
此外,應該回想起三個基本組件所賦予的具體權重 [即 \((A_{\text{act}})\)、\((NB_{p}\times M\epsilon_{p})\) 和 \((NB_{q}\times M\epsilon_{q})\)
---
## Attitude and the prediction of behavior
## 態度與行為的預測
Mc)] of the theory must be determined; indeed, these weights are expected to vary from. behavior to behavior. That is, with respect to some behaviors, the attitude toward the behavior may be a more important determinant of behavior than either type of normative belief and the motivation to comply with these norms, while for other types of behaviors the weight- ings may be reversed. Similarly, while some behaviors may be primarily determined by per- sonal normative beliefs, other behaviors may be more under the control of social normative beliefs.
理論中的權重必須被確定;事實上,這些權重預期會因行為而異。也就是說,對於某些行為,對該行為的態度可能比任何類型的規範信念和遵循這些規範的動機更重要,而對於其他類型的行為,這些權重可能會顛倒。同樣,雖然某些行為可能主要由個人規範信念決定,但其他行為可能更多地受到社會規範信念的控制。
Furthermore, just as the weightings of the three components may vary with the type of behavior that is being considered, they may also vary for different individuals. That is, for some individuals, behavior may be primarily determined by norms and motivation to comply with the norms, while for other individuals behavior may be primarily determined by their attitudes toward the behavior. For example, it is possible that the behavior of high authoritarian individuals may be largely under the control of normative beliefs and motivation to comply with the norms.
此外,正如三個組成部分的權重可能因所考慮的行為類型而異,它們對不同個體也可能有所不同。也就是說,對於某些個體,行為可能主要受到規範和遵循規範的動機的影響,而對於其他個體,行為可能主要受到他們對該行為的態度的影響。例如,高權威主義個體的行為可能在很大程度上受到規範信念和遵循規範的動機的控制。
Thus, even though an "outside" variable (e.g., attitude toward a stimulus object) may be related to one of the basic determinants of behavior (e.g., attitude toward the behavior), the outside variable may still be unrelated to the actual performance of that behavior.
因此,即使一個「外部」變數(例如,對刺激物的態度)可能與行為的基本決定因素之一(例如,對該行為的態度)有關,這個外部變數仍然可能與該行為的實際表現無關。
SUMMARY
摘要
The present approach clearly indicates that traditional measures of attitude (i.e., toward a given object, person, or class of people) are _not_ likely to be related to behavior in any consistent fashion. Indeed, the main purpose of the present paper has been to reconsider, and seriously question, the basic assumption that an individual's attitude toward an object is a major determinant of his behavior with respect to the object. While other investigators have also argued that traditional attitude measures will not predict behavior, they have usually questioned the measure of attitude rather than the assumption of an attitude-behavior relationship. More specifically, they have argued that most traditional measures of attitude are oversimplified; that most measures only consider an individual's "affective feelings," and fail to take his cognitions and conclusions into account. Thus most investigators have attempted to resolve the attitude-behavior problem by expanding the definition of attitude to include affective, cognitive, and conative components.
目前的方法清楚地表明,傳統的態度測量(即對特定對象、個人或人群的態度)不太可能以任何一致的方式與行為相關。事實上,本論文的主要目的是重新考慮並認真質疑一個基本假設,即個體對某一對象的態度是其在該對象上行為的主要決定因素。雖然其他研究者也曾主張傳統的態度測量無法預測行為,但他們通常質疑的是態度的測量,而不是態度與行為之間關係的假設。更具體地說,他們認為大多數傳統的態度測量過於簡化;大多數測量僅考慮個體的「情感感受」,而未能考慮其認知和結論。因此,大多數研究者試圖通過擴展態度的定義來解決態度-行為問題,以包括情感、認知和意向成分。
In contrast to this, the present paper has attempted to show that beliefs (cognitions) and behavioral intentions (conations) can best be viewed as determinants or consequents of attitude; that rather than being viewed as parts of attitude, these variables should be viewed as independent phenomena that are related to, and serve as indicants of, an individual's attitude. Furthermore, it was argued that even if an individual's beliefs about an object and/or his behavioral intentions toward the object are considered, it is not likely that behavioral prediction will be improved. First, these variables may be highly correlated with traditional measures of attitude, and thus they will not explain additional variance in behavior. Second, even if this is not the case (i.e., even if beliefs and behavioral intentions that are unrelated to attitude are selected for consideration), they may still be unrelated to the behavior.4 That is, viewing the attitude-behavior relationship within the framework of a multiattitude object-multimethod approach, it becomes clear that the most important determinants of behavior may be other variables than an individual's beliefs about, attitude toward, or general behavioral intentions toward, a given object. Indeed, this approach clearly indicates that behavior toward an object may be completely determined by situational or individual difference variables, rather than any variable associated with the stimulus object per se. In other words, this approach points out that behavior toward a given object is a function of many variables, of which attitude toward the object is only one.
與此相反,本論文試圖表明,信念(認知)和行為意圖(意向)最好被視為態度的決定因素或結果;這些變數應被視為獨立的現象,與個體的態度相關,並作為個體態度的指標,而不是態度的一部分。此外,論文還指出,即使考慮到個體對某一對象的信念和/或他對該對象的行為意圖,行為預測的改善也不太可能。首先,這些變數可能與傳統的態度測量高度相關,因此不會解釋行為中的額外變異。其次,即使情況並非如此(即,即使選擇與態度無關的信念和行為意圖進行考慮),它們仍可能與行為無關。4 也就是說,從多態度對象-多方法的框架來看態度與行為的關係,最重要的行為決定因素可能是其他變數,而不是個體對某一對象的信念、態度或一般行為意圖。事實上,這種方法清楚地表明,對一個對象的行為可能完全由情境或個體差異變數決定,而不是與刺激對象本身相關的任何變數。換句話說,這種方法指出,對某一對象的行為是多種變數的函數,其中對該對象的態度僅僅是其中之一。
Footnote 4: It should be recalled that the size of the relationship between a behavioral intention and a behavior is primarily a function of the specificity of the behavioral intention.
腳註 4:應該記住,行為意圖與行為之間關係的大小主要取決於行為意圖的具體性。
Consistent with this position, a theory of behavioral prediction was suggested. In contrast to previous attempts to resolve the attitude-behavior problem by expanding the definition of attitude, the proposed theory attempts to understand the attitude-behavior relationship by (1) considering a limited set of variables that function as primary determinants of behavior, and (2) analyzing the relations between these variables and traditional methods of attitude.
與此立場一致,提出了一種行為預測理論。與以往通過擴展態度定義來解決態度-行為問題的嘗試不同,所提出的理論試圖通過(1)考慮一組有限的變數,這些變數作為行為的主要決定因素,以及(2)分析這些變數與傳統態度方法之間的關係,來理解態度-行為關係。