Exploring the Impact of Age,Race,and Stereotypes on Perceptions of Language Performance and Patronizing Speech

Abstract

Two experiments tested whether age and racial stereotypes influence communication. Specifically, both studies sought to understand if older African American targets would experience a communicative double jeopardy. In the first experiment, participants assessed targets’ language performance and beliefs about their own speech style (i.e., patronizing speech style). Age (participant and target) interacted with stereotype to influence ratings of language competence, and an interaction of target race, stereotype, and participant age influenced the elicitation of patronizing speech. In the second experiment, participants assessed communication competence and patronizing speech. Age groups of the targets and the participants, rather than racial groups, significantly influenced perceptions of both ratings of communication competence and the adoption of a patronizing speech style. Implications for the Age Stereotype in Interaction Model of intergenerational communication and future research on intersectionality are discussed.

Keywords

intergenerational communication age stereotypes patronizing speech language in adulthood communication competence

Many scholars assert that the style and effectiveness of an intergenerational communicative exchange is influenced by a number of different variables that are largely a function of stereotyping of older individuals (Harwood & Giles, 1996; Hummert & Shaner, 1994; Ryan, Hummert, & Boich, 1995). This research has limited its scope of investigation to European Americans, thus neglecting an important cultural component in the study of patronizing speech: the influence of race. Interracial communication, like intergenerational communication, may also be challenging because of the influence of stereotypes (Fujioka, 1999; Kochman, 1994; Kochman 1981; Mallinson & Brewster, 2005; Teal & Street, 2009).

Few studies have discerned the influence of race on the quality of communication toward older adults. For the most part, race and age continue to be addressed separately in communication literature. Fortunately, there is some research to provide a framework in which to begin an examination of the influence of race on communication in later life. Hummert (1994) proposed the Age Stereotypes in Interaction Model to detail features of the interactants and contexts, their link to stereotypes, and influence on communication.

This model does not address race. Adding a level of intergroup communication to the intergenerational interaction has the potential to lead to a compounding of negative stereotypes, creating a double jeopardy situation for the older minority individual. This study examines the potential impact of stereotypes associated both with race and with age, and as such, seeks to expand the Age Stereotypes in Interaction Model (ASI; Hummert, 1994; Hummert & Shaner, 1994).

Review of the Literature

Much of the intergenerational communication research is based on the Communicative Predicament of Aging Model (CPA model; Harwood, Giles, Fox, Ryan, & Williams, 1993; Ryan, Giles, Bartolucci, & Henwood, 1986), which explains the process of using age as a group identifier. The model, derived from communication accommodation theory (see Giles, 2016, for a recent review), proposes that age-related cues trigger stereotyped expectations about interactions with older people. For example, wrinkles and gray hair activate assumptions about an older person’s language abilities during an interaction. These assumptions lead the younger person to modify his or her speech in order to accommodate the presumed difficulties of the older person. The older person is then constrained by the modifications and assumes that their competence is diminishing. The older person may lose self-esteem and adopt an old identity. The younger person’s expectation is reinforced by these behaviors, and both communicators become reluctant to engage in future interactions (Nussbaum, Pecchioni, Robinson, & Thompson, 2000; Ryan et al., 1986).

Patronizing speech is related to the perception of helplessness or low functionality of the message target. That is, negative stereotypes lead a perceiver to believe that normal adult speech will be ineffective with some individuals and that extensive accommodations are necessary for understanding (Culbertson & Caporael, 1983; Giles, 2016; Ryan et al., 1986). According to Ryan and her colleagues, patronizing speech is distinguished from normal adult speech by the presence of simplification strategies. These strategies include slow, loud speech; short, simple sentences; exaggerated intonation, a demeaning emotional tone, and a low quality of talk. This speech style is seen as patronizing and unsatisfying by most older adults (Giles, Fox, & Smith, 1993; Ryan, Bourhis, & Knops, 1991). Not only is patronizing speech inappropriate, it is disconfirming and a potential detriment to the receiver, as it assumes the receiver is incompetent and powerless (Nussbaum et al., 2000). The older adult may lose self-esteem, start to doubt his/her competence, and behave in ways that reinforce the expectation of powerlessness (Giles, Fox, Harwood, & Williams, 1994).

According to Hummert, Garstka, Ryan, and Bonnesen (2004), age stereotypes play a significant role in the decision process of whether or not to modify one’s speech. Hummert, Garstka, Shaner, and Strahm (1994) identified seven stereotypes of older adults, three positive and four negative. The negative stereotypes are generally associated with decrement; consequently, studies have shown that the stereotypes of older adults encourage the modification of one’s speech style even when the stereotype does not indicate that a communicative deficit actually exists (Hummert & Shaner, 1994; Hummert, Shaner, Garstka, & Henry, 1998). The CPA model demonstrates that when intergenerational communicators come together, negative expectations result in patronizing speech (Harwood et al., 1993; Ryan et al., 1986). Since positive and negative stereotypes of older adults exist, however, Hummert (1994) developed a more comprehensive model for inter- and intragenerational encounters. The ASI model details features of the interactants, the context of the encounter, how those features may trigger stereotypes (positive or negative) and impact communication (modified or not; Hummert, 2007; Rittenour & Cohen, 2016).

Several studies support the tenets of the ASI model by finding that not all older adults are communicated with in a patronizing way (Hummert & Shaner, 1994; Hummert et al., 1998). Hummert and Shaner (1994) provide evidence that negatively stereotyped targets, but not positively stereotyped targets, will receive patronizing speech. Hummert et al. (1998) revealed that the degree and type of patronizing speech (overly directive or nurturing) is influenced by context and participant age. Results found a significant increase in the amount of directive patronizing speech to the more positive target while in the hospital context. Anderson, Harwood, and Hummert (2005) found age salience and type of relationships (grandparent vs. acquaintance), among other variables, to influence the stereotyping process. Rittenour and Cohen (2016) studied young adult reactions to age progressions finding support for the ASI model because “the physical characteristics of age activated people’s negative stereotypes of older adults more broadly” (p. 283).

