Aging and the Social Ecology of Everyday Interpersonal Perception: What is Perceived, in Whom, and Where?

Abstract

Objectives

Despite a proliferation of research in interpersonal perception and aging, no research has identified the nature of the social and emotional perceptions made by aging individuals in everyday life. In this study, we aimed to identify the social ecological features that characterize everyday interpersonal perception across the adult lifespan.

Method

Three studies were conducted. Study 1 identified and compared the targets and locations of young, middle-age, and older adults’ everyday interpersonal perceptions; these perceptions were categorized into types in Study 2. Study 3 applied these categorizations to identify and compare the social ecology surrounding aging individuals’ interpersonal perceptions.

Results

Everyday interpersonal perceptions were directed toward familiar others and occurred in familiar locations, although the specific familiar targets and locations sometimes varied significantly with age. However, the types of perceptions made in everyday life did not vary significantly between age groups.

Discussion

Aging individuals make similar types of interpersonal judgments, but the targets and locations of these judgments may change with age. Future studies on interpersonal perception and aging will need to account for these features of the aging individual’s social ecology to provide an accurate assessment of the aging process.

Interpersonal perception—judging other people’s feelings and traits—is an integral component of healthy social functioning across the lifespan. Indeed, older adults who are more skilled in knowing how other people feel report higher levels of well-being than their less skilled counterparts (Phillips, Scott, Henry, Mowat, & Bell, 2010). Although there is some consensus that normative aging is characterized by declines in interpersonal perception (Isaacowitz, Livingstone, & Castro, 2017), researchers continue to disagree about the factors driving these declines. Recent arguments suggest that studies in interpersonal perception and aging rely on traditional laboratory paradigms that lack ecological validity and result in lower performance for older adults (Isaacowitz & Stanley, 2011; Kunzmann & Isaacowitz, 2017). We address this possibility by considering the nature of interpersonal perception in old age.

Past research has neglected the aging individual’s social ecology. In real life, social interactions involve specific social partners (e.g., romantic partners, family members, and friends) and these interactions are embedded within specific settings (e.g., the physical location). This contextual information—including the familiarity of the social partner and the location where the perception takes place—can be used to identify how a social partner is feeling or behaving. Because older adults are more likely than younger adults to utilize contextual information when making interpersonal judgments (Noh & Isaacowitz, 2013; 2015), such information is critical for our understanding of age-related differences in interpersonal perception; laboratory paradigms may or may not tap into the types of interpersonal perception skills employed by aging individuals in daily life, resulting in an incomplete understanding of the interpersonal perception aging process.

The present study is a first step in identifying the nature of everyday interpersonal perception in adulthood. We present findings from three adult lifespan studies that utilized retrospective reports to describe and compare the social ecology of young, middle-age, and older adults’ everyday interpersonal perception.

Aging and the Social Ecology of Interpersonal Perception

The perceptions we make about others in daily life involve a variety of common features including the specific people judged (target), the type of judgment made (type), as well as other supplemental contextual information, such as the physical location in which the judgment took place (location). These features may have been neglected by previous research because they are difficult to standardize and measure in the laboratory. Yet, by not including these features, past studies make several assumptions about the aging individual’s social ecology, and subsequently, about interpersonal perception in old age.

First, most studies on interpersonal perception rely on paradigms that use actors as perception targets, assuming that strangers are the typical or most important target in the aging individual’s social ecology. Second, most studies have focused on the perception of basic emotional states (e.g., happy, sad, angry, fear, surprise, and disgust), implying that such perceptions are the most frequent or important type of perception made by aging individuals. Finally, because past studies generally omit context from their stimuli (for exceptions, see Noh & Isaacowitz, 2013; 2015), there is an assumption that the locations where interpersonal perceptions take place are irrelevant to the process of interpersonal perception in old age. These assumptions may constrain our understanding of the specific environmental features involved in judging others’ feelings and traits, including the extent to which these features vary systematically with age.

Targets

Recent conceptualizations of interpersonal perception emphasize the target, or the person conveying their feelings or traits (Castro, Cheng, Halberstadt, & Grühn, 2016). Real-world interpersonal perception involves specific targets: you may judge how angry your spouse feels during an argument or whether your physician appears overly concerned about your test results. You may even make judgments about people you do not know: you may judge whether the grocery store cashier seems tired or whether the person driving next to you is honking their horn out of concern or irritation. Thus, the targets of interpersonal perception in everyday life may span a dimension of familiarity, ranging from individuals that we are very familiar with to individuals we regularly interact with but do not know well to complete strangers.

