A Psychological Study of the Inverse Relationship Between Perceived Risk and Perceived Benefit

Please download to get full document.

View again

of 12
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Document Description
Judgments of risk and judgments of benefit have been found to be inversely related. Activities or technologies that are judged high in risk tend to be judged low in benefit, and vice versa. In the present study, we examine this inverse relationship
Document Share
Document Tags
Document Transcript
  Risk Analysis, Vol. 14, No. 6 1994 A Psychological Study of the Inverse Relationship Between Perceived Risk and Perceived Benefit Ali Siddiq Alhakamil and Paul Slovic2 Received March 8, 1994; revised June 28 1994 Judgments of risk and judgments of benefit have been found to be inversely related. Activities or technologies that are judged high in risk tend to be judged low in benefit, and vice versa. In the present study, we examine this inverse relationship in detail, using two measures of relationship between risk and benefit. We find that the inverse relationship is robust and indicative of a con- founding of risk and benefit in people’s minds. This confounding is linked to a person’s overall evaluation of an activity or technology. Theoretical and practical implications of this risk-benefit confounding are discussed. 1. INTRODUCTION 1.1. Background KEY WORDS: erceived risk; perceived benefit. Analytic approaches to c-cision making treat risk and benefit as distinct concepts. The benefit one gains from driving to work is, presumably, qualitatively dif- ferent from the risk. In some cases, such as with medical technologies, risks and benefits are not conceptually dis- tinct or independent-the benefits are reductions in phys- ical ailments and death. However, such cases of nonin- dependence are relatively infrequent. Studies in which people have been asked to judge risks and benefits have consistently observed an inverse (negative) correlation between perceived risk and per- ceived benefit across diverse hazards. The first such re- port, from the psychometric study by Fischhoff ef al.>’) found that, for 30 items examined, mean perceived risk and mean perceived benefit were inversely related; that is, the greater the perceived benefit, the lower the per- 1 Imam Muhammad Ibn Saud Islamic University Psychology Depart- To whom correspondence should be addressed at Decision Research, ment, P.O. Box 15593, Riyadh, 11454 Saudi Arabia. 1201 Oak Street, Eugene, Oregon 97401. ceived risk, and vice versa. For example, people tended to judge alcoholic beverages, handguns, and smoking as very low in benefit and very high in risk; on the other hand, they perceived prescription antibiotics, railroads, and vaccinations as having high benefit and relatively low risk. Simiiar results were found by Slovic et aZ.,@ who surveyed a representative sample of the adult population in Canada. Survey respondents were asked first to rate the riskiness of each of 33 items; they then rated the benefit of each item. Both risk and benefit were rated on 7-point scales. A chart of the means of the perceived risk and benefit of the 33 hazards (Fig. 1 clearly shows that per- ception of risk was inversely related to perception of ben- efit. The correlation between risk and benefit means was -.23. Although there is variability in the relationship, in general one sees that the higher the perceived risk, the lower the perceived benefit, and vice versa. Is the inverse relationship depicted in Fig. 1 merely a correct reflection of a fact about the world? Or might it reflect, at least in part, a confounding of risk and ben- efit in people’s minds? The present study attempts to answer this second question. Evidence that risks and benefits are indeed con- founded in the mind comes from new analyses that we 1085 0272-4332/94/1U)(1-1085 07.00/1 1994 Society for Risk Analysis  1086 Cigarette Smoking Pesticides Nuclear Power Sleeping Pills Tranquilizers Antidepressants Heartsurgery CancerDNgs Birth Control Pills Food Additives Automobiles Cleansers Menopause Drugs Nonprescnption Drugs X-Rays Antihypertensives Airplane Travel Art Sweeteners Biotech DNgS Prescnption Drugs Antibiotics Antiarthritics Laxatives Insulin Aspirin Appendectomy Acupunture Vaccines Herbal Medicines Vitamin Pills Alcohol IUDs AlDSDNgS Alhakami and Slovic 1 4 7 Degree of RisWBenefit Fig. 1. Means of the perceived risk and perceived benefit ratings by Slovic er aL (1991). performed on the data from Canada. Table I presents the correlations across 1261 survey