Health Education Research, Vol. 16, No. 5, 533-540,
October 2001
© 2001 Oxford University Press
Is optimistic bias influenced by control or delay?
School of Psychology, Faculty of Health and Behavioral Sciences, Deakin University, Geelong, Victoria 3217, Australia
Correspondence to: V. A. Clarke
 |
Abstract |
|---|
 Top Abstract IntroductionMethodResultsDiscussionReferences |
|---|
Optimistic bias is a commonly observed but poorly explained phenomenon. Our aim was to determine whether optimistic bias varied according to the nature of the event. Two event characteristics were explored: control and delay. A sample of 100 participants aged 18–30 years was randomly selected from the local residential telephone directory. Respondents were interviewed over the telephone. The highly structured interview schedule assessed respondents' perceptions of their own risk, and the risk of an average person of their age and sex for experiencing four negative life events: developing skin cancer, being involved in a serious car accident as the driver, being involved in a serious car accident as a passenger and having to wear a hearing aid. It also assessed respondents' perceptions of control and delay for each event. Data analysis using a repeated-measures MANOVA showed that optimistic bias occurred for all four events. Optimistic bias was significantly greater for the two events high in control (skin cancer and accident as the driver) than for those low in control (accident as a passenger and hearing aid). Delay was not related to the magnitude of optimistic bias. These findings have implications for health promotion campaigns and self-protective behaviors.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Optimistic bias refers to the pervasive phenomenon of perceiving oneself as being less likely than an average other to experience negative events (Weinstein, 1980
). On an individual level, this perception may be realistic (Hoorens, 1996), as hereditary factors or behavioral interventions may reduce an individual's susceptibility. On the group level, it is logically impossible for everyone's chances of the occurrence of a negative event to be below average.
The demonstration of unrealistic optimism may have both beneficial and harmful consequences. Unrealistic optimism may be beneficial in that it may aid in maintaining a relatively high level of self-esteem (Perloff, 1988). It is also suggested that people are motivated to see themselves as invulnerable in order to reduce anxiety (Weinstein, 1982). This may be because vulnerability often creates symptoms of emotional distress such as acute anxiety, depression, helplessness and excessive fear. In addition to maintaining a high self-esteem and reducing anxiety, illusions of invulnerability may be adaptive because they enable people to go about their daily life without being overcome by fear (Perloff, 1988).
However, although unrealistic optimism can be beneficial, when things do not go in the expected direction, the demonstration of unrealistic optimism may also be dysfunctional because individuals assert that they are less likely than others to experience illness, injury and other negative events, and therefore these beliefs may interfere with the individual's taking of precautions to reduce their risk (Weinstein, 1980Weinstein, 1984Weinstein, 1987; Harris and Middleton, 1994). Numerous authors have reported a positive correlation between personal optimism, endangering behavior and neglect of precaution [e.g. (Weinstein, 1982Weinstein, 1983; Dolinski et al., 1987)]. For example, Dolinski et al. reported that after the Chernobyl disaster, Polish students who perceived themselves as invulnerable to radiation were less likely to report taking precautions, relative to students who thought that they were at no less risk than others (Dolinski et al., 1987). This perception of invulnerability may lead people to ignore legitimate risks in their environment and to fail to take measures to offset those risks. If people believe that negative events are less likely to happen to them, then perhaps they will pay less attention to risk-related information and will thus be less likely to engage in self-protective behaviors. An awareness of optimistic bias and of factors that influence this phenomenon is important for those involved in health promotion.
Optimistic bias has been found for many negative events, including skin cancer (Clarke et al., 1997), automobile accidents (Greening and Chandler, 1997), sexually transmitted diseases (Cohn et al., 1995), attempting suicide (Weinstein, 1987), heart disease, emphysema and lung cancer (Williams and Clarke, 1997), and risky behaviors, such as having unprotected sex (Cohn et al., 1995).