Adding race into the ASI model provides a depiction of the double jeopardy situation an older minority individual may experience. “Older blacks perceive more discrimination than do older whites” (Barnes et al., 2004, p. 315). Race may be a salient factor influencing perceptions of incompetence and encouraging overaccommodative speech.

Five decades after Martin Luther King’s historic “I Have a Dream” speech in Washington, D.C., a new survey by the Pew Research Center finds that fewer than half (45%) of all Americans say the country has made substantial progress toward racial equality and about the same share (49%) say that “a lot more” remains to be done. (Pew Research Center, 2013).

Although the majority of respondents agreed that Blacks and Whites get along well (73% of Black respondents; 81% of White respondents), “still, about a third of all blacks (35%) say they had been discriminated against or treated unfairly because of their race in the past year” (Pew Research Center, 2013).

According to Fiske (2000), discrimination is a behavioral bias whereas stereotyping is a cognitive bias. As such, it is important to recognize that negative stereotypes of African Americans continue to exist in media portrayals (e.g., Fujioka, 1999; Millard & Grant, 2006) and in interpersonal contexts (e.g., Mallinson & Brewster, 2005). The negativity of racial stereotypes includes perceptions of African Americans as lazy, uneducated, and aggressive. Understanding how racial stereotypes interact with age stereotypes can provide valuable information for the study of intergroup communication.

The considerable literature on interracial communication between African Americans and European Americans focuses on the styles within the two groups rather than factors that impact interactions between them (Hecht, Collier, & Ribeau, 1993; Orbe, 1995). Investigators agree that African Americans tend to be louder and use more intense language than European Americans (Donahue, 1985; Hecht et al., 1993; Kochman, 1994). In the patient–physician communication literature, Street, O’Malley, Cooper, and Haidet (2008) examined the role of concordance by race, and while they found that “racial concordance does appear to orient patients toward some common ground with the physician . . . other factors, however, may be more influential determinants of perceived personal similarity” including age (p. 203).

Martin, Hecht, and Larkey (1994) found an absence of strategies for improving conversations with African Americans among European Americans, and when presented with an unsatisfying conversation with Whites, they noticed that Blacks will withdraw or avoid conversation altogether. Much like the negative spiraling effects of CPA, investigators explain this response as acquiescence to the power that is implied by the perception of European Americans. Nussbaum et al. (2000) and Ryan, Kennaley, Pratt, and Shumovich (2000) note comparable reactions among elderly adults, as they will typically disengage from communicative interactions when such exchanges become difficult or dissatisfying.

Additionally, older people are often assumed to have more receptive and expressive communication deficits. Ryan, Kwong See, Meneer, and Trovato (1992) developed the Language in Adulthood Questionnaire, a scale that identifies changes in perceptions of language among healthy adults of different ages. Their study indicated that younger adults perceive their language performance more positively than older adults on most items. Hummert, Garstka, and Shaner (1995) found similar results based on age but also found that negatively stereotyped targets were believed to be less competent in language performance than positively stereotyped targets. To date, researchers have ignored the possibility that race might interact with age to influence beliefs about communication.

The ASI model (Hummert, 1994, 2007; Hummert et al., 2004; Rittenour & Cohen, 2016) provides a framework to explain the relationship between age of interactants, stereotypes of older adults, communication beliefs, and patronizing speech. Since past research suggests that there are negative assessments based on race, age and race may combine to form an even more complicated communicative predicament. Therefore, this study tested the following hypothesis: Beliefs about language competence and patronizing speech will be influenced by an interaction of the target’s age, the target’s race, and the target’s stereotype, such that negatively stereotyped older African Americans will be rated as the least competent and most likely to receive patronizing speech.

Since the interracial communication research (Donahue, 1985; Hecht et al., 1993; Kochman, 1994) suggests that conversational styles may differ between African Americans and European Americans, we ask the following research question: Will participants’ race influence ratings of their own speech style?

Study 1

This hypothesis and research question were examined in a 2 × 2 factorial design, with two independent within-subjects factors, target age (young or old) and stereotype consistency (young positive/old negative or young negative/old positive). Additional independent variables included participant age (young or old), participant race (African American or European American, hereafter AA and EA, respectively), and target race (AA or EA). Dependent variables were the participants’ beliefs about the targets’ language competence and the speech style participants reported they would use with the targets.

Participants

Participants included 82 young adults (18-40 years old; M = 25.62 years, SD = 7.66), 50% EA and 50% AA, and 82.9% female. In addition, there were 86 older adults (55-89 years old; M = 72.09 years, SD = 8.57), 53.5% EA and 46.5% AA, and 72.1% female. Considering years of education, young adults (M = 14.95 years, SD = 2.59) were more educated than the older adults (M = 13.01 years, SD = 2.46). This difference was statistically significant, F(1, 161) = 24.99, p < .0001. There was no significant difference in education between EAs and AAs and no significant interaction between age and race. All participants were volunteers from communication courses, community organizations, and churches. Although course credit was offered to students, participation in the study was strictly voluntary for everyone. An alternative extra credit assignment was also available.

Stimulus Materials

Photographs

Eight simple headshots of young and old AA and EA women displaying neutral expressions in a neutral background were selected for a pilot test. This neutrality was important, as it controlled for confounding factors such as smiles and settings (e.g., nursing homes, beaches, and otherwise suggestive backgrounds) found in previous research to influence the association with particular stereotypes (Hummert, Garstka, & Shaner, 1997; Hummert et al., 1998). The pilot test verified that participants could identify the age and race of the women and that four of the women were not associated with either positive or negative stereotypes of older adults. Based on this pilot, four photographs were selected for the experiment: 1 young EA, 1 young AA, 1 old EA, and 1 old AA.