respondents between perceived risk and perceived benefit for each of the 33 hazards studied in that survey. Negative correlations were present for 32 of the 33 hazard items, and 30 of these 32 correlations were statistically significant. This means, for example, that persons who perceived nuclear power as high in risk tended to see it as low in benefit, and vice versa. Why this inverse relationship is stronger for some hazards than for others is another question that the present study seeks to answer. Correlation has a limitation as a measure of the in- verse relationship, however. For example, if everyone perceived an activity to be high in risk and low in benefit the correlation might well be zero, despite the consistent discrepancy between the two kinds of perceptions. Therefore it seems necessary to us to calculate a second measure of the inverse relationship. This measure is the “distance” between perceived risk and perceived ben- efit, defined as the absolute difference between the two judgments for a particular item. These absolute differ- ences were calculated and then averaged across respon- dents for each item in the Canadian survey.(2) he high- est mean distance was obtained with cigarette smoking (D = 4.41 , which was judged to be of low benefit (mean = 1.83 and high risk (mean = 6.00 . High dis- tance was also characteristic of alcohol (D = 3.05 , with a mean benefit of 2.56 and a mean risk of 5.02. Other items exhibiting high distance scores were those judged as high in benefit and low in risk, such as vaccines (D = 3.21, mean benefit = 5.92, and mean risk = 2.93 , in- sulin D = 3.13, mean benefit = 6.07, and mean risk = 3.21 , and antibiotics (D = 2.88, mean benefit = 5.98, and mean risk = 3.46 . A relatively small degree of distance between risk and benefit judgments was characteristic of menopause drugs, biotechnology drugs, heart surgery, antidepres- sants, laxatives, and cleansers. The small distance be- tween perceived risk and perceived benefit suggests that individual respondents judged the risk and benefit at  Inverse Relationship 1087 Table I. Correlations Between Perceived Risk and Perceived Benefit for 33 Items [N = 1261; Canadian study@ ] Item Correlation Nuclear power Alcohol IUDs Nonprescription drugs Herbal medicines Aspirin Acupuncture Food additives Pesticides Birth control pills Artificial sweeteners Sleeping pills Tranquilizers Cigarette smoking Biotechnology drugs Cleansers Antidepressants Insulin Laxatives Antihypertensives Vaccines Menopause drugs X-rays Vitamin pills Antiarthritics Cancer drugs Antibiotics Appendectomy Airplane travel Prescription drugs AIDS drugs Automobiles Heart surgery -.33* -.30* -.27* -.24* -.24* 23’ -.23 -.23 -.23 -.22* -.22* -.20* -.20* -.20* -.20* -.18* -.17* -.15* -.14* -.14* -.13* -.12* -.12* -.12* -.12* -.11* -.11* -.lo* -.lo* -.08* - 05 -.03 .02 ‘Significant at .01 evel. about the same level on the 7-point scale. The ordering of items based on this distance measure is not signifi- cantly correlated with the ordering based on the risk- benefit correlations shown in Table I. The rank-order correlation between the two measures of risk-benefit dif- ferences is -.04. If risks and benefits are confounded within people’s minds, are there differences between persons in this re- gard? Again the answer is yes, based on additional anal- yses of the Canadian survey data. The correlation be- tween risk and benefit judgments was computed across the 33 hazard items for each of the 1261 respondents. Figure 2 presents the distribution of these correlations. Negative correlations were characteristic of 83 of the respondents, demonstrating the pervasiveness of this pat- tern of perception. Although the size of these correla- tions varied considerably across respondents, more than .73 o .aa 57 o .72 .41 to .56 iii .25 o .40 4 -09t0.24 x z -.25 o -.41 to -.40 .56 -.57 o .72 -.73 o -.a8 0 100 200 300 Frequency Fig. 2. Frequency distribution of the risk-benefit correlation within each respondent and across hazards by Slovic et al. (1991). one-quarter of the sample exhibited correlations more negative than -.41. Tiemann and Tiemannt3) using a very different methodology, also observed a tendency of some persons, but not others, to judge risks as high and benefits low, or vice versa. 1.2. Theoretical Issues The negative correlations found in previous studies suggest that people fail to consider the dimensions of risk and benefit separately. These negative correlations may be considered an interesting manifestation of the halo effect. The halo effect was first mentioned by Wellsc4) and later named by Thorndike.