Although optimistic bias has been found with a wide range of negative events, little attempt has been made to identify the characteristics or the events that yield differing amounts of optimistic bias. A recent study of both positive and negative events explored respondents' reasons for the differences in their two absolute ratings (Gouveia and Clarke, 2001). It was found that most explanations reflected the respondents' belief that s/he exerted a greater control over his/her behavior. Existing research addressing the issue of control has yielded conflicting findings, with some studies finding that it is an important factor [e.g. (Weinstein, 1980; McKenna, 1993; Harris, 1996; Klar et al., 1996)] and others finding little or no support [e.g. (Weinstein, 1982Weinstein, 1983Weinstein, 1984Weinstein, 1987; Harris and Middleton, 1994)]. This issue will be explored further in the current study.
Most studies of optimistic bias focus on perceived risk. However, recent research has shown that optimistic biases occur for other health belief variables including age of onset of the disease. This had been found both for women in relation to breast cancer and for men in relation to prostate cancer (Clarke et al., 2000). People consider that others will experience the negative event sooner than they will personally experience it. Similarly, studies of younger adults show that they estimate the age of onset of skin cancer as later for themselves (Clarke et al., 1997). Shifting the focus from the self–other judgement to the nature of the event, the present study explores the issue of delay further by examining the amount of optimistic bias shown for an event perceived to have a delay of onset.
Perceived delay of onset refers to an individual's belief regarding the length of the time between engaging in a risky behavior and the occurrence of the negative event or outcome associated with that risky behavior. The occurrence of an event is seen as having a long delay when there is a long time-span between a risky behavior and the consequence of that behavior. For example, a fairly lengthy time period is associated with lung cancer due to the uptake of smoking (Williams and Clarke, 1997). An event perceived to have a short delay can be characterized as an event where the consequences of a risky behavior occur soon after the risk is performed.
Research has suggested that the risks associated with negative events can be separated into long- and short-term risks. For example, Moore and Rosenthal claimed that for some events there are lengthy intervals, or ‘lead times, between risky behaviors (e.g. sunbathing without protection) and their consequences (e.g. skin cancer), while other events such as car accidents have little time delay between the risky behavior (driving recklessly) and the subsequent effect of that behavior (car accident) (Moore and Rosenthal, 1992). Moore and Rosenthal found that differences in ‘lead times’ (or delay of onset) did not affect risk perceptions (Moore and Rosenthal, 1992). More specifically, they reported that similarly sized optimistic biases were produced for the long-term risk of skin cancer and the short-term risk of having a car accident.
Optimistic bias has been measured using either comparative or absolute judgements. Comparative judgements involve asking participants to compare their risk to the risk of another individual in relation to a particular event (Hoorens, 1995). Absolute judgements involve asking participants to rate their own risk in relation to an event and then to rate the risk of an average person (Hoorens, 1995). Questions phrased in absolute terms do not require respondents to make a direct comparison and so are less susceptible to the comparative bias (Hoorens and Buunk, 1993). Although absolute judgements have been used less frequently than comparative judgements (Strecher et al., 1995), they were used in the present study as they provide a more conservative measure.
The present study
The aim of the present study was to use highly structured telephone interviews with a random sample of young adults to determine whether optimistic bias assessed using absolute judgements for four negative events varied according to the control of the event or the delay of onset of the negative consequence. The variables of control and delay of onset were manipulated through the selection of events. Four negative life events were selected: developing skin cancer, being involved in a serious car accident as the driver, being involved in a serious car accident as a passenger and having to wear a hearing aid in the future. Skin cancer and having an accident as the driver are more controllable than having to wear a hearing aid or having an accident as a passenger. Developing skin cancer and wearing a hearing aid are more distal than the two accident conditions (passenger and driver). The perceptions of control and delay of onset were validated through ratings of the delay of onset and controllability of those events.
|
Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Selection of participants and participants
The only selection criterion for participants was that they be between 18 and 30 years of age. Persons under 18 require parental consent to participate in a telephone interview, hence it was not expedient to interview them. The upper age was restricted to 30 years of age to ensure a considerable delay before the onset of the distal events (getting skin cancer, wearing a hearing aid). Telephone numbers of potential respondents were randomly selected from the Geelong residential telephone directory, using the program ‘Australia On Disk’, November 1997 version. There were 100 participants. Of these, 52 were female and 48 were male. Their ages ranged from 18 to 30 years, with a mean age of 23.5 years (SD 3.7 years).