Stereotype Consistency

The positive and negative stereotype selected followed a model from past research. Hummert et al. (1998) operationalized stereotypes associated with aging by using six Golden Ager traits (positive) and six Despondent traits (negative). The traits used to define the Despondent stereotype were sad, depressed, afraid, hopeless, neglected, and lonely. The six traits used to describe the Golden Ager were active, sociable, well-informed, productive, independent, and future oriented. Stereotype descriptions were counterbalanced, thus having older targets identified with positive traits and younger targets identified with negative traits (stereotype-inconsistent), as well as having older targets identified with negative traits and younger targets identified with positive traits (stereotype-consistent). Stereotype consistency focused on age stereotypes only.

Questionnaires

Demographics

Participants were asked to indicate their race, age, educational level, gender, and if English was their first language. Data from participants who indicated English was not their first language were not analyzed.

Language in Adulthood Questionnaire

Ryan et al. (1992) created the Language in Adulthood Questionnaire (LIA) to measure perceived language abilities. The original questionnaire consists of 10 items that assess a participant’s perception of each target’s receptive language skills (e.g., losing track of the topic in conversation), nine items that assess expressive skills (e.g., using fewer difficult words in conversation), and one item reflecting overall conversation skill. Using a 7-point Likert-type scale, respondents were asked to circle a number from 1 (strongly agree) to 7 (strongly disagree) expressing agreement or disagreement with the statements.¹

Six statements assessing stereotypical beliefs associated with AAs were randomly listed among statements on beliefs about language skills from the LIA. The AA stereotypes included descriptions of being dishonest, poor, loud, lazy, uneducated, and aggressive (Fujioka, 1999). The Likert-type scale enabled participants to identify these terms as not at all likely, but the negativity of the terms fit the experimental design (i.e., measuring negative perceptions to determine their link to perceptual outcomes).

Vocal Characteristics Beliefs Scale

Participants’ beliefs about their speaking style with the stimulus person were assessed by rating eleven vocal characteristics: fast, hesitant, understandable, expressive, loud, wavering, thin, shrill, high-pitched, exaggerated intonation, and exaggerated pronunciation (Hummert & Shaner, 1994). Seven characteristics are associated with older voices (fast, wavering, hesitant, thin, loud, understandable, and expressive), and four are associated with patronizing speech style (shrill, high pitch, exaggerated intonation, and exaggerated pronunciation). Participants rated their speech toward each target on a Likert-type scale from 1 (not at all) to 7 (extremely).

Procedure

Data were collected with groups of participants from sets of four different packets. Each packet began with the demographic information. Then participants were presented with one photograph followed by the LIA and belief scale and then a second photograph followed by the same scales. Every participant responded to one older adult and one young adult, and every participant responded to the positive stereotype and to the negative stereotype. Age and stereotype combinations were balanced so that respondents saw young-positive/old-negative or old-positive/young-negative. Presentation order was counterbalanced. In addition, participants received a packet with either EA photos or AA photos; that is, race of target was a between subjects factor.

Results

Reliability of the LIA

Ryan et al. (1992) and Hummert et al. (1995) reported problems with achieving adequate reliability on both dimensions of the LIA. Similar problems were found in our analysis. Although the receptive dimension achieved adequate reliability (α = .84), the expressive dimension did not (α = .64). Based on Hummert et al.’s past work, we conducted a confirmatory factor analysis of the LIA, resulting in two factors identical to those reported by Hummert et al., χ²(171) = 2962.99, p < .0001. The first factor, memory, included seven items and was reliable (α = .92); the initial factor loadings for these items ranged from .51 to .84 and accounted for 40% of the variance. The second factor was hearing which included 6 items (α = .86); the initial factor loadings for these items ranged from .36 to .64 and accounted for 11% of the variance. Seven additional items were analyzed in their own separate analyses, including one question to measure conversation more generally.

Testing of the Hypothesis

Three repeated measures multivariate analyses of variance (MANOVAs) examined the impact of all independent variables (target age, target race, target stereotype consistency, participant race, and participant age group) on language competence and patronizing speech.² The first MANOVA examined the impact of these factors on the memory and hearing dimensions of the LIA while the second MANOVA tested the impact of these factors on seven individual items of the LIA. The third MANOVA tested the impact of these factors on the eleven vocal characteristics used to measure participant beliefs about their own speech style.

LIA Results

The hypothesis for this study predicted that target age, target race, and target stereotype consistency would affect participant ratings of language competence. The first MANOVA tested the impact of these factors on the memory and hearing dimensions of the LIA and this test was not significant, F(2, 127) = .240, p = .79, η² = .004.

A three-way interaction of target age, age group of participant, and stereotype consistency was significant, F(2, 127) = 4.12, p < .02, η² = .06. Univariate tests revealed this effect to be significant on hearing, F(1, 128) = 6.74, p < .01, η² = .05, but not on memory, F(1, 128) = .47, p = .49, η² = .004. Older adults rated the negatively stereotyped older target as having the most hearing problems, and older adults rated the positively stereotyped young adult targets as having the least hearing problems.

In addition, there was a significant interaction of target age and stereotype consistency, F(2, 127) = 25.26, p < .0001, η² = .28. Univariate tests revealed this effect to be significant on memory, F(2, 128) = 49.60, p < .0001, η² = .28. Negatively stereotyped older targets were rated as having the most memory problems, and positively stereotyped young adults were rated as having the least memory problems. Means are presented in Table 1.