@) Halo occurs when individuals judge objects, people, or things in terms of general attitudes toward them. For example, when a person’s overall impression of another individual is favorable, then his or her perceptions of that indivi- dual’s attributes (such as intelligence, ability, physical appearance, etc.) tend to be favorable. Nisbett and Wil- sod6 manipulated people’s positive attitude toward an instructor using videotaped interviews with the same in- dividual. In one condition, the instructor was described as warm and friendly; in another condition, the instructor was described as cold and distant. The results indicated that persons who rated the “warm” instructor judged his appearance, mannerisms, and accent to be favorable, whereas those who rated the “cold” instructor found each of these attributes to be unfavorable.  loss Alhakami and Slovic The psychological literature contains several theo- ries that can explain the nonindependence among di- mensions (halo effect). These fall into three broad categories: (a) theories that explain the halo effect as a result of the way concepts (including risks) are repre- sented in the mind-cognitive consistency theories, for example; b) theories that explain halo as a result of the influence of attitudes and affect on cognition; and (c) theories explaining halo as due to the way information is processed. Cognitive consistency theories assert that people operate under a strong need for consistency among their Thus, when people consider an activity or technology to be beneficial they may, to be consistent, also tend to view the technology as having low risk. The halo effect may also be caused by people’s reliance on general evaluative attitudes or affective states when making riskbenefit judgments.(g) When the attitude is fa- vorable, the activity or technology being judged may be seen as having high benefit and low risk. On the other hand, when the item being evaluated is viewed unfavor- ably, with negative affect, it may be seen as having low benefit and high risk. Our general attitudes or affective states may thus “confound” the riskbenefit judgment. Information-processing theories(1o) mply that halo will be influenced by the familiarity with or knowledge of the technology and the concreteness or specificity of the dimensions people rate. Greater familiarity with what is being rated and greater specificity lead to less halo. The halo effect may also be a function of the salience or the availability of instances about risks and benefits.(’l) That is, when the benefit dimension is salient, it may inhibit the recall of instances of risk, and vice versa, leading to an inverse relationship and large disparity between risk and benefit judgments. 2. OBJECTIVES AND METHODS OF THE PRESENT STUDY As noted earlier, new analyses of the data from the Canadian survey by Slovic et al.@) ound negative cor- relations between perceived risk and perceived benefit across respondents for 32 of 33 items studied. These ranged between -.33 and -.03. In the present study we investigate whether such negative relationships will rep- licate in another set of items based on hazardous activ- ities and technologies. In addition, we attempt to identify the factors that determine the interdependence between risk and benefit judgments and to understand why some items have higher negative relationship than others. We shall also supplement correlational measures of relation- ship with measures of distance, defined above as the absolute value between scaled judgments of risk and benefit. High distance indicates that people view the item as having high benefit and low risk, or vice versa. 2.1. Methods Participants in this study were 100 students from the University of Oregon who participated to fulfill the requirements for an introductory psychology class. Each participant evaluated 40 different items (ac- tivities and technologies) with regard to their perceived benefits and perceived risks. These items were drawn from previous risk-perception research(1z) nd were se- lected to encompass a wide range of hazard types. Risk was judged first, for all 40 items, followed by benefit judgments. Risk was defined in terms of the overall risk to U.S. society. Specifically, the following question was asked: “In general, how risky do you consider each of the fol- lowing items to be for the United States society as a whole?” The scale ranged from 1) “not at all risky” to (7) “very risky.” Respondents were asked to answer the following question about