Interview schedules
A highly structured interview schedule was used to assess people's perceptions of risk, control and delay of onset of negative events. There were two interview schedules, counterbalanced to eliminate order effects: one schedule asked participants to rate their own risk first; one asked them to rate the risk of an average person of their age and sex first. Each schedule consisted of four components: demographics (age and sex), questions about perceived likelihood of occurrence for the self and for an average person of the same age and sex for the four negative events, questions relating to perceived control, and questions assessing perceptions of delay of onset. Each interview took approximately 10 min to complete.
Measures
Perceived likelihood ratings for the self and for an average person
Perceived likelihood of occurrence of each event was measured by asking respondents to estimate the risk of experiencing the event, both for themselves and for an average person of their age and sex. When estimating likelihood, respondents were asked to use a scale ranging from 0 to 100, where 0 meant no risk and 100 meant that the event would certainly happen.
Control
Perceptions of control were measured by asking respondents to estimate the degree of control they perceived themselves to have over experiencing each of the four events. Perceived control was rated on a scale ranging from 0 to 100, where 0 meant no personal control over the event and 100 meant that the event was under complete personal control.
Delay
Perceptions of delay were assessed by asking respondents to estimate the age at which they thought they would be most likely to experience each of the four events. Delay ratings were calculated by subtracting the respondent's current age from the age they specified for each event. Responses greater than the average life expectancy, 80 years (American Cancer Society, 1998), were scaled down to 80 years, as were responses that suggested that an individual would ‘never’ experience a particular event. Responses that indicated that the event could occur at ‘anytime’, ‘now’ or ‘whenever’, or at an age younger than the respondent's current age were recoded with a value of ‘0’, indicating that there was no delay between a respondent's current age and the time at which they may experience that particular event. Fewer than 10% of responses required recoding of imprecise responses.
|
Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Response rate
Of 620 telephone numbers dialed, 112 yielded an eligible person. Of those eligible, 12 chose not to participate, resulting in a consent rate of 89.3%.
Descriptive statistics
Table I presents the mean ratings for perceptions of risk, perceived control and perceived delay, for each of four events. Lower means represent lower perceptions of risk, less control and shorter delay. An examination of the mean scores for self- and average-risk estimates shows that participants rate their own risk of experiencing each of the four events as lower than the risk of an average person of their age and sex.
|
Data screening
The data were screened to ensure that the statistical assumptions of normality, linearity and homogeneity of variance–covariance were satisfied. Distributions that were skewed as a result of univariate outliers were recoded back to the value next closest to the mean for that distribution. Distributions that remained skewed after the recoding of univariate outliers were transformed as recommended by Tabachnick and Fidell (Tabachnick and Fidell, 1996
).
Validation of assumptions of control and delay of target events
Examination of the mean scores for control and delay for each of the four events shows that the events differ on these dimensions (see Table I
). Four variables were computed to confirm the assumption that the event differed in perceived control and delay. Each new variable represented an average value for events selected as high or low in control and either long or short in delay. For example, the first variable, high control, was the mean of the scores for the two events perceived as being high in control (skin cancer and accident as a driver). To assess whether these mean scores were significantly different, paired sample t-tests were used. A Bonferroni adjustment resulted in a criterion level of P = 0.025. As predicted, high-control events (skin cancer and car accident as a driver) were significantly different from the low-control events (hearing aid and car accident as a passenger), t(99) = 3.15, P < 0.025. Similarly, long-delay events (skin cancer and hearing aid) differed significantly from short-delay events (accident as a passenger and accident as a driver), t(99) = 18.57, P < 0.025.
Possible confounding variables: age, sex and presentation order
The data were analyzed to assess the effects of demographic and procedural variables on optimistic bias. A three-way (2x2x2) multivariate analysis of variances (MANOVA) was conducted with age, sex and counter-balanced presentation order as the independent variables. Age was dichotomized into two groups, 18–23 and 24–30 years, using the median split. The dependent variables were the mean difference scores between own risk and average risk for each of the four negative events. There were no significant effects, so age, sex and presentation order were omitted from further analyses.