The familiarity of targets is likely to vary with age. The social convoy model suggests that individuals select social partners who provide emotional and social support (Kahn & Antonucci, 1980). Extending this idea to lifespan development, socioemotional selectivity theory suggests that these social convoys become more selective with age, given age-related shifts in the prioritization of emotional well-being (Carstensen, Isaacowitz, & Charles, 1999). Consistent with these theoretical models, older adults’ social networks are smaller and more selective than those of younger adults (Carstensen, 1992; Wrzus, Hänel, Wagner, & Neyer, 2013) and consist primarily of close or familiar others (Antonucci & Akiyama, 1987). Moreover, interactions with family members are more frequent than interactions with non-family members in old age (Sander, Schupp, & Richter, 2017). Consequently, as compared to younger adults, older adults may be more experienced in perceiving familiar targets, who are more likely to be older adults’ social partners.

Types

There are many different types of judgments we might make about others. Consider the earlier example of an argument with one’s spouse: During such an argument, you may look to your spouse’s face, voice, or body for cues that indicate their emotional state and conclude after identifying the knitted brows, snarled lips, and clenched fists that your spouse is likely feeling angry. This example illustrates the most commonly studied type of interpersonal perception—the judgment of someone else’s emotional state. However, this is only one among many interpersonal perceptions made in daily life. For example, individuals readily make judgments about others’ personality and character traits (Hareli & Hess, 2010; Todorov & Uleman, 2002). In the case of a spousal argument, you might judge whether your spouse is being dominant, aggressive, or rude. The extent to which different types of interpersonal perceptions are more or less frequent and more or less important with age remains unclear.

Younger and older adults encounter different social situations that may afford different perceptual opportunities. Both the quantity and quality of social interactions change with age, as older adults experience less social interactions overall, and less interpersonal arguments and tensions specifically, than do younger adults (Almeida, 2005). Moreover, social situations may be less stressful in old age (Charles, Piazza, Luong, & Almeida, 2009; Lefkowitz and Fingerman 2003) and may be characterized by greater positivity (Murry & Isaacowitz, 2016) and less negativity (Miller, Charles, & Fingerman, 2009). Older adults, therefore, may have limited opportunities to perceive other people’s negative feelings and traits in everyday life. Instead, older adults may have greater opportunities to make positive interpersonal perceptions. According to socioemotional selectivity theory, older adults shift their priorities toward maximizing positive affect and wellbeing due to viewing their time as limited (Carstensen et al., 1999). Consequently, the salience of positive emotions increases with age (Ferrari, Bruno, Chattat, & Codispoti, 2017; Mather & Carstensen, 2005) and older adults may direct their resources toward making positive interpersonal perceptions (e.g., judging others’ happiness or kindness) and away from making negative interpersonal perceptions (e.g., judging others’ anger or greed) as compared to younger adults.

Locations

Although context may support older adults’ interpersonal perception skills (Ngo & Isaacowitz, 2015; Stanley & Isaacowitz, 2015), no work has identified the physical contexts in which judgments about others take place. Interpersonal perceptions likely occur in both familiar—one’s home—and unfamiliar—a new restaurant—settings. The frequency and importance of these settings may vary with age. Given that older adults interact more with family members than non-family members (Sander et al., 2017), older adults’ interpersonal perceptions may occur most frequently in familiar locations—their own homes or the homes of family members. In contrast, younger adults may experience greater contextual mobility and diversity in their daily lives, and thus, their social ecology may be equally characterized by familiar and unfamiliar settings.

Research Objectives

The goal of the present research was to describe the social ecology of interpersonal perception in adulthood. Consistent with past research on interpersonal perception and aging, we identified and compared the social ecology of young and old adults. Given the paucity of work on interpersonal perception in midlife, we also investigated middle-age adults, allowing for a comprehensive assessment of the adult lifespan.

To achieve our goal, we conducted three studies. Study 1 compared the targets and locations of young, middle-age, and older adults’ everyday interpersonal perceptions. Study 2 categorized these perceptions into types. Study 3 used these categorizations to identify and compare the social ecology surrounding young, middle-age, and older adults’ everyday interpersonal perceptions.

Study 1

Consistent with socioemotional selectivity theory and the social convoy model, we hypothesized that the frequency of specific targets and locations would vary with age: Familiar targets (e.g., spouse/partners, family members and friends) and familiar locations (e.g., perceivers' own home) would be more frequent for older adults than for younger and middle-age adults.

Method

Participants

Six hundred and twelve adults were recruited on Amazon Mechanical Turk (MTurk). (Different participants were sampled in each study. The same inclusion criteria [US residency, 90% approval rate] were applied. In each study, participants were recruited on MTurk and directed to a Qualtrics survey). Evidence of the validity and reliability of Mturk as an online data collection platform in aging studies is accumulating (Williamson et al., 2013). Participants included young adults ages 18–39 (N = 219; M_age = 29.83, SD = 5.10; 71.2% female), middle-age adults ages 40–64 (N = 198; M_age = 49.23, SD = 7.00; 67.7% female), and older adults ages 65–81 (N = 195; M_age = 67.48, SD = 3.07; 55.7% female). All participants resided in the United States and had MTurk worker approval rates ≥ 90%, further supporting their reliability as participants. Participants reported a range of ethnicities (77.94% White, 8.99% Black, 4.41% Asian American/Pacific Islander, 2.94% Hispanic/Latino, 0.82% Native American, 4.90% Multiracial) and income levels (Median = $49,000; range: $0 to $1,000,000) and were in good health (M = 2.30, SD = 0.93).

Procedure

Participants first completed a five-item demographic screening questionnaire that assessed participants’ country of residence, sex, income, age, and ethnicity. Age quotas were used to balance sampling across age groups. Adults who met age requirements were directed to an informed consent page. Following consent, participants were randomly administered 10 blocks of questions pertaining to 10 different interpersonal perceptions. Questions were self-paced to allow participants sufficient time to respond to all items. After completing the 10 blocks of questions, participants reported their date of birth (to serve as a reliability check for age), religious affiliation, and health status. Participants were debriefed and compensated $1.50 for their participation.

Measures

Demographics

Participants first reported their country of residence, sex, annual household income (in U.S. dollars), age (in years), and ethnicity. At the end of the survey, participants reported their date of birth (mm/dd/yyyy) and their rating of their health at the present moment on a five-point scale ranging from 0 (Poor) to 4 (Excellent).

Interpersonal perceptions

Forty interpersonal perceptions were identified through piloting and discussion. We first generated a diverse list of interpersonal perceptions by asking lab members to document the interpersonal judgments they made throughout a single day; lab members were also encouraged to ask their friends and family for similar input. As our goal was to generate as many different judgment types as possible for preliminary investigation, these pilot responses were then aggregated and supplemented with additional discussion among the authors. This process resulted in the final list of 40 interpersonal perceptions (see supplemental materials). To reduce redundancy and participant burden, participants were randomly assigned 10 interpersonal perception blocks, resulting in a between-subjects design with 130 to 160 participants per perception and approximately equal amounts of young, middle age, and older adults per perception.

Participants were presented with a general definition of interpersonal perception (“judgments that people make about others in everyday life including judgments about types of people and judgments about behaviors that people display”). Participants were then presented with 10 blocks of questions pertaining to specific interpersonal perception events. After reading a brief layperson’s definition of the perception (generated during the piloting procedures discussed above), participants were asked to recall a time in the past 3 months when they made a particular judgment about someone else. Example questions and layperson definitions are provided as supplemental materials. After anchoring their report to a specific interpersonal perception event, participants provided information regarding the target and location of the perception.

Target

Participants were asked to indicate the target of each interpersonal perception from a predetermined list of parent, child, spouse/partner, other relative, friend, coworker, neighbor, physician, pharmacist, nurse, sales clerk, receptionist, wait staff, mail clerk, customer service representative, telemarketer, stranger, or other person. This list was generated through piloting as described above. Participants could also indicate “no one” if they did not recall making a given perception in the past 3 months. A final response choice of “I don’t understand what perceiving someone’s [quality/attribute] means” was provided in case participants failed to understand the meaning of the attribute. Participants indicated a failure to understand the interpersonal perception in only 1.5% of reported perceptions, indicating that participants generally understood the instructions.

Location

Participants responded to an open-ended question about the location where each interpersonal perception took place. Responses were typed into text entry boxes and independently coded by two research assistants. Codes were first generated using an open coding scheme and then integrated into coherent and well-defined categories. Location was coded using five categories: the perceiver’s home, someone else’s home (like a relative, friend, neighbor, or coworker), work or school, a specific community setting (restaurant, bar, store, bank, driving on road, park, or church), or some other location (anything that did not fit into the four previous categories). Inter-rater reliability on 30% of codes was high (M_kappa = 0.84; range = 0.74–0.95).

Results

We compared the frequencies of different targets and locations across age groups using contingency table analyses and Bonferroni post hoc comparisons. Table 1 presents the target and location frequencies for each age group. Subscript letters denote age group comparisons; groups with different letters varied significantly in target/location frequency (i.e., frequency was higher for one group) whereas groups with the same letters did not vary significantly. Bayes factors were calculated in JASP to quantify evidence for the alternative hypothesis (frequencies vary across age groups) relative to the null hypothesis (frequencies are the same across age groups); Bayes factors >30 indicate very strong evidence for the alternative hypothesis, whereas factors <0.03 indicate very strong evidence for the null hypothesis (Wagenmakers, Wetzels, Borsboom, & van der Maas, 2011).

Table 1.

Target and Location Frequencies by Age Group (Study 1)

	Age Group
	Young Adults (N = 219)	Middle-Age Adults (N = 198)	Older Adults (N = 195)
Target
Parent	182a	90b	20c
Child	258a	300b	287b
Spouse/partner	377a	249b	244b
Relative	71a	100b	123b
Friend	236a	150b	178a, b
Coworker	172a	191a	71b
Neighbor	30a	39a	92b
Physician	14a	22a	25a
Pharmacist	3a	10a	11a
Nurse	6a	16a,b	17b
Sales clerk	20a	24a	34a
Receptionist	6a, b	1b	11a
Wait staff	12a	15a,b	27a
Mail clerk	3a	5a	6a
Customer service representative	15a,b	9b	26a
Telemarketer	5a	8a,b	17b
Stranger	109a	66b	75a, b
Other	72a	73a	36b
No one	560a	597b	605b
Target total	2,151	1,965	1,905
Location
Perceiver’s home	695a	526b	515b
Someone Else’s home	136a	147a,b	168b
Work or school	241a	248a	97b
Community setting	371a	324a	401b
Other	121a	121a	117a
Location total	1,564	1,366	1,298

Note. Different subscript letters denote significant age group comparisons (at p < .05). To account for multiple pair-wise comparisons, a Bonferroni adjustment was used.

Target

Age group was significantly associated with target frequency, χ²(36) = 375.08, p < .001, and there was very strong evidence for the alternative hypothesis relative to the null hypothesis, BF₁₀ >30. There was mixed evidence for a familiar target effect: some familiar targets (e.g., other relative) were more frequently reported by older adults than by younger adults whereas other familiar targets (e.g., spouse/partner) were more frequently reported by younger adults than by older adults. Moreover, some familiar targets (e.g., friend) did not vary significantly between younger and older adults’ reports. Middle-age adults often fell between these two groups and reported similar familiar target frequencies as older adults.

Location

Age group was significantly associated with location frequency, χ²(8) = 97.77, p < .001, and there was very strong evidence for the alternative hypothesis relative to the null hypothesis, BF₁₀ >30. There was mixed evidence for a familiar location effect: Older adults’ interpersonal perceptions took place more frequently in someone else’s home as compared to younger adults, however, younger adults more frequently reported that interpersonal perceptions took place in their own home as compared to older adults. Moreover, older adults reported community settings more frequently than did younger adults. Again, middle-age adults fell between the two groups and were generally more similar to older adults in their reported location frequencies.

Discussion

Our hypothesis regarding age-related differences in the familiarity of targets was not supported—familiar targets were sometimes more frequently reported by younger adults than by older adults. However, there was a familiarity effect within older adults: familiar others like a spouse/partner, relative, or friend were more frequently reported than non-familiar others as the targets of older adults’ everyday interpersonal perceptions. These patterns are consistent with theoretical perspectives of increased social selectivity in old age (e.g., Antonucci & Akiyama, 1987; Carstensen, 1992) and with recent findings demonstrating that the frequency of interactions with familiar social partners (i.e., family members) is stable across the lifespan but that interactions with non-familiar social partners (i.e., neighbors, acquaintances) decrease with age (Sander et al., 2017).

Our hypothesis regarding familiar locations was similarly not supported: interpersonal perceptions that took place in the participant’s own homes were more frequently reported by younger adults than by middle-age or older adults. Notably, older adults’ interpersonal perceptions took place more frequently in community settings than in others’ homes, and community settings were more frequently reported by older adults than by middle-age or younger adults, suggesting that community settings may be relevant for older adults’ everyday interpersonal perception.

Although these findings describe the targets and locations of the aging individual’s social ecology, information regarding the types of interpersonal perceptions made in everyday life remains unclear. Because our goal in Study 1 was to generate a diverse list of all the possible interpersonal perceptions made in daily life, the categorization of these perceptions into meaningful age-invariant types requires validation. Moreover, age group comparisons among different types of interpersonal perceptions were not possible in Study 1 due to the unbalanced between-subjects design (where perceptions were randomly assigned between participants). Studies 2 and 3 were designed to address these limitations, respectively.

Study 2

The 40 interpersonal perceptions from Study 1 were categorized into types according to their quality (emotional or social), timing (state or trait), and valence (positive, negative, neutral).

Method

Participants

Fifty young adults (M_age = 30.08, SD = 4.59; range: 21–39; 48% female), 48 middle-age adults (M_age = 49.04, SD = 7.00; range: 40–63; 58.3% female), and 50 older adults (M_age = 68.12, SD = 3.21; range: 65–77; 68.8% female) were recruited for participation on MTurk. Participants were predominantly White (87.8%) and varied in their levels of income (Median = $40,000; range: $0–$175,000) and health (M = 3.35, SD = 0.95).

Procedure

Participants completed the five-item demographic questionnaire used in Study 1. Following consent, participants completed three sorting activities; activity order was randomized across participants. In each activity, participants were asked to sort the 40 interpersonal perceptions from Study 1 into three categories. One activity asked participants to indicate whether each perception was about something social, emotional, or neither social or emotional (termed herein as quality categorization), another activity asked participants to indicate whether each perception was about a state, a trait, or neither a state or trait (termed herein as timing categorization), and the remaining activity asked participants to indicate whether each perception was about something positive, negative, or neutral (termed herein as valence categorization). After completing the sorting activities, participants reported their date of birth and health status. Participants were debriefed and compensated $1.50 for their participation.

Measures

Participants completed the same demographic information as in Study 1. Next, participants categorized the 40 interpersonal perceptions as to their quality, timing, and valence. Participants were first presented with instructions for the sorting task (e.g., “On the next page, you will see a list of perceptions that you might make about someone else. Your job is to classify them into “Social,” “Emotional,” or “Neither”) and definitions of each sorting category (e.g., “Social: These perceptions describe how someone interacts with or relates to other people”). Participants were then instructed to drag each perception to its appropriate category; comprehension of these instructions was verified with an instructional sorting task (e.g., “Sort the labels ‘Social,’ ‘Emotional,’ and ‘Neither’ into the category boxes of the same name”).

Results

To determine how young, middle-age, and older adults typically categorize interpersonal perceptions, we first calculated the frequencies for the categorization of each of 40 interpersonal perceptions as “Social” “Emotional” or “Neither,” as “State” “Trait” or “Neither,” and as “Positive” “Negative” or “Neutral” for each age group. We identified the interpersonal perceptions that were classified into the same category by at least 2/3 of participants across and within each age group. From this list, we identified the two interpersonal perceptions that were classified at the highest rates (i.e., classified similarly by the largest number of participants across age groups) for the following six categories: Social (dominance, playfulness), Emotional (anger, sadness), State (fatigue, fear), Trait (creativity, easygoingess), Positive (happiness, kindness), and Negative (greed, rudeness) (Too few interpersonal perceptions were classified as “Neither” or “Neutral” to be included in this step.). Finally, we utilized logistic regression to determine whether the quality, timing, and valence, categorization odds varied by age; across models, age group was not significant (ps: 0.101–0.786), confirming that there were no significant age group differences in the categorization of these perceptions. Table 2 presents the final list of twelve interpersonal perceptions; full frequency data are presented in supplemental materials.

Table 2.

Interpersonal Perception Quality, Timing, and Valence Categorization (Study 2)

	Categorization
	Quality	Timing	Valence
Perception
Anger	Emotion	State	Negative
Creativity		Trait	Positive
Dominance	Social	Trait
Easygoingness	Social	Trait	Positive
Fatigue		State
Fear	Emotion	State	Negative
Greed			Negative
Happiness	Emotion	State	Positive
Kindness		Trait	Positive
Playfulness	Social		Positive
Rudeness			Negative
Sadness	Emotion	State	Negative

Discussion

Young, middle-age, and older adults shared some understanding of the different types of interpersonal perceptions made in everyday life, as at least one-third of the original 40 interpersonal perceptions were similarly categorized according to their quality, timing, or valence across age groups. Whether or not these twelve different types of interpersonal perceptions vary in importance and frequency with age was the focus of Study 3.

Study 3

Given the findings from Study 1, we predicted that some familiar targets—other relative—would be more frequent in older adults’ reported interpersonal perceptions than those of younger adults, and that other familiar targets—spouse/partner—would be more frequent in the perceptions reported by younger adults than older adults. Given the surprising location results of Study 1, we made no specific predictions regarding age differences. Consistent with theoretical perspectives that emotional—and particularly positive emotional—goals are prioritized in old age (e.g., Carstensen et al., 1999), we predicted that emotional and positive perceptions would be rated by older adults as more important and more frequent than social or negative perceptions and in comparison to younger adults.

Method

Participants

Fifty-two young adults (M_age = 30.04, SD = 5.19; range: 20–39; 50% female), 53 middle-age adults (M_age = 48.43, SD = 5.95; range: 40–59; 64.2% female), and 51 older adults (M_age = 67.31, SD = 3.19; range: 65–780; 56.9% female) were recruited for participation on MTurk. Participants were mostly White (85.3%) and of moderate income levels (Median = $50,500; range: $2,000 to $225,000) and reported good health (M = 3.31, SD = 0.91).

Procedure

Procedures were identical to Study 1 with one revision: Participants received 12 randomized blocks of questions pertaining to the 12 interpersonal perceptions validated in Study 2.

Measures

Measures were identical to Study 1 except that participants also rated the importance and frequency of each interpersonal perception. After reporting the target and location, participants rated the degree of importance in making a given interpersonal perception to their well-being on a five-point scale (1= Not at all important to 5= Very important). Participants were then asked to “Think of the last week. How often in your interactions with other people have you had to judge someone’s [quality/attribute]?” Frequency ratings were made on a five-point scale (1= Never to 5= All the time). Mean importance and frequency ratings for each interpersonal perception by age group are presented in Table 3.

Table 3.

Means (and Standard Deviation) for Importance and Frequency Ratings by Age Group (Study 3)

	Age Group
	Young Adults (N = 52)	Middle-Age Adults (N = 53)	Older Adults (N = 51)
Importance ratings
Anger	3.98 (1.15)	3.81 (1.11)	3.61 (1.31)
Creativity	3.40 (1.13)	3.12 (1.38)	3.18 (1.22)
Dominance	3.50 (1.21)	3.44 (1.16)	3.43 (1.30)
Easygoingness	3.39 (1.21)	3.60 (1.08)	3.20 (1.23)
Fatigue	3.48 (1.26)	3.45 (1.30)	3.70 (1.36)
Fear	3.79 (1.21)	3.57 (1.22)	3.69 (1.21)
Greed	3.21 (1.45)	3.52 (1.28)	2.91 (1.20)
Happiness	4.19 (1.03)	4.17 (1.07)	4.14 (1.15)
Kindness	4.00 (1.07)	4.00 (1.02)	4.06 (1.01)
Playfulness	3.59 (1.22)	3.54 (1.24)	3.43 (1.43)
Rudeness	3.60 (1.18)	3.42 (1.24)	3.20 (1.28)
Sadness	3.73 (1.24)	3.91 (1.02)	3.67 (1.16)
Frequency ratings
Anger	2.29 (1.00)	2.09 (0.93)	1.92 (0.91)
Creativity	2.27 (1.19)	2.17 (1.07)	2.12 (0.97)
Dominance	2.15 (1.04)	1.93 (1.00)	1.98 (0.97)
Easygoingness	2.46 (1.08)	2.40 (1.04)	2.28 (1.06)
Fatigue	2.62 (1.05)	2.40 (1.08)	2.22 (1.03)
Fear	1.79 (1.09)	1.70 (0.93)	1.77 (0.84)
Greed	1.87 (1.10)	1.72 (0.86)	1.60 (0.86)
Happiness	3.10 (1.07)	3.08 (0.85)	2.88 (1.02)
Kindness	2.87 (1.10)	2.79 (1.22)	2.75 (1.13)
Playfulness	2.77 (1.17)	2.45 (1.15)	2.12 (0.93)
Rudeness	2.29 (1.00)	2.26 (0.90)	1.94 (0.97)
Sadness	2.23 (0.99)	2.19 (1.08)	1.98 (1.05)

Results

As in Study 1, we compared the frequencies of different targets and locations across age groups using contingency table analyses and Bonferroni comparisons. Target and location frequencies are presented in Table 4 and age group comparisons are noted using subscript letters Bayes factors were calculated to quantify support for age-related differences and similarities. We then used multilevel modeling to test whether specific types of interpersonal perceptions were rated by participants as being more important or frequent and whether these differences varied between age groups.

Table 4.

Target and Location Frequencies by Age Group (Study 3)

	Age Group
	Young Adults (N = 52)	Middle-Age Adults (N = 53)	Older Adults (N = 51)
Target
Parent	59a	32b	13c
Child	90a	110a	96a
Spouse/partner	129a	98b	89b
Relative	26a	32a	53b
Friend	73a	65a	67a
Coworker	56a	42a	17b
Neighbor	9a	19a	18a
Physician	5a	3a	2a
Pharmacist	1a	2a	1a
Nurse	2a	1a	2a
Sales clerk	2a	9a	6a
Receptionist	2a	0a	1a
Wait staff	5a	6a	4a
Mail clerk	0a	0a	3a
Customer service representative	5a	12a	5a
Telemarketer	0a	2a	4a
Stranger	35a	37a	33a
Other	7a	21b	18a,b
No one	115a	144a	175b
Target total	621	635	607
Location
Perceiver’s home	197a	173a	149a
Someone Else’s home	59a,b	52b	71a
Work or school	67a	49a,b	25b
Community setting	131a	145a	121a
Other	27a	38a	37a
Location total	481	457	403

Note. Different subscript letters denote significant age group comparisons (at p < .05). To account for multiple pairwise comparisons, a Bonferroni adjustment was used.

Target

Although age group was significantly associated with target frequency, χ²(38) = 121.52, p < .001, there was very strong evidence for the null hypothesis relative to the alternative hypothesis, BF₁₀ < 0.03. As in Study 1, some familiar targets (e.g., relatives) were more frequently reported by older adults than by younger adults whereas other familiar targets (e.g., spouse/partner) were more frequently reported by younger adults than by older adults. Moreover, there were no significant age differences in the frequency of friend targets. Middle-age adults often fell between the two groups and generally did not vary from the other groups.

Location

Although age group was significantly associated with location frequency, χ²(8) = 26.35, p < .001, there was very strong evidence for the null hypothesis relative to the alternative hypothesis, BF₁₀ < 0.03. In contrast to Study 1, younger and older adults did not vary significantly in the reported frequencies of perceptions made in their own home, someone else’s home, or the community setting.

Type

Multilevel models were conducted to predict age group differences in participants’ importance and frequency ratings for different types of interpersonal perceptions. Fully unconditional models were first conducted in SAS Version 9.4; results revealed that 23% of the variance in importance ratings was between-people (τ₀₀ = 0.35, z = 6.77, p < .001) and 77% was within-people (σ² = 1.17, z = 28.53, p < .001). Similarly, 32% of the variance in frequency ratings was between-people (τ₀₀ = 0.38, z = 7.46, p < .001) and 68% of the variance was within-people (σ² = 0.81, z = 29.27, p < .001). These results indicated sufficient within- and between-person variability for subsequent models.

The high correlation between quality and timing categorizations (i.e., all emotion perceptions were also categorized as state perceptions; see Table 2) resulted in a negative covariance matrix; we thus omitted the state/trait categorizations from all models to ensure model convergence and estimation. This action is further justified by our lack of hypotheses regarding age differences in the perception of states vs. traits.

We conducted two intercepts- and slopes-as outcomes models (Raudenbush & Bryk, 2002) with randomly varying slopes predicting importance and frequency ratings with quality categorization (γ₁₀) and valence categorization (γ₂₀) at Level 1, age group (γ₀₁) at Level 2, and the cross-level interaction terms between quality categorization and age group (γ₁₁) and between valence categorization and age group (γ₂₁). Models were conducted using Bayesian estimation in Mplus version 7.4 (Muthén & Muthén, 1998–2015). Results are presented in Table 5.

Table 5.

Unstandardized Coefficients, Posterior Standard Deviations, and 95% Credibility Intervals for Importance and Frequency Ratings Multilevel Models (Study 3)

	Importance Ratings β0			Frequency Ratings β0
Fixed Effects	B	SD	95% CI	B	SD	95% CI
Intercept γ00	3.18***	0.14	2.94 to 3.46	1.60***	0.14	1.36 to 1.88
Quality γ10	0.65***	0.12	0.40 to 0.88	0.49***	0.11	0.27 to 0.68
Valence γ20	0.39***	0.12	0.13 to 0.61	1.02***	0.12	0.79 to 1.24
Age group γ01	−0.16	0.12	−0.39 to 0.06	−0.22*	0.12	−0.47 to 0.02
Quality × age γ11	0.08	0.11	−0.13 to 0.30	0.12	0.10	−0.08 to 0.31
Valence × age γ21	0.05	0.10	−0.15 to 0.25	0.00	0.09	−0.16 to 0.18
Random effects
Between τ00	0.32***	0.07	0.20 to 0.47	0.35***	0.06	0.24 to 0.47
Quality τ11	0.13***	0.06	0.03 to 0.25	0.07***	0.04	0.02 to 0.15
Valence τ22	0.19***	0.08	0.06 to 0.36	0.27***	0.07	0.15 to 0.40
Within σ00	0.95***	0.06	0.84 to 1.05	0.56***	0.03	0.50 to 0.63

Note. Models were fit using Bayesian estimation. Coefficients were tested using a one-tailed p-value based on the posterior distribution. These p-values reflect the probability of detecting the inverse effect: the proportion of the posterior distribution that is below or above zero for positive and negative coefficients, respectively. Credibility intervals that do not contain zero denote significant effects following a traditional two-tailed significance test. For more information on Bayesian estimation in multilevel modelling, see Muthén (2010).

*p < .05. ***p < .001.

There were no significant age group differences in participants’ ratings of importance or frequency. Across participants, emotional perceptions were rated higher in importance and frequency than were social perceptions. Similarly, positive perceptions were rated higher in importance and frequency than negative perceptions. Contrary to our predictions, these main effects of quality and valence were not further qualified by age group: importance and frequency ratings for different types of perceptions did not vary between age groups.

Discussion

As in Study 1, younger and older adults’ everyday interpersonal perceptions frequently involved familiar other targets, although the frequency of specific familiar targets varied with age such that other relatives were more frequently reported by older adults than younger adults and spouses/partners were more frequently reported by younger adults than by older adults. Similarly, everyday interpersonal perceptions were reported as occurring most frequently in participants’ own homes but this did not vary significantly across age groups in Study 3. Indeed, in comparison to Study 1, most age group frequency comparisons in Study 3 were non-significant, and Bayes factors were smaller than 0.03 indicating very strong evidence for the null hypothesis relative to the alternative hypothesis. Together, these results highlight both age-related differences and similarities in the reported targets and locations surrounding everyday interpersonal perceptions.

