Introduction
Western cultures are witnessing rapidly increasing numbers of older people, many of whom report sexual interest, desire, and capacity for sexual activity that extends far into later life (Kontula & Haavio-Mannila, Reference Kontula and Haavio-Mannila2009; Lindau et al., Reference Lindau, Schumm, Laumann, Levinson, O’Muircheartaigh and Waite2007; Lindau & Gavrilova, Reference Lindau and Gavrilova2010; Nicolosi et al., Reference Nicolosi, Laumann, Glasser, Edson, Moreira and Paik2004). Despite the many positive health and well-being outcomes associated with maintaining a sexual life for older adults (Weeks, Reference Weeks2002), sexual expression in later life often receives little support. Young adults appear to hold the most negative views of sexuality in older adults compared to children and middle-aged adults (Gordon & Avery, Reference Gordon and Avery2004; Hagestad & Uhlenberg, Reference Hagestad and Uhlenberg2005; Kite, Stockdale, Whitley, & Johnson, Reference Kite, Stockdale, Whitley and Johnson2005). Older adults are defined here as individuals 65 years or older, in line with related research (Lai & Hynie, Reference Lai and Hynie2011; Sharpe, Reference Sharpe2004), whereas young adults refer to those aged 18 to 24 (White, Hingson, Pan, & Yi, Reference White, Hingson, Pan and Yi2011). Historically, young adults’ attitudes towards the sexuality of older adults have been primarily negative (Angiullo, Whitbourne, & Powers, Reference Angiullo, Whitbourne and Powers1996; Palmore, Reference Palmore2001). However, in keeping with a general trend towards more liberal, permissive, and accepting sexual attitudes among young people (Andersen & Fetner, Reference Andersen and Fetner2008; Treas, Reference Treas2002), some recent surveys of young adults suggest a movement towards more positive attitudes (Bouman, Arcelus, & Benbow, Reference Bouman, Arcelus and Benbow2006; Lai & Hynie, Reference Lai and Hynie2011; Pratt & Schmall, Reference Pratt and Schmall1989).
People who endorse positive attitudes towards the sexual lives of older people frequently reveal incongruent behaviours that hinder sexual expression or reflect avoidance, opposition, or disapproval (Mahieu, Van Elssen, & Gastmans, Reference Mahieu, Van Elssen and Gastmans2011). For example, in their review, Mahieu et al. (Reference Mahieu, Van Elssen and Gastmans2011) found that despite reporting positive and accepting attitudes towards the sexuality among older adults, nurses often responded unfavourably to the sexual behaviour of nursing home residents by mocking the residents, or using physical barriers to prevent sexual interactions. The inconsistency between attitudes and behaviour may reflect reliance on measures of explicit attitudes – that is, measures assessing participants’ self-reports. Self-reports often reflect socially desirable responding: in this case, to express positive attitudes towards sexuality and aging. One method for overcoming these inherent demand characteristics (i.e., changes in a participants’ behaviour because of their awareness of what the researcher is investigating) is to assess implicit attitudes – those attitudes operating beyond conscious awareness. In the current study, we identified and compared young adults’ explicit and implicit attitudes towards the sexuality of older adults.
Explicit Attitudes towards Sexuality among Older Adults
Ageism refers to prejudice and discrimination towards older persons generally (Butler, Reference Butler1969). Even young children express aversion to and demonstrate avoidance of older adults (Isaacs & Bearison, Reference Isaacs and Bearison1986). People are especially likely to hold ageist attitudes regarding sexual expression because of pervasive societal messages that link sexuality with youth and fertility (Sharpe, Reference Sharpe2004). In addition, many people hold narrow and/or negative views towards diversity in sexual expression generally (Tebbe & Moradi, Reference Tebbe and Moradi2012; Vincent, Parrott, & Peterson, Reference Vincent, Parrott and Peterson2011), which likely influence their views of sexuality among older adults. Even people who hold positive attitudes towards older adults and towards sexuality generally still may reveal negative attitudes towards sexuality among older adults (Aizenberg, Weizman, & Barak, Reference Aizenberg, Weizman and Barak2002; Bouman & Arcelus, Reference Bouman and Arcelus2001; O’Hanlon & Brookover, Reference O’Hanlon and Brookover2002; Tomlinson, Reference Tomlinson1998; Walker & Harrington, Reference Walker and Harrington2002; Walker, Osgood, Richardson, & Ephross, Reference Walker, Osgood, Richardson and Ephross1998). To gain insights into young people’s attitudes towards sexual expression in older adults, it is important to assess attitudes towards both sexual expression (i.e., negativity towards sex) and older adults generally (i.e., ageism).
Surveys of younger cohorts have revealed fairly consistent and pervasive negative attitudes regarding the sexuality of older people (Fraboni, Saltstone, & Hughes, Reference Fraboni, Saltstone and Hughes1990; Rupp, Vodanovich, & Credé, Reference Rupp, Vodanovich and Credé2005; Stuart-Hamilton & Mahoney, Reference Stuart-Hamilton and Mahoney2003). Young adults especially have long viewed the sexuality of older people as “ugly ... grotesque, and incongruous” (Bouman et al., Reference Bouman, Arcelus and Benbow2006, p. 151) or non-existent altogether (Sharpe, Reference Sharpe2004; West, Reference West1975). In one of the first studies to address this topic, college students’ attitudes were assessed using a sentence-completion task (Golde & Kogan, Reference Golde and Kogan1959). Almost all students indicated that sex for older adults was unimportant, inappropriate, or irrelevant.
A few studies, however, have found that younger generations express generally positive attitudes towards the sexuality of older adults, or at least reveal no signs of ageism. In one study, 293 college students evaluated a scenario depicting either younger or older adults (age 25 vs. 75) engaged in eight sexual behaviours using semantic differential scales (e.g., “good-bad,” “nice-awful”; Pratt & Schmall, Reference Pratt and Schmall1989). The target’s age had little to no impact on the students’ evaluations indicating that there was no systematic bias against older adults. In another study, 305 young adult university students rated younger (their own age) and older (65 years or older) men’s and women’s likely interest in a range of sexual activities (Lai & Hynie, Reference Lai and Hynie2011). Participants perceived older targets to be significantly less interested in the sexual activities than the younger targets. However, both older men and women were rated as somewhat to very interested in both traditional and experimental forms of sexual activity.
A qualitative study of 277 university students had participants respond to a film about the sexual lives of women over age 65 (Allen & Roberto, Reference Allen and Roberto2009). Students’ narratives indicated that they endorsed views supportive of expressing sexual needs no matter one’s age. An online survey asked 606 college students to rate the acceptability of various sexual behaviours for women in response to vignettes describing the sexual lives of women at age 75 (Allen, Petro, & Phillips, Reference Allen, Petro and Phillips2009). Participants reported moderately high levels of acceptability as long as the women were depicted as being both physically and mentally healthy. These findings suggest a trend among young people to endorse more-permissive sexual attitudes compared to views expressed in the past (Treas, Reference Treas2002).
Are young adults’ attitudes towards the sexuality of older adults generally positive as indicated by these studies? Are young people becoming less ageist with respect to sexual expression? Researchers have encountered difficulties in detecting ageism, particularly in studies that have relied on self-reports of attitudes because of the social pressures to disavow prejudicial views (De Paula Couto & Wentura, Reference De Paula Couto and Wentura2012; Levy & Banaji, Reference Levy, Banaji and Nelson2002). Each of the aforementioned attitude studies used self-report methodologies and often required participants to make direct comparisons between older and younger targets. Thus, these studies may have over-estimated positive attitudes towards the sexuality of older adults by invoking strong self-presentation or social desirability effects in respondents.
Measuring Implicit Attitudes
Greenwald and Banaji (Reference Greenwald and Banaji1995) have criticized the use of explicit attitudes because they require cognitive effort, are moderated by higher order processes, and are plagued by issues of socially desirable responding, resulting in overly positive depictions of internalized views. Explicit attitudes are particularly problematic when addressing socially sensitive topics (O’Sullivan, Reference O’Sullivan2008), such as attitudes towards sexuality and aging. By contrast, implicit attitudes are involuntary and effortless responses to stimuli occurring outside of conscious awareness (Greenwald & Banaji, Reference Greenwald and Banaji1995; Turner & Fischler, Reference Turner and Fischler1993). Implicit attitude measures assess automatic responses to stimuli, overcoming limitations of self-presentation inherent in explicit measures. Research suggests that implicit attitude measures are more predictive of behaviour outcomes than are explicit measures in such diverse areas as prejudice and racism (Greenwald, McGhee, & Schwartz, Reference Greenwald, McGhee and Schwartz1998), weight discrimination (Chambliss, Finley, & Blair, Reference Chambliss, Finley and Blair2004), and disability biases (Pruett & Chan, Reference Pruett and Chan2006). Because explicit attitudes represent the views that individuals are able to access consciously and are willing to report, these attitudes typically only partially resemble implicit attitudes – that is, attitudes that are operating beyond conscious awareness. Indeed, researchers have found that explicit and implicit attitudes in other domains are typically only moderately correlated (Greenwald, Nosek, & Banaji, Reference Greenwald, Nosek and Banaji2003).
In this current study, we employed the Implicit Association Test (IAT) to assess young adults’ implicit attitudes towards sexuality among older adults. The IAT was created to measure automatic associations (i.e., preferential responses) between counterbalanced target stimuli by using a number of response discrimination tasks (Greenwald et al., Reference Greenwald, McGhee and Schwartz1998). Faster sorts between two stimuli indicate a bias towards, or greater ease in making, an association between those stimuli, whereas slower sorts indicate more difficulty (cognitive interference) in making links between two stimuli. Two studies have used the IAT to study attitudes towards sexuality (Geer & Robertson, Reference Geer and Robertson2005; Thompson & O’Sullivan, Reference Thompson and O'Sullivan2012) and found that both men and women have negative implicit attitudes to sexual words and pictures. In addition, three studies that used the IAT to investigate attitudes towards aging among young adults (Gast & Rothermund, Reference Gast and Rothermund2010; Jelenec & Steffens, Reference Jelenec and Steffens2002; Nosek, Banaji, & Greenwald, Reference Nosek, Banaji and Greenwald2002) found that young people have an automatic bias favouring stimuli depicting younger as compared to older adults. Researchers have examined implicit attitudes in relation to other areas of sexuality, such as attitudes towards homosexuality (Jonathan, Reference Jonathan2008), and the sexual double standard (Sakaluk & Milhausen, Reference Sakaluk and Milhausen2012), but not with regard to general sexual attitudes, nor with regard to attitudes towards the sexuality of older adults specifically.
In our study, we created two IATs to assess attitudes towards aging (i.e., ageism); one assessed attitudes towards older adults generally (the Aging Nonsexual IAT) and the other towards older adults engaged in sexual activities specifically (the Aging Sexual IAT). We hypothesized that participants would demonstrate an automatic bias favouring younger over older adults for both IATs (hypothesis H1). However, because sexuality is a sensitive social topic and thus likely to generate some discomfort relative to more generalized topics, we expected that this difference would be greater for sexual than for non-sexual activities for both men and women (hypothesis H2).
We also created two separate IATs to assess sexual attitudes (i.e., negativity towards sex); one assessed attitudes towards sexual activities in young adults (the Sexual Young Adult IAT, and the other assessed attitudes towards sexual activities among older adults (the Sexual Older Adult IAT). Again, because sex is a socially sensitive topic, we expected a bias in favour of general as compared to sexual activities for both younger and older adults (hypothesis H3). However, because of societal messages linking sexuality and youth, we expected that this bias would be greater for older targets as compared to younger ones (hypothesis H4). In addition, in keeping with gender role socialization, we expected that the bias in favour of general over sexual activities would be greater for women than for men (hypothesis H5). Women tend to demonstrate less tolerance for (and more negative attitudes overall towards) the sexuality of older adults compared to men (e.g., Allen et al., Reference Allen, Petro and Phillips2009). These findings correspond to research indicating that women generally hold more conservative (less permissive) sexual attitudes compared to men (Oliver & Hyde, Reference Oliver and Hyde1993; Petersen & Hyde, Reference Petersen and Hyde2010).
Predictors of Attitudes towards Sexuality among Older Adults
We also sought to examine whether general attitudes towards aging and/or attitudes towards sexuality are associated with explicit and implicit attitudes towards sexuality and older adults. There is some evidence that both types of attitudes are associated with explicit attitudes towards sexuality and older adults, although researchers have not examined their association with implicit attitudes. For example, sexual attitudes have been shown to predict greater acceptance among college students of the sexual expression of older adults (Allen et al., Reference Allen, Petro and Phillips2009). Similarly, ageist attitudes have been shown to be related to traditional views about sexual interests among older adults (Lai & Hynie, Reference Lai and Hynie2011). Our final hypothesis (H6) was that individuals with more-permissive aging and sexual attitudes would demonstrate more-positive attitudes towards older adults’ sexuality on both the explicit and implicit measures.
The Current Study
In this study, we sought to identify and compare young adults’ explicit and implicit attitudes towards the sexual lives of older adults. We assessed both implicit and explicit attitudes with the aim of developing a better understanding of attitudes towards the sexuality of older adults and possible discrepancies between individuals’ reported versus internalized views. Factors likely associated with these attitudes, namely (explicit) ageist and sexual attitudes, were also explored.
Method
Participants
The study participants were 134 students who were attending an Eastern Canadian university. Because of the heterosexual nature of the stimuli used in the study, nine participants indicating a sexual orientation other than heterosexual were removed from data analyses. We also excluded five participants because of missing data (greater than 10%) and 18 participants who had an error rate (i.e., categorized stimuli incorrectly) of 30 per cent or greater on the IAT. Thus, our final sample size consisted of 120 participants (50 men and 70 women).
All participants were aged 18 to 24 (M = 19.2 years). Participants were predominantly White (91.4%) and English-speaking (93.3%). Almost all (93.3%) were born in Canada. Over half of our participants indicated that they were in a committed relationship at the time of the study (8.3% married or cohabitating and 44.2% dating someone exclusively), 8.3% were dating someone but not exclusively, and 39.2% indicated that they were not dating anyone.
Measures
Background Questionnaire. Participants provided their background information including age, gender, ethnicity, relationship status, and sexual orientation.
Fraboni Scale of Ageism (Fraboni et al., Reference Fraboni, Saltstone and Hughes1990). The 29-item Fraboni Scale of Ageism assesses attitudes towards aging, in particular the affective components of prejudice and discrimination against older adults. Respondents indicated their agreement with each item on a 5-point Likert scale ranging from strongly disagree (1) to strongly agree (5). We reversed the scale scoring for ease of interpretation so that higher scores indicated more-permissive attitudes towards older adults (Explicit Aging Attitudes). Examples of items included “I do not like it when older people try to make conversation with me”, and “I personally would not want to spend time with an older person.” Research indicates that the scale has strong reliability and validity (Fraboni et al., Reference Fraboni, Saltstone and Hughes1990; Rupp et al., Reference Rupp, Vodanovich and Credé2005; α = .85 in the current study).
Brief Sexual Attitudes Scale – Permissiveness Subscale (Hendrick & Hendrick, Reference Hendrick and Hendrick1987; Hendrick, Hendrick, & Reich, Reference Hendrick, Hendrick and Reich2006). The Brief Sexual Attitude Scale is a multidimensional measure of sexual attitudes that assesses the following dimensions: conservatism/permissiveness, instrumentality, communion/relationality, and sexual practices. We used the 10-item Permissiveness Subscale (e.g., “It is okay for sex to be just good physical release”, “I do not need to be committed to a person to have sex with him/her”; Explicit Sexual Attitudes) as our measure of tolerance. The other subscales assessed constructs irrelevant to our hypotheses. Participants indicated their agreement with each item on a 5-point Likert scale ranging from strongly agree (1) to strongly disagree (5). We reversed scoring on this scale so that higher scores indicated more-permissive attitudes. The Permissiveness Subscale and the measure as a whole have strong psychometric properties (Hendrick et al., Reference Hendrick, Hendrick and Reich2006; α = .90 in the current study).
Attitudes towards Sexuality of Older Adults Scale. Because no appropriate explicit measure could be found, we adapted a measure of attitudes towards sexuality in disabled people to measure explicit attitudes towards sex among older adults (Scotti, Slack, Bowman, & Morris, Reference Scotti, Slack, Bowman and Morris1996). Respondents indicated the extent to which they believed it was acceptable for older adults to engage in each of 16 behaviours (e.g., masturbation, prolonged kissing with a partner in public, oral sex without a condom) on a 5-point Likert scale ranging from completely unacceptable (1) to completely acceptable (5). The original measure has strong psychometric properties (Scotti et al., Reference Scotti, Slack, Bowman and Morris1996), but because it was modified significantly, it is unclear to what extent those data can be applied to the current measure. Internal consistency for the modified scale was high (α = .86).
Implicit Association Test (Greenwald et al., Reference Greenwald, McGhee and Schwartz1998). For the current study, we created two groups of two IATs. The first group assessed biases in attitudes towards aging by contrasting photographic stimuli of older people with stimuli of younger people. In particular, the first (Aging Nonsexual IAT) compared stimuli of younger people to older people engaging in general activities (i.e., playing chess, grocery shopping, having dinner). The second (Aging Sexual IAT) contrasted stimuli of younger people to older people engaging in sexual activities with an individual of the other gender (i.e., mutual masturbation, oral sex, penile-vaginal intercourse). The second group of IATs was designed to measure attitudes towards sex by contrasting the same stimuli of general activities with the stimuli of sexual activities. We used both sexual and general activities to assess bias in sexual and non-sexual contexts as well as to control for discomfort that many individuals feel in responding to sexual stimuli. Specifically, the third IAT (Sexual Young Adult IAT) contrasted stimuli of younger people engaging in general activities to stimuli of young people engaging in sexual activities, whereas the fourth IAT (Sexual Older Adult IAT) contrasted stimuli of older people engaging in general activities to older people engaging in sexual activities. Four IAT tests were required to perfectly counterbalance both older and younger targets (to assess ageism in attitudes) with general versus sexual activities (to assess negativity towards sex) because each IAT captures biases in one dimension only.
Each IAT protocol was programmed using Inquisit 3 for Windows from Millisecond Software (www.millisecond.com). The IATs were presented on a personal computer and followed the standard paradigm (Greenwald et al., Reference Greenwald, McGhee and Schwartz1998). This experimental procedure incorporated five blocks of stimuli, with the third and the fifth blocks providing the critical data. The first block included 20 trials that introduced the participant to the categorical stimuli. The second block was identical, but required that the participant sort the attribute dimension (e.g., pleasant and unpleasant) rather than the categorical dimension (e.g., older people and younger people). The third block was the first test block, in which the categorical dimension was paired with the attribute dimension (e.g., sexual with pleasant and neutral with unpleasant). A total of 40 discrimination trials were used to collect the critical data. The fourth block, similar to blocks one and two, switched the location of the attribute labels, so that they appeared in reverse corners of the screen. The fifth block was the final test block and was identical to the third block, but required opposite associations between both dimensions (e.g., sexual with “unpleasant” and neutral with “pleasant”). In sum, participants completed three practice blocks with 20 discrimination trials each and two critical trial blocks with 40 discrimination trials each.
During the IAT blocks, stimuli were presented in the middle of the screen with a black background. We used neutral (e.g., couple playing chess, couple shopping together) and sexual (e.g., couple engaging in oral sex, couple engaging in penile-vaginal intercourse) categorical images selected from a research database. We used pleasant (e.g., “joy” and “beautiful”) and unpleasant (e.g., “nasty” and “horrible”) words for the attribute dimension also selected from a research database. The category names were presented in the upper corners of the screen, one in each corner. Participants were asked to associate the stimuli with one of the two categories according to a given rule. They pressed the “e” key to indicate the stimuli should be sorted to the left category and the “i” key to sort to the right category. The stimulus remained on the screen until the participant correctly associated it with a given category. According to the most recent sorting rule presented, if it was incorrectly sorted, the program flashed a red “x” on the screen until it was sorted correctly. Administration of all IAT protocols took approximately 40 minutes in total.
Computation of IAT Scores . The IAT generates a difference score (i.e., D score) that is calculated using a standard scoring procedure (Schnabel, Asendorpf, & Greenwald, Reference Schnabel, Asendorpf and Greenwald2008). Specifically, the algorithm computes a D score for the two test blocks by calculating the difference between the response latencies divided by the total standard deviation (Nosek, Greenwald, & Banaji, Reference Nosek, Greenwald and Banaji2005). The algorithm addresses two types of errors: errors from exceptionally fast response times (before stimuli are presented) and exceptionally slow response times (suggesting distraction or task interruption). In particular, latencies that were quicker than 300 ms were recoded to equal 300 ms, and latencies exceeding 3,000 ms were recoded to equal 3,000 ms.
D scores are reported on a continuum ranging from –2.0 to +2.0. Negative D scores in the Aging Nonsexual IAT reflect a stronger automatic bias favouring general activities among young adults as compared to general activities among older adults, whereas positive D scores reveal a stronger automatic bias favouring general activities among older adults as compared to general activities in younger adults. In the Aging Sexual IAT, negative scores reflect a bias in favour of sexual activities in young adults over sexual activities in older adults, whereas positive scores reflect a bias in favour of sexual activities in older adults over sexual activities in young adults. In the Sexual Young Adult IAT, negative scores indicate a bias favouring general activities in younger adults over sexual activities in younger adults; positive scores indicate that the bias is reversed. Finally, for the Sexual Older Adults IAT, negative scores reveal a bias favouring general activities among older adults over sexual activities in older adults, and positive scores mean that the bias is reversed. In each case, zero indicates no automatic bias. This scoring procedure produces higher internal consistencies, higher correlations with other implicit measures, higher correlations with explicit measures, and is more resistant to extraneous variables than the original scoring algorithm that did not incorporate error responses (Greenwald et al., Reference Greenwald, Nosek and Banaji2003).
Piloting of IAT Stimuli
For piloting purposes, 25 participants were asked to categorize the 40 IAT images into their proper categories (i.e., older sexual, younger sexual, older general, younger general) as fast as possible. The stimuli that were placed in the correct category the most efficiently were the ones chosen for actual data collection. Following procedures used successfully in another IAT image-based study (Thompson & O’Sullivan, Reference Thompson and O'Sullivan2012), efficiency was calculated using a response latency metric and the number of errors made when associating the stimuli with the target categories. Five out of ten pictures in each category were chosen for creation of the four IATs we used in this study.
Procedure
Participants were recruited from introductory psychology courses via an academic online scheduling system. Recruitment advertisements indicated that participants would be asked to complete a questionnaire package as well as an assortment of computerized tests evaluating their associations with regard to sexuality across the lifespan. Upon arriving for the study, participants were tested individually in a private research space. After providing informed consent, in line with standard definitions used in psychological research (Orimo et al., Reference Orimo, Ito, Suzuki, Araki, Hosoi and Sawabe2006), respondents were told that the term “older people” was used to refer to “anyone equal to or over the age of 65”. Each participant completed the four separate IATs, with each IAT taking approximately 10 minutes to complete, followed by the package of questionnaires. All participants received extra credit in their introductory psychology course and were debriefed about the purpose of the study. The study was approved by our institution’s ethics review board.
Results
Implicit and Explicit Attitudes towards Sex among Older Adults
Participants reported very positive explicit attitudes towards sex among older adults, with a mean score of 4.26 per item out of a possible 5 (SD = 8.95). In terms of implicit attitudes, as predicted, participants’ Aging Nonsexual IAT (M = –.46, SD = .37) and Aging Sexual IAT (M = –.48, SD = .39) scores both revealed an implicit bias favouring younger people over older people (H1) for both general and sexual activities. In addition, participants’ Sexual Young Adult IAT (M = –.36, SD = .48) and Sexual Older Adult IAT (M = –.54, SD = .44) scores revealed an implicit bias favouring general activities as compared to sexual activities among both younger and older adults (hypothesis H3). To determine the extent of association between the implicit and explicit measures, we examined the bivariate correlations between Explicit Aging Attitudes and our two implicit measures of attitudes towards sex among older people (Aging Sexual IAT and Sexual Older Adults IAT) (see Table 1). Contrary to what we expected, neither of the correlations was significant (r = –.01 and r = .12, p > .05 for the Aging Sexual IAT and Sexual Older Adults IAT respectively).
Table 1: Correlations among variables used in this study
a N = 120
b Gender: Men = 0, Women = 1
c Aging Nonsexual IAT = younger people vs. older people engaging in general activities; Aging Sexual IAT = younger people vs. older people engaging in sexual activities; Sexual Young Adult IAT = general activities vs. sexual activities among younger adults; Sexual Older Adult IAT general activities vs. sexual activities among older adults
d * p < .05; ** p < .01
Differences in Implicit Aging Attitudes
To determine the effects of gender and target age (hypothesis H2) on the first IAT group (Aging Nonsexual IAT and Aging Sexual IAT), we conducted a 2 (IAT) × 2 (gender) mixed-design ANOVA. Contrary to our predictions, participants’ scores on the Aging Nonsexual IAT (older versus young people engaging in general activities) and the Aging Sexual IAT (older versus younger people engaging in sexual activities) were not significantly different – F(1, 188) = .70, p = .40 – indicating that the bias in favour of younger over older people did not vary significantly as a function of the activity in which they were depicted engaging. In addition, men’s and women’s responses were also not significantly different from one another: F(1, 118) = 3.37, p = .07.
Differences in Implicit Sexual Attitudes
We used the same design to assess the effects of gender and target age (Sexual Older Adult IAT) on the second IAT group (Sexual Young Adult IAT and Sexual Older Adult IAT). The results indicated that the main effect for IAT was significant: F(1, 118) = 207.80, p < .001. Participants’ scores on the Sexual Young Adult IAT (general versus sexual activities in older adults) were significantly lower than were their scores on the Sexual Older Adults IAT (general versus sexual activities in younger adults). This finding suggests that participants’ bias in favour of general activities over sexual activities was significantly stronger when viewing photos of older people, in line with hypothesis H4. Men’s and women’s responses were not significantly different from one another – F(1, 118) = .90, p = .34 – which is contrary to hypothesis H5.
Predicting Implicit and Explicit Attitudes towards Sex among Older Adults
We used three separate hierarchical multiple regression analyses to test the extent to which gender, Explicit Aging Attitudes, and Explicit Sexual Attitudes were associated with both explicit and implicit attitudes towards sexuality among older adults (hypothesis H6). The first multiple regression analysis predicted explicit attitudes towards sexuality and aging (see Table 2). Examination of the zero-order correlations indicated that individuals with more-positive Explicit Aging Attitudes reported more-positive explicit attitudes towards sexuality and aging. Results of the multiple regression analysis indicated that the set of predictors was significant: R 2 = .08; F(3, 116) = 3.15, p = .03. Only Explicit Aging Attitudes uniquely accounted for a significant amount of the variance.
Table 2: Multiple regression analysis predicting explicit attitudes towards sexuality and aging
a N = 120
b F (3, 116) = 3.15, p = .03. Gender: Men = 0, Women = 1
c * = p < .05, ** = p < .01
The second multiple regression analysis predicted participants’ scores on the Aging Sexual IAT (younger versus older people engaging in sexual activities) (see Table 3). Examination of the zero-order correlations indicated that only gender was significantly correlated with the Aging Sexual IAT scores with women demonstrating somewhat less bias than did men with respect to sexual activities in young compared to older adults. However, this gender difference was small (Ms = –.57 and –.42 for men and women respectively). Of note, the predictors were not significantly associated with the Aging Sexual IAT scores: R 2 = .04; F(3, 116) = 1.70, p = .17.
Table 3: Multiple regression analysis predicting Aging Sexual IAT Scores
a N = 120
b F (3, 116) = 1.70, p = .17
c Aging Sexual IAT = younger versus older adults engaged in sexual activity
d Gender: Men = 0, Women = 1
e * = p < .05, ** = p < .01
The third multiple regression predicted scores on the Sexual Older Adult IAT (older adults engaged in general versus sexual activities) (see Table 4). Examination of the zero-order correlations indicated that Explicit Sexual Attitudes were significantly associated with the Sexual Older Adult IAT scores, and the regression itself revealed that the set of predictors was marginally related to Sexual Older Adult IAT scores: R 2 = .06; F(3, 116) = 2.41, p = .071. Only Explicit Sexual Attitudes added uniquely to the equation. Participants with restrictive sexual attitudes were more likely to reveal an implicit bias favouring general versus sexual activities among older adults.
Table 4: Multiple regression analysis predicting Sexual Older Adult IAT Scores
a N = 120
b F (3, 116) = 2.41, p = .07
c Sexual Older Adult IAT = Older adults engaged in general versus sex activities
d Gender: Men = 0, Women = 1e * = p < .05, ** = p < .01
Discussion
In this study, we examined and compared young adults’ explicit and implicit attitudes regarding the sexuality of older adults. Past research assessing such attitudes has relied entirely on self-reports, which are known to be influenced by demand characteristics (Colton & Covert, Reference Colton and Covert2007; Evans & Rodney, Reference Evans and Rodney2008). We believe that the study of attitudes regarding the sexuality of older adults may be especially prone to such demand characteristics given the highly sensitive nature of the topic as well as social pressure to hide or suppress ageist views about the elderly. This study incorporated a relatively complex design, but may be the first to begin to parse out the associations of both explicit and implicit measures to the study of attitudes towards the sexuality of older adults.
Consistent with other recent research (Bouman et al., Reference Bouman, Arcelus and Benbow2006; Lai & Hynie, Reference Lai and Hynie2011; Pratt & Schmall, Reference Pratt and Schmall1989), we found that, when asked, young adults report positive views about the sexuality of older adults. However, our research counters these findings by demonstrating that their implicit attitudes towards sexuality and aging, in fact, are negative, not positive. This is a notable discrepancy. By virtue of being beyond conscious awareness and control, the IAT results lead us to question the validity of measures of explicit attitudes of sexuality and aging, which reflect what an individual deliberately chooses to reveal. Our findings that participants’ implicit attitudes towards sexuality and aging were unrelated to self-reported explicit attitudes towards either sexuality or aging alone reinforces the disconnection noted between explicit and implicit measures. Future research employing self-report measures should, at the very least, incorporate standard measures of socially desirable responding to control for this potential bias.
Explicit Attitudes towards Sexuality and Aging
The young adults in our sample reported explicit attitudes towards the sexuality of older adults that corresponded to a rating between acceptable and completely acceptable across a wide range of sexual activities. Nonetheless, there was some variability in participants’ responses such that some participants were less accepting than were others. Our results provide some insights into the extent to which attitudes towards sexuality and attitudes towards aging influence attitudes towards sexual expression in older adults. We found that (explicit) attitudes regarding the sexual lives of older adults could be accounted for best by explicit attitudes towards aging; in fact, no meaningful connection was found with explicit sexual attitudes. This finding suggests that young people with more ageist attitudes, or at least those who are willing to endorse explicitly more negative attitudes towards aging, apply these attitudes to sexual expression in older adults. That is, there is some carry-over in attitudes towards aging. Neither the valence nor intensity of individuals’ attitudes towards sexuality was related to the extent to which they exhibited bias towards sexual expression in older adults. Thus, to change views about sexuality and aging, it may be necessary to change views about aging generally rather than about sexuality specifically.
Implicit Attitudes towards Sexuality, Aging, and Sexuality among Older Adults
The results of this study also shed light on young adults’ implicit attitudes towards sexuality, aging, and the combination of sexuality and aging (i.e., sexual expression by older adults). The design allowed us to examine the relative contribution of sexual and aging attitudes to those about the sexuality of older adults. Participants’ implicit attitudes indicated a moderate bias against older adults in favor of younger adults with regard to both their sexual lives and in more-general contexts. That is, when presented with younger adults and older adults engaged in general activities (the Aging Nonsexual IAT) as well as when presented with younger adults and older engaged in sexual activities (the Aging Sexual IAT), participants demonstrated a bias favouring younger adults. This finding reveals that young people hold ageist attitudes. This may be particularly true for men because we found that the men had a somewhat stronger bias against older people engaged in sexual activity than did the women. These results are consistent with past research investigating general ageist attitudes (Gast & Rothermund, Reference Gast and Rothermund2010; Jelenec & Steffens, Reference Jelenec and Steffens2002; Nosek et al., Reference Nosek, Banaji and Greenwald2002), and extends them to the sexual context, but are inconsistent with our predictions regarding gender differences.
The finding that young adults hold ageist implicit attitudes is particularly problematic because this form of prejudice has been linked to the daily communication of ageist stereotypes (Lagacé, Tanguay, Lavallée, Laplante, & Robichaud, Reference Lagacé, Tanguay, Lavallée, Laplante and Robichaud2012), emotional rejection of older people (Lookinland & Anson, Reference Lookinland and Anson1995), and neglect (Winterstein, Reference Winterstein2012). On the other hand, this finding may simply reflect greater comfort, ease, and perhaps familiarity with experiences captured by sorting tasks depicting young adults because the participants themselves are young adults. An essential next step in furthering our understanding of young people’s attitudes towards older adults is to examine how the bias noted in implicit attitudes is expressed behaviourally, if at all, in ways that reflect discrimination. This could be done experimentally either by presenting participants with hypothetical scenarios depicting opportunities to interact with younger compared to older adults or by conducting experiments in which participants would indicate behaviourally their preferences (e.g., how closely they sit near an older adult along a row of empty chairs).
Our findings also indicate that young people hold attitudes reflecting a notable bias against sexual activities in favor of more generic activities for both younger (the Sexual Young Adult IAT) and older adults (the Sexual Older Adult IAT). These results are consistent with past IAT research assessing attitudes towards sexuality that has used sexual words compared to neutral words and sexual photos compared to romantic photos (Geer & Robertson, Reference Geer and Robertson2005; Thompson & O’Sullivan, Reference Thompson and O'Sullivan2012). However, it extends this work by generalizing the findings across age groups. That is, our results replicate work indicating a bias against sexual activities depicted among young people and show that this bias is also present with respect to sexual activity in older adults. A bias favouring general over sexual activities is not particularly surprising; sexual issues, imagery, and information have long been regarded as socially sensitive topics, producing any of a range of emotions, such as anxiety, arousal, embarrassment, or curiosity. This connection is developed early in life. For instance, children learn at a young age that their bodies should be viewed as shameful or dirty (Kaestle & Allen, Reference Kaestle and Allen2011) and to keep sexual interest private (de Graaf & Rademakers, Reference de Graaf and Rademakers2011). Thus, it is likely that participants experienced greater degrees of discomfort, and hence cognitive interference, when pairing sexual activities than when pairing general activities.
Conversely, it was likely easier for our participants to sort stimuli containing no sexual content because the generic content of those stimuli introduced no cognitive interference. In keeping with this interpretation, we found that individuals with more negative explicit sexual attitudes demonstrated a stronger automatic bias favouring general compared to sexual activities in older adults. Surprisingly, men and women responded to the tasks comparing general and sexual activities similarly. In contrast, consistent with past research (Petersen & Hyde, Reference Petersen and Hyde2010), we found that the men reported more positive explicit sexual attitudes. This association suggests that sexual socialization and gender role expectations affect men’s and women’s reports of their sexual attitudes, but not their (implicit) responses beyond their conscious control. Indeed, we did not find gender differences in implicit sexual attitudes towards young adults (the Sexual Young Adult IAT).
We found a bias against sexual expression in older adults both when compared to sexual expression in younger adults (the Aging Sexual IAT) and when compared to older adults engaging in nonsexual activities (the Sexual Older Adult IAT). Combining stimuli representing both sexuality and aging produced greater bias than did stimuli for either sexuality or aging alone. Although the bias towards general as compared to sexual activities was present for both age groups, this finding was stronger for older adults (the Sexual Young Adults IAT) than for younger adults (the Sexual Older Adult IAT) among both the men and the women. The bias we found against both sex and aging interact to produce an even greater effect when presented with the combination of age and sex. The social psychological literature indicates that multiple, intersecting, stigmatized identities, such as older adult and sexual person, often compounds the impact of stigma and, in some cases, the resulting discrimination (Nabors, Reference Nabors, Nettles and Balter2012).
Limitations and Future Directions
Like all studies, ours contained a number of limitations that we could not rectify. First, we did not assess behaviour or behavioural intentions. Research is needed to determine whether, consistent with these findings, ageist (implicit) attitudes affect young people’s behaviour with respect to accepting and supporting sexual expression in older adults more so than behaviour in other contexts. Second, our sample consisted of young adults (18–24 years) attending university. Research on individuals currently caring for older adults would provide more-proximal insights into the extent to which prejudicial attitudes are reflected in discriminatory behaviour that has an impact on the sexual lives of those under their care. As individuals age, they may become increasingly more comfortable with the idea of sexuality in later life, as some researchers have suggested (Lai & Hynie, Reference Lai and Hynie2011). It is possible that the IAT responses reflect a degree of familiarity with cultural stereotypes rather than attitudes regarding the associated stimuli. However, the extreme automaticity and the evaluative components embedded in IAT tasks indicate that they capture attitudes, not familiarity or knowledge alone (Gawronski & Bodenhausen, Reference Gawronski and Bodenhausen2006; Nosek & Hansen, Reference Nosek and Hansen2008). In fact, previous studies have found that cultural knowledge has a minimal independent relationship with performance on the IAT (e.g., Nosek & Hansen, Reference Nosek and Hansen2008; Olson & Fazio, Reference Olson and Fazio2004). Finally, frequency and quality of contact with older people, aging knowledge, or experience and comfort with sexuality imagery or tasks are factors that we did not assess but which might help explain attitudes towards the sexuality of older adults.
Conclusion
Despite these limitations, we believe that our study makes three contributions to the field. First, we demonstrated that assessing individuals’ explicit attitudes have almost no predictive utility in understanding their implicit attitudes – that is, those evaluative reactions beyond conscious control or awareness. Second, using a mixed-methods design that incorporated both self-report measures and a series of IAT tests, we showed that young people do indeed reveal negative implicit attitudes (despite explicitly endorsing positive attitudes) regarding the sexual lives of older adults. Third, we revealed that it is young people’s discomfort around sexual matters, rather than biases against older people generally, that best explains the significant negative bias in their reactions towards the sexuality of older adults. Although this is an exploratory study, such information points to the importance of implementing programs that are designed to improve acceptance regarding sexuality and eliminate prejudicial attitudes among young people and those caring for older adults regardless of these individuals’ stated views on the subject. Given our finding that explicit attitudes towards sexuality but not towards aging were associated with attitudes towards sexual expression in older adults, these programs need also to target improving comfort around sexual matters more broadly. As the number of older adults grows in coming decades (Pratt & Schmall, Reference Pratt and Schmall1989), changing these attitudes may be increasingly important.