Occurrence of optimistic bias
The presence of an optimistic bias was identified using a repeated measures MANOVA. There was one within-subjects factor, optimistic bias, which contrasted estimates of self-risk and average-risk for each of the four negative events. All four events were entered as measures in the analysis to enable an assessment of optimistic bias for each event separately. A main effect for self- and average-risk estimates, F(4,96) = 40.55, P < 0.05, provided evidence for an overall optimistic bias. The overall difference between self-risk and average-risk estimates explained 63% of the variance in risk perceptions (see Table II). Across all four events, participants' estimates of their own risk of experiencing the negative events were lower than their estimates of the risk of an average person of their age and sex. To maintain the type 1 error rate at 0.05, a Bonferroni adjustment was used for all univariate analyses, reducing the familywise 
level from 0.05 to 0.0125.
|
Examination of the results of the univariate F-tests in Table II
shows that optimistic biases occurred for all four negative events. Relative to the average person of the same age and sex, respondents saw themselves as having a lower risk of: developing skin cancer, being involved in a car accident as a passenger, having to wear a hearing aid and being involved in a car accident as the driver.
The impact of perceived controllability and perceived delay on optimistic bias
The effects of control and delay on optimistic bias were assessed using a repeated measures MANOVA. There were three within-subjects factors: optimistic bias, control and delay. Each of the within-subjects factors had two levels: optimistic bias (self versus average risk); control (high versus low) and delay (long versus short).
There was a significant overall main effect for control, F(1,99) = 12.86, P < 0.05, with the linear combination of dependent variables for the difference between risk ratings for events perceived to be high and low in control, explaining 12% of the variance. The main effect for delay was not significant, F(1,99) = 0.96, P > 0.05. There was a significant interaction between optimistic bias and control revealing that optimistic bias was greater for the events perceived to be high in control (skin cancer and accident driver) than for the events perceived as being low in control (accident as a passenger and hearing aid). The interaction between optimistic bias and delay was not significant, F(1,99) = 0.06, P > 0.05, indicating that the magnitude of optimistic bias was not affected by long or short delay. The three-way interaction between optimistic bias, control and delay was also non-significant, F(1,99) = 2.88, P > 0.05.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Participants rated their own risk of experiencing each of the negative events as lower than the risk of an average person for all four negative events. There were no significant differences in relation to gender, age or order of presentation, indicating the pervasiveness of the optimistic bias effect across events for all categories of participants. However, the large standard deviations obtained for the difference scores for all four of the events indicates that there are considerable individual differences. Thus, optimistic bias occurs for all of the events, but not necessarily for all participants.
An overall optimistic bias occurred across the negative events showing that participants perceived their personal risk of experiencing the negative events as being lower than the risk of an average person of their age and sex. This finding was strong, explaining 63% of the variance, and is consistent with previous research [e.g. (Weinstein, 1980; Clarke et al., 1997Clarke et al., 2000; Williams and Clarke, 1997) supporting the presence of an overall optimistic bias. The magnitude of the overall effect of optimistic bias (explaining 63% of the variance) contrasts with the relatively small impact of control (explaining 12% of the variance) and the absence of a significant effect for delay. This indicates that optimistic bias is a pervasive phenomenon that appears likely to occur regardless of the specific characteristics of individual events.