benefits: “In general, how beneficial do you consider each of the following items to be for the United States society as a whole?” The scale ranged from (1) “not at all beneficial” to (7) “very beneficial.” After making their riskbenefit judgments, one half of the respondents were presented with a second ques- tionnaire that included 20 of the 40 hazard items. The other respondents received the remaining 20 items. This split-half method was used because of the length of this second task. Each item appeared on a separate page. Re- spondents were asked to judge each item against the set of 25 bipolar scales, shown in Table 11. In the test book- let, the item name (e.g., Nuclear Power, Pesticides, etc.) appeared at the top of the page of rating scales. The scales were selected to represent each of the three factors found by Osgood et ~l.(~) o determine the affective meaning of objects. Specifically, these were the evalu- ation factor, which is defined by scales such as good- bad and pleasant-unpleasant; the potency factor, which is defined by scales such as large-small and strong- weak; and the activity factor (e.g., active-passive, fast- slow). Several other scales were included to represent characteristics that have been found important in deter- mining perception and acceptance of risks (e.g., known- unknown, familiar-unfamiliar, old-new, dread-not dread, voluntary-mpulsory, fatal-not fatal, and con-  Inverse Relationship 1089 Table II. Rating Scales ~ 7; Bad ; ; ; 6; ood ; ; Nice ; ; ; ; 5; ; ; Awful Changing ; ; 3; ; ; ; ; Steady Known ; ; 3; ; ; 6; 7; Unknown Strong ; ; 3; ; ; 6; 7; Weak 6; 7; Fine oarse Dangerous ; ; 3; ; 5; ; ; Safe Poisonous Voluntary ; ; ; ; ; ; ; Compulsory Familiar ; ; ; 4; ; ; 7; Unfamiliar Acceptable ; ; ; ; ; ; ; Unacceptable Controllable ; . ; ; ; 6; ; Uncontrollable Useful ; ; 3; ; ; 6; ; Useless ; ; ; ; 5; 7; Calm . iolent 7; New ; ; 3; ; ; ; ld Active ; ; ; ; ; 6; ; Inactive 7; Beneficial ; ; ; ; ; ; armful 7; Lenient ; ; ; ; ; ; evere Pleasant ; ; ; ; 5; ; 7; Unpleasant 7; Powerless ; ; 3; ; ; ; owerful 7; Valuable Worthless ; ; ; ; ; 6; 7; Unfair Fair ; ; 3; ; 5; ; ; Not fatal Fatal ; ; ; ; 5; 6; ; ; 3; ; 5; ; ; ; ; ; ; ; Not poisonous Pleasurable ; ; ; ; 5; ; 7; Painful 1; ; ; ; ; 6; ; Not dread read trollable-uncontrollable). Respondents received the fol- lowing instructions [adapted from Osgood et a1J9 ]: The second part of this experiment measures the mean- ings of different items (technologies and activities) to you by asking you to judge them against a series of descriptive scales. In performing this task, please make your judgments on he basis of what these technologies and activities mean to you. On each page of this booklet you will find an item to be judged and beneath it a set of scales. You are to rate the item on each of these scales in order. 2.2. Analysis 1. Two measures of the riskbenefit relationship were calculated, one based on correlations and the other on distances: (a) Correlations between risk and benefit judgments were computed across respondents for each activity and technology, and b) the distance between risk and benefit judgments was computed for each re- spondent by taking the absolute difference between the two judgments. These absolute differences were then av- eraged across respondents for each item. 2. The mean of each of the 25 bipolar scales was calculated for each item. 3. Factor analysis was performed on the intercor- relations among the means of the 25 bipolar scales to determine the factors that contributed significant vari- ance to the judgments on these scales. 4. Regression analyses were performed to deter- mine the factors that affect the magnitude and the direc- tion of the correlation between risk and benefit judg- ments and to determine the factors that affect the distance between risk and benefit judgments. 3. RESULTS 3.1. Correlations and Distance Measures Table I11 presents the correlations across respon- dents between perceived risk and perceived benefit for each of the 40 items used in this study. Negative cor- relations were obtained for 38 items. Twenty-six corre- lations were significant at the 01 level. The highest negative correlations were 52 for water fluoridation, -.51 for herbicides, and -SO for DDT. Items that had low negative correlations included surgery, policy work, home appliances, fire fighting, prescription drugs, lasers, radiation therapy, air travel, bicycles, railroads, display screens, and so on.
Similar documents
View more...
Search Related
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks