| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
1 Department of Epidemiology and Public Health
2 Department of Psychology
3 Department of Internal Medicine, Yale University, New Haven, Connecticut.
Address correspondence to Becca Levy, Yale University, Department of Epidemiology & Public Health, 60 College Street, New Haven, CT 06520-8034. E-mail: becca.levy{at}yale.edu
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
AMONG persons aged 65 years and older, the third most common chronic condition is hearing loss, with prevalence rates ranging from 25% among those who are 70 to 74 years old to 50% by age 85 (Bogardus, Yueh, & Shekelle, 2003
; National Center for Health Statistics, 2001). Hearing loss in old age can lead to increased social isolation, self-denigration, loneliness, and depression (Gates & Rees, 1997; Herbst & Humphrey, 1980; Lichtenstein, Bess, & Logan, 1988). Although a wide array of factors has been identified that may contribute to hearing loss in old age, ranging from degeneration of cochlea cells to a decline in speed of information processing (for a review, see Willott, Chisholm, & Lister, 2001), the potential role of social-psychological factors contributing to objectively screened hearing in old age has been neglected.
Nevertheless, among cultures that attach less stigma to old age, several studies have documented that older members of these cultures tend to experience less hearing loss with old age. For example, older Easter Island inhabitants, members of the Polynesian culture that holds predominantly positive stereotypes of aging (Pearson, 1992), demonstrated a hearing advantage over those who had left the island to live in industrialized societies (Goycoolea et al., 1986). Although the authors assumed the hearing loss was due to the greater noise associated with industrialization, a social-psychological explanation cannot be precluded by their evidence. That is, industrialized societies have been found not only to produce more noise but to foster negative age stereotypes (Butler, 2004); therefore, the patterns of hearing performance among the older present and expatriate Easter Islanders may reflect the cultures in which they were immersed.
The fact that sensory perception is susceptible to social influences has been established experimentally. For example, Asch (1955) demonstrated that judgments about the length of lines can be manipulated by the experimenter, so that social conformity takes precedence over the evidence provided by participants' senses. In a follow-up study to Asch (1955), scientists using functional-magnetic-resonance imaging found biological evidence for the involvement of perceptual processes during social conformity (Berns et al., 2005). Recent studies have also found that visual-perception patterns can be influenced by the activation of racial stereotypes (Eberhardt, Goff, Purdie, & Davies, 2004), and activating gender stereotypes can alter speech perception (Strand, 1999).
In our study, we build on these experimental studies by examining whether age stereotypes can contribute to auditory perception. In addition, we explore for the first time whether the process by which stereotypes predict sensory perception can occur in the community and over time.
The assimilation of age stereotypes begins in childhood (e.g., Kwong See & Rasmussen, 2005), and it continues throughout adulthood—before and after these stereotypes become self-relevant (Levy, 2003). The internalized age stereotypes that are brought into old age may be positive as well as negative, but it is the latter that predominate (e.g., Palmore, 1999).
Debilitation is a major theme of age stereotypes (Butler, 2004; Levy, 2003). When older individuals experience a loss of hearing, the activation of stereotypes could be compounded. It may be that, at the same time as a negative age stereotype is triggered, a stereotype that is specific to impaired hearing is also triggered. The latter stereotype includes the traditional association of deafness with mental disability (Becker, 1981). This stereotype may, in turn, lead to what has been called "self-stigmatization," which occurs when individuals carrying negative images of hearing loss find that their own hearing is impaired, and thereupon target themselves with these images (Hetu, 1996). Because hearing loss is also associated with old age, the two types of stereotypes may provide mutual reinforcement.
There are three pathways by which age stereotypes might influence hearing in old age. The social pathway would operate through age stereotypes generating expectations about physical deterioration, such as hearing loss, that become self-fulfilling prophecies (Jussim, Palumbo, Chatman, Madon, & Smith, 2000; Levy, Slade, & Kasl, 2002). Because these expectations may be held by the targeters as well as the targets, they contribute to a feedback loop in which impaired hearing acts as an age cue generating "an impression upon others of confusion and disorientation that may reinforce negative expectations" (Ryan, 1996, p. 96).
Cognition is an additional pathway by which age stereotypes may contribute to reduced hearing. A series of studies has found that the activation of negative age stereotypes held by older individuals can adversely affect their memory performance (Hess, Auman, Colcombe, & Rahhal, 2003; Hess, Hinson, & Statham, 2004; Levy, 1996). Thus, older individuals' ability to draw on the higher level cognitive processes, which could be used to compensate for declines in their ability to process acoustic signals (Pichora-Fuller, 2003; Wingfield, Lindfield, & Goodglass, 2000), may be compromised by negative stereotypes.
The third pathway of potential age stereotype operation is physiological. When the negative-age-stereotype activation occurs, it is likely to be stressful, as suggested by a laboratory study in which older individuals who had been subliminally exposed to these stereotypes showed a significantly heightened cardiovascular response to stress (Levy, Hausdorff, Hencke, & Wei, 2000). At the next stage of this pathway, stress could elevate cytokines (Miller, Cohen, & Ritchey, 2002). Finally, cytokines could negatively affect hearing (Adams, 2002).
In this study, we consider two dimensions of age stereotypes as predictors of hearing loss: externality (i.e., stereotypes related to physical appearance) and negativity. Although negativity as a stereotype dimension has been considered in several studies (e.g., Hess et al., 2003; Hummert, Garstka, Shaner, & Strahm, 1994; Levy, 1996), the dimension of externality has received little attention. However, we found in prior research that external age stereotypes, as well as negative age stereotypes, were associated with worse memory performance by older participants (Levy & Langer, 1994). This finding about externality is consistent with the nature of stereotypes in general, because they frequently operate by producing mental images of the targets (Allport, 1954). It has been demonstrated that mental images can generate the same emotional arousal and causal sequences as the real objects (Kosslyn, 1995). In addition, visually presented stereotypes, in the forms of photographs, have been shown to have robust effects (Poehlman, Uhlmann, Greenwald, & Banaji, 2005).
Although the age-stereotype dimensions of negativity and externality were previously considered as two dimensions of a single concept (Levy & Langer, 1994), in the following study we considered whether they may operate as independent dimensions.
We therefore formulated the following hypotheses. First, older persons with more negative age stereotypes will perform worse on screened hearing at 36 months, after adjusting for baseline-screened hearing, age, and other relevant covariates. Second, older persons with more external age stereotypes will perform worse on screened hearing at 36 months, after adjusting for baseline-screened hearing, age, and other relevant covariates.
|
METHODS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
). We identified potential participants from 3,157 age-eligible members of a health plan in New Haven, Connecticut. The primary inclusion criteria for the Precipitating Events Project were English speaking and requiring no personal assistance with four key activities of daily living (bathing, dressing, transferring from a chair, and walking). The participation rate was 75.2% (Gill et al., 2001). To be included in the analytic sample, participants had to have their age stereotypes measured at baseline, as well as their hearing screened at baseline and 36 months, by nurse researchers who were unaware of the study's hypotheses. Eighty-eight of the original participants died before the 36-month assessment. An additional 120 participants were missing the age-stereotype measure or at least one of the hearing assessments. The 546 participants who were included in the analytic sample did not significantly differ from those who were excluded in terms of age, gender, number of chronic conditions, race, or smoking history, but they did have significantly more years of formal education (12.1 vs 11.6 years; p =.04). The baseline characteristics of the included participants are described in Table 1.
|
Measures
Age stereotypes
At baseline, participants were asked, "When you think of an old person, what are the first five words or phrases that come to mind?" Although this measure had been used previously to assess views of aging (e.g., Levy & Langer, 1994), to confirm that this measure is assessing beliefs that are stereotypical, we had two judges rate responses on a nine-item scale of stereotypicality, ranging from 1 (not at all stereotypical) to 9 (very stereotypical; see Macrae, Bodenhausen, Milne, & Jetten, 1994; Galinsky & Moscowitz, 2000). Because the interrater reliability was high (r =.81), we averaged the ratings. We found that the words were judged to be stereotypical, based on the mean rating obtained for each participant (M = 7.6).
Next, we scored all the age-stereotype words on the two dimensions that acted as independent variables in our study: negativity and externality. Scores on the negativity dimension ranged from 1 (not at all negative, e.g., compassionate) to 5 (very negative, e.g., feeble). The externality dimension ranged from 1 (not at all external, e.g., grouchy) to 5 (very external, e.g., white hair). For both dimensions, we scored neutral words or phrases as 3 (e.g., normal). We rated the participants' five responses independently on each of the two dimensions. Overall scores for each dimension ranged from 5 to 25, for the negativity dimension M = 9.9, SD = 4.9, and for the externality dimension M = 12.3, SD = 3.8. The two raters, who were blind to the participants' characteristics, received an effective reliability of 94%, using the Spearman–Brown formula (Rosenthal & Rosow, 1991).
Many of the age stereotypes that related to appearance were negative (e.g., stooped over), so it is not surprising that our two age-stereotype dimensions were moderately correlated (r =.39, p <.001). However, because some of the appearance-related age stereotypes were positive (e.g., always dresses well) and some of the negative age stereotypes were not related to appearance (e.g., senile), we examined whether the two age-stereotype dimensions independently predict screened hearing at 36 months.
Screened hearing
We screened hearing at baseline and 36 months with a hand-held audioscope, using a protocol that has been found to be valid and reliable with older persons (Lichtenstein et al., 1988; McBride, Mulrow, Aguilar, & Tuley, 1994). An examiner held the audioscope in the participants' ear, with the probe tip sealing the canal. The examiner asked participants who wore hearing aids to remove them before testing. The right ear was tested first, followed by the left ear, with a sequence of three 40-dB pure-tone frequencies, or pitches (500 Hz, 1000 Hz, and 2000 Hz). Participants raised a hand when they heard a tone. At baseline and 36 months, participants received a score of 1 for each tone heard in the right or left ear, with a total possible score ranging from 0 to 6. At baseline, the participants had an average screened-hearing score of 3.53, SD = 2.05, and at 36 months the average score was 2.96, SD = 2.14.
A review of hearing-loss studies found that the audioscope "demonstrated excellent accuracy in determining hearing loss" (Yueh, Shapiro, MacLean & Shekelle, 2003, p. 1979). Three studies have tested the audioscope by comparing it with a more sensitive measure: the formal audiogram, which examines hearing thresholds (Ciurlia-Guy, Cashman, & Lewsen, 1993; Lichtenstein et al., 1988; McBride et al., 1994). Each of these studies used the audioscope at the 40-dB threshold, as was used in our study, with a screening frequency of 2000 Hz or a combination of 1000 and 2000 Hz. The three studies found the audioscope to have strong sensitivity (
.94) and good specificity (.69–.80).
We used the two frequencies validated in previous studies (Ciurlia-Guy et al., 1993; Lichtenstein et al., 1988; McBride et al., 1994). To make these frequencies compatible with the pitches that are common to hearing and understanding everyday speech, we added 500 Hz (Gates & Rees, 1997). A speech-reception-threshold test uses the frequencies of 500, 1000, and 2000 Hz (Mhoon, 1997). These frequencies are also used in the hearing-loss computation suggested by the American Medical Association (Rees, Duckert, & Carey, 1999).
Covariates
We measured all covariates at baseline. Among them were several that are known to affect hearing: demographic variables of age, education, gender, and race (Willott et al., 2001); health-related variables of depressive symptoms (Herbst & Humphrey, 1980), measured by the Center for Epidemiologic Studies–Depression Scale (Radloff, 1977); number of physician-diagnosed chronic conditions (Gill et al., 2001); and smoking history (Fried, 1998), which was scored as present or absent.
We also included a covariate of baseline-screened hearing in order to minimize the possibility that a finding related to age stereotypes predicting screened hearing was simply a reflection of baseline differences in screened hearing. This measure consisted of the 40-dB tones (500 Hz, 1000 Hz, and 2000 Hz) heard in either the right or left ear, which we scored in the same way as our outcome measure of screened hearing at 36 months.
To compensate for the PEP strategy of oversampling persons at increased risk for functional decline, as measured by an older age, lower scores on the Mini-Mental State Exam (MMSE) (Folstein, Folstein, & McHugh, 1975), and slower gait speed (for a full description of assembling the cohort, see Gill et al., 2001), we considered as covariates MMSE score and gait speed, in addition to age. We measured gait speed as the time needed to walk back and forth over a 10-ft course.
Statistical Analysis
We constructed a regression model to determine whether the two age-stereotype variables predicted screened hearing at 36 months, after all the covariates were first entered. We used the two age-stereotype dimensions as continuous variables. We selected covariates that significantly correlated with both the predictors and the outcome variable of screened hearing at 36 months. We conducted a partial F-test to examine whether adding the predictor variables to the regression model predicted significantly more of the variance in the outcome than the covariates alone. In addition, we assessed the significance of all two-way interactions by converting predictor and covariate variables to their z scores for centering purposes and running regression models with the interactions included (Aiken & West, 1991).
To examine the influence of the age stereotypes on screened hearing at 36 months, we calculated predicted screened-hearing scores at 1 SD above and 1 SD below the age-stereotype means and then subtracted the lower score from the higher score, according to the method suggested by Aiken and West (1991). To compare this screened-hearing-difference score to the decline expected to occur as a result of age, we divided this difference score by the final regression model's standardized parameter estimate for age.
To enhance our confidence that the influence of age stereotypes on screened hearing at 36 months was not due to an impairment in screened hearing that existed prior to baseline, we examined whether age stereotypes predicted screened hearing at 36 months in a subset of 75 participants who were able to hear all the tones at baseline.
We performed all analyses with SAS, Version 8.2 (SAS Institute Inc., Cary, NC). We considered a value of p <.05 to be significant.
|
RESULTS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
; see Table 2). In addition, we checked for colinearity among the variables in the model and found that the Pearson correlation coefficients between the covariates did not exceed.21 and the correlations between the covariates and predictors did not exceed.12.
|
We found that the predictor variables continued to significantly predict the outcome of hearing at 36 months, when all possible covariates were entered into the model, including age, baseline-screened hearing, gender, race, education, MMSE, gait speed, smoking history, number of chronic conditions, and depressive symptoms (see Table 2).
We also found that none of the variables, including the two age-stereotype predictors, negativity and externality, significantly interacted in predicting screened hearing in the regression model.
To determine whether the results remained after restricting our analyses to participants who were able to hear all of the tones at baseline, we repeated the model for this subset of 75 individuals. We found that participants in the subset who had more negative age stereotypes showed significantly greater screened hearing loss over the next 36 months, after taking into account all of the covariates, ß = –.11, t = –2.20, p =.03. However, we found that the external age stereotype was no longer a significant predictor of hearing loss in the subset, after taking into account all of the covariates.
When we estimated the influence of the two age-stereotype variables on screened hearing at 36 months (according to the method suggested by Aiken and West, 1991), we found that those with more negative and external age stereotypes scored 0.7 of a point lower in screened hearing than those with more positive and internal age stereotypes. According to our final-regression model, this difference is equivalent to the decline in screened hearing expected to occur as a result of aging during an 8-year period.
|
DISCUSSION |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
), rather than on whether psychosocial factors contribute to hearing in old age. According to the effect sizes generated by the full regression model, the two age-stereotype variables, negativity and externality, had a stronger impact on screened hearing at 36 months in our sample than did other established predictors of hearing loss: depressive symptoms, gender, race, and smoking history (Fried, 1998; Herbst & Humphrey; Willott et al. 2001). The lack of significant interaction between the two stereotype variables indicates that they operate independently in predicting screened hearing. The estimated influence of negative and external age stereotypes on screened hearing (equivalent to the decline in screened hearing due to aging expected over 8 years, as estimated by the regression model) is made more impressive by the outcome measure's inability to pick up the full range of possible change. That is, our screened-hearing measure was able to detect only those changes in hearing thresholds that crossed the 40-dB level used for the screening criterion.
It is unlikely that the influence of negative age stereotypes on screened hearing at 36 months is due to impairment in screened hearing prior to baseline, because the result was similar when we restricted our analysis to the 75 people who were able to hear all of the tones at baseline. Also, we found in the full sample that negative age stereotypes significantly predicted screened hearing at 36 months, after we took into account baseline-screened hearing.
Because functional hearing loss and physiological hearing loss are not always highly correlated (Erdman & Demorest, 1998), in future research it would be helpful to conduct a longitudinal study with multiple assessments of age stereotypes, functional hearing measured by a hearing-handicap scale (e.g., Newman, Weinstein, Jacobson, & Hug, 1991), and physiological hearing measured by a formal audiogram to quantify the thresholds of audibility. There would, then, be a foundation for determining both the contribution of age-stereotype-induced hearing loss to the reinforcement of age stereotypes, and the extent to which the reinforcement may in turn contribute to subsequent hearing loss.
The salient combination of negative and external age stereotypes is consistent with advertising that focuses on reducing outward indications of aging (e.g., wrinkles and gray hair), which in itself may contribute to the pejorative status of these markers. The main target of such campaigns is women—consistent with society's traditionally valuing their appearance more than that of men (Calasanti & Slevin, 2001). In line with this background, we found that the female participants reported holding significantly more external age stereotypes than did the male participants (see Table 1). We also found, however, that gender significantly interacted neither with externality nor negativity in predicting screened hearing. This suggests that the association between age stereotypes and screened hearing is similar among both older men and older women.
In addition, the lack of a significant interaction between either of the two age-stereotype variables and age or baseline-screened hearing suggests that the age-stereotype predictors operate similarly among the different ages of the participants in the sample and at different ranges of the baseline-screened-hearing spectrum.
Both negative and external age stereotypes could have adverse health-behavior consequences. For instance, findings that older individuals, in contrast to younger individuals, are more accepting of hearing loss (Fozard & Gordon-Salant, 2001), and that elders with hearing loss often do not seek professional care (Yueh et al., 2003), might be explained in part by age stereotypes. That is, it awaits further research to determine whether older individuals who hold more negative and more external age stereotypes might assume hearing loss is inevitable and therefore not worthy of medical attention. The lack of support from family and friends, as well as health care professionals, for hearing loss help-seeking (e.g. Mahoney, Stephens, & Cadge, 1996) may also be partially due to their sharing, or even promoting, the sense of inevitability.
In the absence of a routine medical intervention for hearing loss (Bogardus et al., 2003), the possibility of a psychosocial explanation takes on additional importance. Our findings suggest that efforts to understand patterns of hearing in older individuals could benefit from taking age stereotypes into account.
|
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received for publication June 17, 2004. Accepted for publication August 12, 2005.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  A. Terracciano and R. R. McCrae Perceptions of Americans and the Iraq Invasion: Implications for Understanding National Character Stereotypes. Journal of Cross-Cultural Psychology, January 1, 2007; 38(6): 695 - 710. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