Events high in control elicited slightly stronger optimistic biases than events low in control. The discrepancies between risk ratings for the self and for an average person of the same age and sex were greater for skin cancer and being involved in a car accident as the driver, than for being involved in a car accident as a passenger and having to wear a hearing aid. These findings are more conclusive than those noted in the inconclusive research cited earlier as the current study directly assessed the perceived control of the events before demonstrating the impact of control on optimistic bias. Further, the present study used a sample from the general population rather than being limited to a student sample. However, research including qualitative comments enabling participants to explain the differences between their ratings for themselves and for others (Gouveia and Clarke, 2001) suggests that people perceive that they have a greater control over events than does the average person, as they assume that the average person does not take the precautions they personally adopt. This explanation is equally applicable in the present study where participants could argue that they are less likely to get skin cancer as they remain out of the sun and/or protect themselves with clothing or sunscreen, and that they are less likely to have an accident while a driver as they are good drivers. In both cases it is assumed that the average other does not engage in sun-protective behavior and is not a good driver. In the present study the amount of difference between self-risk and average-risk ratings shown in relation to being involved in a car accident as a driver (21.05%) was twice that observed for the passenger condition (9.35%). This is consistent with differences between these two conditions in McKenna's study that used comparative rather than absolute judgements, a study in which the differences were attributed to the illusion of control rather than to optimistic bias (McKenna, 1993).
Given that all events elicited a difference in self-average rating and events high in control elicited a greater difference than events low in control, it is possible that the individual differences in the gap between self-average ratings might reflect individual differences in perceived control. This individual difference factor needs to be explored more fully in later research. This might enable the isolation of a group that is at greater risk than is the so-called average member of the population.
The magnitude of optimistic bias did not differ between proximal and distal events. These findings are consistent with those of Moore and Rosenthal who reported that risk perceptions in first-year psychology students were not affected by whether events were seen as having long or short ‘lead times’ (Moore and Rosenthal, 1992). This is an interesting finding as it demonstrates that, in relation to delay of onset, the characteristic of the event is not related to optimistic bias, although optimistic bias has been shown to occur in people's estimates of the delay in onset of an event (Clarke et al., 1997Clarke et al., 2000).
Practical implications
The results of the present study suggest that people perceive their risks of experiencing four negative events as lower than the risk of an average person of their age and sex. This may affect the self-protective behaviors (Weinstein, 1989; McKenna, 1993; Horswill and McKenna, 1999), the risk-taking behavior of individuals and their response to health promotion messages. For example, if people believe that their risk of being involved in a serious car accident is lower than the risk of others, they may pay less attention while driving. Furthermore, if people do not acknowledge their own susceptibility to harm, then educational campaigns, such as those of the Transport Accident Commission, will be ineffective in changing people's behavior. Similarly, if people think that they are unlikely to be involved in a car accident, then they may not be motivated to engage in behaviors that decrease the likelihood of being involved in such an event (e.g. avoid speeding, tailgating and drinking alcohol and driving). Believing that one's chances of having an accident are minimal, people may also be less inclined to engage in activities which can lessen the severity of an injury if an accident was to occur (e.g. wearing a seat belt). The illusion that people do not need to protect themselves from something that is not going to happen anyway may also adversely affect campaigns aimed at increasing precautionary behavior. For example, the information provided in mass media campaigns, such as the Drink-Driving campaign of the Transport Accident Commissions and National Cancer Council Sun-Smart campaigns, may not be heeded by the population, because the majority of people may see the messages portrayed in these campaigns as being directed at people who are more vulnerable than themselves. The effectiveness of these campaigns may be improved if they emphasize that although people may have the ability to control the risks of negative events, the risks of experiencing such events can only be decreased if this control is exerted. For example, people generally have the ability to control the risks of developing skin cancer (e.g. do not intentionally burn your skin, when outside wear a hat, a T-shirt and sunscreen); however, one's risk of developing skin cancer will not decrease unless s/he actually engages in these risk-decreasing behaviors. Further, the extent to which the majority of people engage in these behaviors needs to be more strongly emphasized.
|
References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
American Cancer Society (1998) American Cancer Society Homepage. http://www.cancer.org/cancerinfo/acs
Clarke, V. A., Williams, T. and Arthey, S. (1997) Skin type and optimistic bias in relation to the sun protection and suntanning behaviors of young adult. Journal of Behavioral Medicine, 20, 207–222.[ISI][Medline]
Clarke, V. A., Lovegrove, H., Williams, A. and Macpherson, M. (2000) Unrealisltic optimism and the health belief model.Journal of Behavioral Medicine, 23, 367–376.[ISI][Medline]
Cohn, L. D., Macfarlane, S., Yanez, C. and Imai, W. K. (1995) Risk perception: differences between adolescents and adults. Health Psychology, 14, 217–222.[ISI][Medline]
Dolinski, D., Wojciech, G. and Zawisza, E. (1987) Unrealistic pessimism. Journal of Social Psychology, 127, 511–516.