Table 1.

Stereotype by Target Age Interaction on the LIA.

Target age	Positive target	Negative target
Memory
Young	2.62	3.87
Old	3.04	4.47
Recognizes more words
Young	4.96	3.91
Old	4.43	3.80
People enjoy her storytelling
Young	4.19	3.14
Old	4.28	3.46
Does most of the talking
Young	3.99	3.30
Old	3.97	3.38
More to blame when others do not understand
Young	3.27	4.03
Old	3.05	3.99
Conversation difficult
Young	3.38	4.23
Old	2.91	3.94

Note. LIA = Language in Adulthood Questionnaire.

The second MANOVA also tested this hypothesis by examining the impact of these factors on seven individual items from the LIA. Again, the predicted three-way interaction of target age, target race, and stereotype consistency was not significant, F(7, 125) = .45, p = .87, η² = .02. However, the target age by stereotype consistency interaction was significant, F(7, 125) = 5.42, p < .0001, η² = .23. Univariate tests revealed significant effects on several questions: recognizes more and more words, F(1, 131) = 15.71, p < .0001, η² = .11; people enjoy her storytelling, F(1, 131) = 16.93, p < .0001, η² = .11; does most of the talking, F(1, 131) = 6.50, p < .01, η² = .05; more to blame when others do not understand, F(1, 131) = 17.63, p < .0001, η² = .12. According to the means, the positively stereotyped young target recognizes more and more words, and the positively stereotyped older target is the best storyteller. The negatively stereotyped young adult is the most to blame when others don’t understand and this conversation is most difficult. Both the positively stereotyped young and older targets were rated as doing most of the talking. Means are presented in Table 1.

Patronizing Speech Results

The second part of the hypothesis predicted a three-way interaction of target age, target race, and stereotype consistency on participant beliefs about their use of patronizing speech. The predicted interaction was not significant, F(11, 105) = .82, p = .62, η² = .08. However, target race, stereotype consistency, and age group of participant was significant, F(11, 105) = 2.15, p < .02, η² = .18. Univariate tests revealed impacts that approach significance on fast, F(1, 115) = 3.32, p = .07 and understandable, F(1, 115) = 4.82, p = .06, η² = .03. Most participants reported slower speech rates to the stereotype-consistent target than to the stereotype inconsistent target (Older EA participants: M = 2.62, M = 2.99; Younger EA participants: M = 2.98, M = 3.08), with older AA participants being the most extreme in their ratings (M = 2.33, M = 3.98). Of all participants, only the AA young adults reported a faster speech rate toward the stereotype-consistent target (M = 3.55) than to the stereotype-inconsistent target (M = 3.23). As for understandable, older AA participant ratings indicated that they would speak more understandably than older EA participants, regardless of stereotype consistency (AA M = 5.58, EA M = 4.81) or stereotype inconsistency (AA M = 5.44, EA M = 4.94). Young participant ratings, however, indicated that young AA participants would speak less understandably to stereotype-consistent target (M = 4.55) than to stereotype-inconsistent target (M = 5.32) and that young EA participants exhibit the opposite pattern, speaking more understandably to the stereotype-consistent targets (M = 5.80) than to the stereotype-inconsistent target (M = 5.03).

There was also a significant interaction effect of target age and stereotype consistency, F(11, 105) = 1.95, p < .04, η² = .17. Univariate tests revealed a significant effect on fast, F(11, 105) = 17.84, p < .0001, η² = .13. Participants responded that they would speak most slowly to the negatively stereotyped older adult target (M = 2.29), then the negatively stereotyped young target (M = 3.15), and that they would speak most quickly to the positively stereotyped old target (M = 3.49) and to the positively stereotyped young target (M = 3.45).

Participant Race and Speech Style

The third MANOVA also served to answer the research question about the impact of participant race on participant ratings of their speech style. The main effect of participant race was significant, F(11, 105) = 2.92, p < .002, η² = .23. Univariate tests revealed significant effects on understandable, F(1, 115) = 6.44, p < .01, η² = .05; thin, F(1, 115) = 4.65, p < .03, η² = .04; and exaggerated pronunciation, F(1, 115) = 5.67, p < .02, η² = .05. The AA participants indicated that they would speak more understandably, with more thinness, and with more exaggerated pronunciation than would EA participants.

Brief Discussion of Implications and Limitations

These results support Hummert’s (1994, 2007; Hummert et al., 2004) ASI model for the elicitation of patronizing speech, and those stereotypes were more frequently related to age than to race. First, the predicted interaction of target age, target race, and stereotype information was not significant, indicating a lack of intersectionality of these variables on either language performance or the elicitation of patronizing speech. Second, multiple interactions of target age and stereotype information were significant, both on dimensions and items of the LIA, as well as the elicitation of patronizing speech. Whether measuring their impact on language performance or speech style adoption, age groups and stereotypes influenced the outcome. Third, the target’s race only influenced the outcome on two patronizing speech items (fast and understandable) and within the context of participant age and stereotype information. Finally, the significant main effect suggests that race operates independently to influence perceptions. Participant race, not target race, influenced beliefs about the participants’ speech style, such that AA participants believed they would use a speech style with more patronizing speech characteristics than EA participants reported.

These findings are a significant contribution to our understanding of intergenerational encounters but must be tempered by two methodological concerns. First, the LIA is a scale specifically developed to measure age differences in communication behavior. The results of this study are consistent with past research in intergenerational communication by Hummert and various colleagues (e.g., Hummert, Garstka, & Shaner, 1995; Hummert, Garstka, Shaner, & Strahm, 1995). Even so, within the design of this study, the LIA may overemphasize age differences; thus, it would be important to replicate these results with a scale less sensitive to the variable of age. This limitation is addressed in a second experiment; the LIA is replaced with the Communication Competence scale (Wiemann, 1977).

Second, the design of this experiment confounds age and stereotype, requiring stereotype information to be considered as consistent (young-positive, old-negative) versus inconsistent (young-negative, old-positive). As the ASI model suggests, age information should be enough to elicit positive or negative stereotypes of the target (Hummert, 1994, 2007). In an experimental test of this model, then, it is unnecessary and inappropriate to provide stereotype details; rather, simply providing age and race information should provoke positive or negative perceptions, particular beliefs about communication, and result in (un)modified communication behavior. In the second experiment, stereotype characteristics were not provided; this simpler and clearer design was used to test whether age and race would influence beliefs about communication competence and the use of a patronizing speech style. This design change also required a simplification of the hypotheses being tested; therefore, this study tested the following hypotheses:

Hypothesis 1: Older targets will be rated as less competent and more likely to receive patronizing speech than will younger targets.

Hypothesis 2: African American targets will be rated as less competent and more likely to receive patronizing speech than will European American targets.

Hypothesis 3: Target race and target age will interact such that older the African American target will be rated as the least competent and most likely to receive patronizing speech.

Study 2

To test these relationships, a 2 (target race) × 2 (target age) × 2 (participant race) × 2 (participant age) repeated measures design was used. Target and participant race included AA and EA; target and participant age included young and older adults. Each participant received a questionnaire packet containing a photograph of a woman and two scales to rate an imaginary conversation with her. The photographs from the first study were used (young EA, young AA, old EA, old AA), and the order of presentation was counterbalanced. The first scale differed from the first study, replacing the Language in Adulthood Scale with the Communicative Competence Scale (Wiemann, 1977). This is a 36-item Likert-type scale measuring communicative dimensions of interaction management, empathy, affiliation/support, behavioral flexibility, and social relaxation. Participants indicate their agreement with statements such as “This person is an effective conversationalist,” “This person is a good listener,” and “This person is not afraid to speak with people in authority.” Higher scores are indicative of greater communicative competence. The second scale rated the participants’ perceived speech style when interacting with the target and was the same scale used in the first study (described earlier and in Hummert & Shaner, 1994).

Participants

Participants included 117 young adults ranging in age from 18 to 35 years (M age = 23 years), and 71% were female. For the study’s design, only AAs and EAs participated resulting in 52% AA and 48% EA. Participants also included 103 older adults ranging in age from 55 to 92 years (M age = 72 years), and 73% were female. Of the two racial groups, 57% were AA and 43% were EA.

The two age groups differed significantly in their reported years of education, such that the younger group was more educated than the older group, F(1, 216) = 99.75, p < .0005.

Results

The first hypothesis predicted that older targets would be rated as less competent and more likely to receive patronizing speech than young targets. A MANOVA with target age as the IV and competence and patronizing speech variables as the DVs revealed partial support for this hypothesis. The multivariate test revealed a significant impact of target age, F(12, 172) = 2.18, p < .02. Univariate tests revealed that older targets were not rated as less competent than younger targets, though this nonsignificant effect did approach statistical significance, F(1, 12) = 3.04; p = .083, and the means were in the predicted direction (young adult competence M = 126.54; older adult competence M = 119.43). The second part of the hypothesis was supported but only on one of the eleven patronizing speech measures, exaggerated pronunciation, F(1, 12) = 8.78, p < .003. Participants believed they would speak with more exaggerated pronunciation to the older target (M = 3.81) than to the younger target (M = 2.88).

The second hypothesis predicted a main effect of the target’s race on ratings of competence and patronizing speech. This hypothesis was not supported, F(12, 172) = .46, p = .94. Participants did not vary their ratings of competence and use of a patronizing speech style based on race alone.

The third hypothesis predicted an interaction of the target’s age and target’s race on the ratings of competence and patronizing speech. This hypothesis was not supported, F(12, 172) = .91, p = .54.

An unexpected interaction of participant age and target age achieved statistical significance, F(12, 172) = 3.42, p < .0001. The univariate tests revealed a significant effect on three patronizing speech measures, fast, F(1, 12) = 30.68, p < .0001; shrill F(1,12) = 7.69, p < .006; and high-pitched, F(1, 12) = 4.59, p < .03. The pattern of results reveals higher ratings on each of these dimensions within the participants’ own age groups. That is, young participants reported that they would speak more quickly to a younger target (M = 3.86) than to an older target (M = 2.11), and older participants reported that they would speak more quickly to the older target (M = 3.20) than to the younger target (M = 2.35). Similarly, the young participants believed they would speak with more shrillness to the young target (M = 1.88) than to the older target (M = 1.61), and older participants believed they would speak with more shrillness with the older target (M = 2.73) than to the younger target (M = 2.15). Finally, the same in-group age pattern was found for high-pitched; young participants believed they would speak with more high pitch to the younger target (M = 1.98) than to the older target (M = 1.73), and older participants believed they would speak with more high pitch to the older targets (M = 2.75) than to the younger targets (M = 2.21).

Summary

Results of this second experiment indicate higher age than race salience. Two of the hypotheses were not supported, indicating that race did not interact with age to create a double jeopardy situation. The lack of support for Hypothesis 2 further indicates that the target’s race did not influence perceptions of communication competence nor the use of a patronizing speech style. Instead, this experiment supported the influence of age in predicted and unpredicted ways. Specifically, the results supported Hypothesis 1 predicting a main effect of age on communication competence and the use of patronizing speech. The interaction of participant age and target age on the use of a patronizing speech style emphasize age salience and reveal in in-group bias in speech accommodation.

Discussion

First and most significantly, the results of these studies clearly emphasize the impact of age group, whether participant or target age. Almost every single statistically significant result included an age variable. Second, these results do not strongly support a double jeopardy situation for older AAs; instead, race only modestly impacts beliefs about speech style. These results have significant implications for the ASI model (Hummert, 1994, 2007; Hummert et al., 2004) and future research examining intergenerational communication.

The impact of age is quite apparent with the significant interaction of target age, age group of participant and stereotype consistency. Older participant ratings of targets’ hearing were more differentiated than the young participant ratings, suggesting that older adults have a more complex understanding of hearing ability across the lifespan. According to the National Institute on Deafness and Other Communication Disorders (2006), 40% to 50% of adults older than 75 experience some hearing loss; yet only 0.17% of those 18 years or younger have hearing loss, indicating that older adults do, in fact, have more experience with hearing difficulties (either their own or their peers).

The first study’s hypothesis predicted that target race, target age, and stereotype consistency would influence ratings of patronizing speech. This test was not significant, but there was an interaction between target race, stereotype, and age group of participant, indicating that young and older adult participants responded to race and stereotype differently. Although the effect sizes were rather small, the means suggest complicated relationships among these variables, relationships that differ when you examine the impact on ratings of fast versus ratings of understandable. Specifically, older AA participants were the most extreme in their ratings of fast for stereotype-consistent and stereotype-inconsistent targets, and most participants rated their speech as being more slow for the stereotype-consistent targets (further supporting past research by Hummert & Shaner, 1994; Hummert et al., 1998). Young AA participants, however, reported that they would speak more quickly AND less understandably to the stereotype-consistent targets, suggesting that these participants were not impacted by the age-stereotype association. Past intergenerational communication research has not yet included AA participants, and this finding suggests that there may be very interesting cultural differences and age differences between EA participants and AA participants. This warrants further study.

The interaction of target age and stereotype was the most consistent finding in the first study. This interaction was significant for the LIA dimensions, the individual items from the LIA, and vocal characteristics of patronizing speech. These findings are especially important because the effect sizes were larger than the effect sizes for any of the other tests. The consistency and size of this interaction effect suggest these two variables to be very influential. Most of these results are consistent with past research, such as the finding that negatively stereotyped older target was expected to have the worst memory and that the positively stereotyped older target was expected to be the best storyteller (Hummert et al., 1995; Ryan et al., 1992).

Interestingly, young negatively stereotyped targets were downgraded more severely than their older adult counterparts on various communication variables, such as being to blame when others do not understand and that the conversation would be difficult to hold with such a person. This is inconsistent with past research regarding the LIA (e.g., Ryan et al., 1992), but past research has not included stereotype as a variable (when also examining young adults and performance on the LIA). Expectancy-violation may explain these findings. In general, younger adults are perceived more positively than older adults, so when they are described negatively, it is such a difference from the expectation, that these individuals are more negatively evaluated as a result. Ryan et al. did find that younger respondents reported fewer language problems than did older respondents, and young and old respondents both perceived typical 75-year-old targets to have more problems than typical 25-year-old targets.

The finding that the positively stereotyped young target would recognize more and more words may be a result of the targets appearing to be college-aged and of the participants believing that college students are more likely to be learning new words than those who are not in college. Ryan et al. (1992) also found the young to be favored by this vocabulary item on the LIA. Similarly, positively stereotyped young and old targets were believed to do most of the talking. Participants may have believed that these individuals were more engaged in life and therefore, more likely to reveal that engagement through interaction. In this study, “doing most of the talking” may not have been perceived as a communication problem, as also found by Hummert et al. (1995).

A significant interaction of target age and stereotype consistency revealed that participants would speak most slowly to the negatively stereotyped older target and most quickly to the positively targets, regardless of age. This result is consistent with Hummert and Shaner (1994) and with Hummert’s (1994, 2007) ASI model depicting the differential influence of positive vs. negative stereotype information. Normal adult speech results from positive stereotype activation, but overaccommodative speech (e.g., slower speech rate) results from negative stereotype activation.

Intersectionality suggests that various demographic features, such age, gender, and race interact in multiplicative ways, but the results from both of these studies reveal little evidence of age and race intersecting on ratings of competence and speech style. In fact, race had minimal impact on the variables studied here. Three hypotheses included race (one in the first study and two in the second study), and none of these hypotheses were supported by the results.

The first study found a main effect of participant race on the vocal characteristics used to measure patronizing speech. Participants report using different speech styles; the AA participants said that they would speak more understandably, with more thinness, and with more exaggerated pronunciation than would EA participants. Popp, Donovan, Crawford, Marsh, and Peele (2003) found that perceptions of Black students’ speech included that it was emotional, loud, demanding, less socially appropriate, and accompanied by exaggerated gestures. In contrast, this study involves self-reporting of communication behaviors but does find some similarities. If speech styles do differ for AAs and EAs, does that also mean that patronizing speech may take on a different form? Research on patronizing speech has thus far focused on EA/White speakers and targets (e.g., Culbertson & Caporael, 1983; Harwood et al., 1993; Hummert, 2007; Hummert & Shaner, 1994).

Results of these studies support a robust impact of age group and little impact for racial group. The quasi-experimental nature of these studies, however, deserves caution in the interpretation of results. Three concerns should be noted and addressed by future studies examining intersectionality and employing a variety of methods in communication research.

First, the self-report instruments were rather simplistic. Hummert et al. (1994, 1998) used this patronizing speech measure but in combination with psycholinguistic characteristics of actual messages produced by participants. More realistic stimulus materials (i.e., not black and white photographs) and more realistic message measures could be used in a similar experiment. Having participants interact with trained confederates could enable researchers to tease apart the effects of age, race, and stereotype in a more realistic scenario. These formats would also allow content analysis of actual speech and a wider array of patronizing speech measures, such as speech rate, volume, length, emotional tone, and simplification.

Second, this study only examined perceptions of female targets and two racial/ethnic groups. Since there is some evidence for perceiving older males and females differently (e.g., Hummert et al., 1995), including gender in another experiment would be an important step. Examining other racial/ethnic groups would broaden our understanding of how race may interact with age for other cultural communities. Positive and negative stereotypes associated with race/ethnicity might also influence the relationship between age and race, as being explored by Kang, Chasteen, Cadieux, Cary, and Syeda (2014) in their work on multiply-categorizable individuals. Furthermore, other components of the ASI model were not included in this study (e.g., cognitive complexity, quality of contact), and the impact of these features may vary across cultural groups thus impacting stereotype activation in different ways. Future research could capture this with more comprehensive designs.

Third, stereotypes used in the first study were developed with data from EA participants (Hummert et al., 1994). AAs may have slightly or drastically different stereotypes associated with older adults, and these stereotypes (with their appropriate traits) should be identified. Since it has been 20 years since Hummert and colleagues identified these traits and stereotypes, it would be appropriate to combine age and racial/ethnic groups to determine if those stereotypes have changed over time or if they vary across racial/ethnic groups. Focus groups or mixed method approaches may best enable researchers to understand how age, race (as well as gender, education, and other variables) intersect to affect the communication outcome. These projects could connect research by Hummert et al. (1994, 2007; Hummert et al., 2004) to research by Street and colleagues (Street et al., 2008; Teal & Street, 2009).

As a process of intergroup communication, intergenerational communication research could be informed through more systematic analysis of intercultural interactions. These studies do not provide evidence suggesting a need to add race into the ASI model, but they are only a first step in understanding the possible intersections in communication behavior. Additional research needs to determine if the in-group/outgroup processes are similar or if they vary across racial/ethnic groups. Regardless, these studies reveal that age continues to play a significant role in the elicitation of stereotypes and speech accommodation. Further research on these variables will lead to a more complex understanding of intergenerational communication in today’s diverse world.

Footnotes

Acknowledgements

The authors thank Carolyn Atkinson, Harold Kinney, and Darlene Simmons for their help with data collection; Nneka Logan for her help with data entry; and Teri Garstka for her help with data analysis. The authors are also grateful for the constructive criticism provided by two anonymous reviewers and Howie Giles; their insights were valuable in the revision of this article.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The second study was funded by a Research Initiation Grant from the Georgia State University Research Foundation, first author as Principal Investigator.

Notes

Author Biographies

Jaye L. Atkinson (PhD, University of Kansas) is an associate professor in the Department of Communication at Georgia State University. Her research focuses on the intersection between stereotypes of older adults and communication in interpersonal, mediated and sport contexts. She has published in communication, gerontology, and sport journals.

Robin G. Sloan is an MA in communication from Georgia State University. Since 2004, she has worked at the Centers for Disease Control and Prevention in Atlanta, where she is a health communication specialist, assisting in the development of training and educational products.

References

Anderson

Harwood

Hummert

M. L.

(2005). The grandparent-grandchild relationship: Implications for models of intergenerational communication. Human Communication Research, 31, 268-294.

Barnes

L. L.

Mendes

Leon

C. F.

Wilson

R. S.

Bienias

J. L.

Bennett

D. A.

Evans

D. A.

(2004). Racial differences in perceived discrimination in a community population of older blacks and whites. Journal of Aging and Health, 16, 315-337.

Culbertson

G. H.

Caporael

L. R.

(1983). Baby talk speech to the elderly: Complexity and content of messages. Personality and Social Psychology Bulletin, 9, 305-312.

Donahue

W. A.

(1985). Ethnicity and mediation. In Gudykunst

W. B.

Stewart

Ting-Toomey

(Eds.), Communication, culture and organizational processes (pp. 134-154). Beverly Hills, CA: Sage.

Fiske

S. T.

(2000). Stereotyping, prejudice, and discrimination at the seam between the centuries: Evolution, culture, mind, and brain. European Journal of Social Psychology, 30, 299-322.

Fujioka

(1999). Television portrayals and African-American stereotypes: Examination of television effects when direct contact is lacking. Journalism and Mass Communication Quarterly, 76, 52-75.

Giles

(Ed.). (2016). Communication accommodation theory: Negotiating personal relationships and social identities across contexts. Cambridge, England: Cambridge University Press.

Giles

Fox

Harwood

Williams

(1994). Talking age and aging talk: Communicating through the life-span. In Hummert

M. L.

Wiemann

J. M.

Nussbaum

J. F.

(Eds.), Interpersonal communication in older adulthood. Interdisciplinary theory and research (pp. 130-161). Newbury Park, CA: Sage.

Giles

Fox

Smith

(1993). Patronizing the elderly: Intergenerational evaluations. Research on Language and Social Interaction, 26, 129-149.

10.

Harwood

Giles

(1996). Reactions to older people being patronized: The roles of response strategies and attributed thoughts. Journal of Language and Social Psychology, 15, 395-421.

11.

Harwood

Giles

Fox

Ryan

E. B.

Williams

(1993). Patronizing young and elderly adults: Response strategies in a community setting. Journal of Applied Communication Research, 21, 211-226.

12.

Hecht

M. L.

Collier

M. J.

Ribeau

S. A.

(1993). African-American communication. Newbury Park, CA: Sage.

13.

Hummert

M. L.

(1994). Stereotypes of the elderly and patronizing speech. In Hummert

M. L.

Wiemann

J. M.

Nussbaum

J. F.

(Eds.), Interpersonal communication in older adulthood: Interdisciplinary theory and research (pp. 162-184). Thousand Oaks, CA: Sage.

14.

Hummert

M. L.

(2007). As family members age: A research agenda for family communication. Journal of Family Communication, 7, 3-21.

15.

Hummert

M. L.

Garstka

T. A.

Ryan

E. B.

Bonnesen

J. L.

(2004). The role of age stereotypes in interpersonal communication. In Nussbaum

J. F.

Harwood

(Eds.), The handbook of communication and aging (2nd ed., pp. 91-114). Mahwah, NJ: Lawrence Erlbaum.

16.

Hummert

M. L.

Garstka

T. A.

Shaner

J. L.

(1995). Beliefs about language performance: Adults’ perceptions about self and elderly targets. Journal of Language and Social Psychology, 14, 235-259.

17.

Hummert

M. L.

Garstka

T. A.

Shaner

J. L.

(1997). Stereotyping of older adults: The role of target facial cues and perceiver characteristics. Psychology and Aging, 12, 107-114.

18.

Hummert

M. L.

Garstka

T. A.

Shaner

J. L.

Strahm

(1994). Stereotypes of the elderly held by young, middle-aged, and elderly adults. Journal of Gerontology: Psychological Sciences, 49, 240-249.

19.

Hummert

M. L.

Garstka

T. A.

Shaner

J. L.

Strahm

(1995). Judgments about stereotypes of the elderly: Attitudes, age associations, and typicality ratings of young, middle-aged, and elderly adults. Research on Aging, 17, 168-189.

20.

Hummert

M. L.

Shaner

J. L.

(1994). Patronizing speech to the elderly as a function of stereotyping. Communication Studies, 45, 145-158.

21.

Hummert

M. L.

Shaner

J. L.

Garstka

T. A.

Henry

(1998). Communication with older adults: The influence of age stereotypes, context, and communicator age. Human Communication Research, 25, 124-151.

22.

Kang

S. K.

Chasteen

A. L.

Cadieux

Cary

L. A.

Syeda

(2014). Comparing young and older adults’ perceptions of conflicting stereotypes and multiply-categorizable individuals. Psychology and Aging, 29, 469-481.

23.

Kochman

(1981). Black and white styles in conflict. Chicago, IL: University of Chicago Press.

24.

Kochman

(1994). Black and White cultural styles in pluralistic perspective. In Weaver

G. R.

(Ed.), Culture, communication, and conflict: Readings in intercultural relations (pp. 285-304). Needham Heights, MA: Ginn.

25.

Mallinson

Brewster

Z. W.

(2005). “Blacks and bubbas”: Stereotypes, ideology, and categorization processes in restaurant servers’ discourse. Discourse and Society, 16, 787-807.

26.

Martin

J. N.

Hecht

M. L.

Larkey

L. K.

(1994). Conversational improvement strategies for interethnic communication: African American and European American perspectives. Communication Monographs, 61, 236-254.

27.

Millard

J. E.

Grant

P. R.

(2006). The stereotypes of black and white women in fashion magazine photographs: The pose of the model and the impression she creates. Sex Roles, 54, 659-673.

28.

National Institute on Deafness and Other Communication Disorders. (2006). http://www.nidcd.nih.gov/health/statistics/hearing.asp

29.

Nussbaum

J. F.

Pecchioni

L. L.

Robinson

J. D.

Thompson

T. L.

(2000). Communication and Aging. Mahwah, NJ: Lawrence Erlbaum.

30.

Orbe

(1995). African American communication research: Toward a deeper understanding. Western Journal of Communication, 59, 61-78.

31.

Pew Research Center. (2013). King’s dream remains an elusive goal; Many Americans see racial disparities. Retrieved from http://www.pewsocialtrends.org/2013/08/22/kings-dream-remains-an-elusive-goal-many-americans-see-racial-disparities/

32.

Popp

Donovan

R. A.

Crawford

Marsh

K. L.

Peele

(2003). Gender, race, and speech style stereotypes. Sex Roles, 48, 317-325.

33.

Rittenour

C. E.

Cohen

E. L.

(2016). Viewing our aged selves: age progression simulations increase young adults’ aging anxiety and negative stereotypes of older adults. International Journal of Aging and Human Development, 82, 271-289.

34.

Ryan

E. B.

Bourhis

R. Y.

Knops

(1991). Evaluative perceptions of patronizing speech addressed to elders. Psychology and Aging, 6, 442-450.

35.

Ryan

E. B.

Giles

Bartolucci

Henwood

(1986). Psycholinguistic and social psychological components of communication by and with the elderly. Language and Communication, 6, 1-24.

36.

Ryan

E. B.

Hummert

M. L.

Boich

L. H.

(1995). Communication predicaments of aging: Patronizing behavior toward elder adults. Journal of Language and Social Psychology, 14, 144-167.

37.

Ryan

E. B.

Kennaley

D. E.

Pratt

M. W.

Shumovich

M. A.

(2000). Evaluations of staff, residents, and community seniors of patronizing speech in the nursing home: Impact of passive, assertive, or humorous responses. Psychology and Aging, 15, 272-285.

38.

Ryan

E. B.

Kwong See

Meneer

W. B.

Trovato

(1992). Age-based perceptions of language performance among younger and older adults. Communication Research, 19, 423-443.

39.

Street

R. L.

O’Malley

K. J.

Cooper

L. A.

Haidet

(2008). Understanding concordance in patient-physician relationships: Personal and ethnic dimensions of shared identity. Annals of Family Medicine, 6, 198-205.

40.

Teal

C. R.

Street

R. L.

(2009). Critical elements of culturally competent communication in the medical encounter: A review and model. Social Science and Medicine, 68, 533-543.

41.

Wiemann

J. M.

(1977). Explication and test of a model of communicative competence. Human Communication Research, 3, 195-213.