Greening, L. and Chandler, C. C. (1997) Why it can't happen to me: the base rate matters, but overestimating skill leads to underestimating risk. Journal of Applied Social Psychology, 27, 760–780.
Gouveia, S. O. and Clarke, V. A. (2001) An explanation of optimistic bias in relation to positive and negative events. Health Education, in press.
Harris, P. (1996) Sufficient grounds for optimism? The relationship between perceived controllability and optimistic bias. Journal of Social and Clinical Psychology, 15, 9–52.
Harris, P. and Middleton, W. (1994) The illusion of control and optimism about health: on being less at risk but no more in control than others. British Journal of Social Psychology, 33, 369–386.
Hoorens, V. (1995) Self-favoring biases, self-presentation, and the self-other asymmetry in social comparison. Journal of Personality, 63, 793–817.
Hoorens, V. and Buunk, B. P. (1993) Social comparison of health risks: locus of control, the person-positivity bias, and unrealistic optimism. Journal of Applied Social Psychology, 23, 291–302.
Horswill, M. S. and McKenna, F. P. (1999) The effect of perceived control on risk taking. Journal of Applied Social Psychology, 29, 377–391.
Klar, Y., Medding, A. and Sarel, D. (1996) Nonunique invulnerability: singular versus distributional probabilities and unrealistic optimism in comparative risk judgments. Organizational Behavior and Human Decision Processes, 67, 229–245.
McKenna, F. P. (1993) It won't happen to me: unrealistic optimism or illusion of control? British Journal of Psychology, 84, 39–50.
Moore, S. M. and Rosenthal, D. A. (1992) Australian adolescents' perceptions of health-related risks. Journal of Adolescent Research, 7, 177–191.[Abstract]
Perloff, L. S. and Fetzer, B. K. (1986) Self other judgements and perceived vulnerability to victimization. Journal of Personality and Social Psychology, 50, 502–510.[ISI]
Strecher, V. J., Kreuter, M. W. and Kobrin, S. C. (1995) Do cigarette smokers have unrealistic perceptions of their heart attack, cancer, and stroke risks? Journal of Behavioral Medicine, 18, 45–54.[ISI][Medline]
Tabachnick, B. G. and Fidell, L. S. (1996) Using Multivariate Statistics, 3rd edn. Harper Collins, New York.
Weinstein, N. D. (1980) Unrealistic optimism about future life events. Journal of Personality and Social Psychology, 39, 806–820.[ISI]
Weinstein, N. D. (1982) Unrealistic optimism about susceptibility to health problems. Journal of Behavioral Medicine, 5, 441–460[ISI][Medline]
Weinstein, N. D. (1983) Reducing unrealistic optimism about illness susceptibility. Health Psychology, 2, 11–20.
Weinstein, N. D. (1984) Why it won't happen to me: perceptions of risk factors and susceptibility. Health Psychology, 3, 431–457.[ISI][Medline]
Weinstein, N. D. (1987) Unrealistic optimism about susceptibility to health problems: conclusions from a community-wide sample. Journal of Behavioral Medicine, 10, 481–498.[ISI][Medline]
Weinstein, N. D. (1989) Effects of personal experience on self-protective behavior. Psychological Bulletin, 105, 31–50.[ISI][Medline]
Williams, T. and Clarke, V. A. (1997) Optimistic bias in beliefs about smoking. Australian Journal of Psychology, 49, 106–112.
Received on October 10, 2000; accepted on March 13, 2001
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. D. Windsor, K. J. Anstey, and J. G. Walker Ability Perceptions, Perceived Control, and Risk Avoidance Among Male and Female Older Drivers J. Gerontol. B. Psychol. Sci. Soc. Sci., March 1, 2008; 63(2): P75 - P83. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||