With regard to the types of interpersonal perceptions made in everyday life, there were no significant age group differences. Across ages, emotional perceptions were rated as more important and more frequent than social perceptions and positive perceptions were rated as more important and more frequent than negative perceptions. These results suggest potential age-related similarities in the types of interpersonal perceptions made in everyday life, challenging theoretical perspectives that specify increased focus on positive emotions in old age.

General Discussion

The current research sought to identify and compare the social ecology surrounding young, middle-age, and older adults’ everyday interpersonal perceptions. In contrast to theories which suggest increased prioritization on emotions—and particularly positive emotions—in old age (e.g., Carstensen et al., 1999), we found that younger and older adults did not vary in the types of judgments they made about others in everyday life. That is, across age groups, interpersonal perceptions were primarily concerned with judging other’s emotions and/or positive qualities or attributes. Thus, positive emotional perceptions may be equally relevant for younger and older adults. Moreover, across age groups, we found that familiar targets and familiar locations were most frequently reported, although the specific familiar target and location sometimes varied with age (with some familiar targets and locations more frequently reported by younger adults than older adults). These results challenge theoretical perspectives that suggest age-related increases in social selectivity and familiarity with age (Antonucci & Akiyama, 1987; Carstensen, 1992). Collectively, these findings suggest that the ecological features surrounding everyday interpersonal perception do not reflect age-related shifts in emotional goals or social selectivity, and that there are overall age-related similarities in the who, what, and where surrounding everyday interpersonal perception. Below, we discuss the implications of these findings for research on interpersonal perception and aging.

It appears that the social ecology of adults of all ages are not well-represented in current research. Most studies on interpersonal perception and aging tend to focus on negative interpersonal perceptions (e.g., the perception of anger), and age differences in perception accuracy are often strongest for these perceptions. However, our results suggest that negative interpersonal perceptions may constitute the least relevant perceptions in everyday life for adults of all ages, at least as indicated by retrospective ratings of importance and frequency. Previous studies also tend to rely on paradigms that use non-familiar actors as targets and include no location information. Because everyday interpersonal perceptions seem to involve familiar others in familiar locations, current paradigms may provide inaccurate assessments of everyday skills. Indeed, older adults’ interpersonal perception skills benefit from both target familiarity (Stanley & Isaacowitz, 2015) and inclusion of location scene (Noh & Isaacowitz, 2013, 2015), suggesting a need for these features to be better integrated into traditional laboratory paradigms. Future studies should consider the role that such features play in predicting age-related differences in interpersonal perception, for example, by incorporating these features into stimuli and examining whether doing so improves older adults’ interpersonal perception skill.

Because some specific targets and locations of older adults’ interpersonal perceptions may vary from those of younger adults, it is possible that age differences in interpersonal perception accuracy reflect differences in these social ecological features rather than differences in skill. To the extent that younger and older adults’ everyday interpersonal perceptions involve different people and occur in different locations, a one-size-fits-all measurement approach is likely to exacerbate age differences, as older adults may by unmotivated by decontextualized tasks (i.e., those that use strangers as targets and include no background scenes) that do not represent the social ecology of everyday life (Isaacowitz & Stanley, 2011; Kunzmann & Isaacowitz, 2017). Instead, interpersonal perception tasks could be more contextualized, perhaps by asking participants to think of specific people in whom they have made such judgments and the locations where these judgments took place. Such actions may result in a more accurate assessment of older adults’ interpersonal perception skill and, thus, a more nuanced understanding of interpersonal perception and aging. At the very least, our results suggest that these social ecological features should be taken into account when measuring aging individuals’ interpersonal perception in the laboratory.

Our results are not without limitations. First, these studies were conducted online, which may present data reliability and validity concerns (Buhrmester, Kwang, & Gosling, 2011). To address these concerns, we followed common recommendations and included only people living in the United States who had satisfactory approval ratings and who provided matching information on age and date of birth questions, and we excluded people (N = 35) who attempted to complete the studies more than once. Second, in Studies 1 and 3, older adults reported no one targets more frequently than did young and middle-age adults, suggesting either that the interpersonal perceptions examined were less relevant to the daily lives of older adults or that older adults had more difficulty in recalling these recent interpersonal perception events. Indeed, because our study assessed participants’ retrospective reports of recent interpersonal perceptions, our results may more accurately reflect the social ecology surrounding recalled everyday interpersonal perceptions. However, recall biases would exacerbate age-related positivity effects and such effects were not supported in our study. It is possible that participants reported interpersonal perception events that were most salient or memorable, and these may vary systematically with age (e.g., Carstensen & Turk-Charles, 1994; Fung & Carstensen, 2003); assessments of frequencies—for example, of reported targets and locations—may not reflect actual occurrence frequencies in daily life. Future work may build upon our design by utilizing interval-based experience-sampling methods, daily diaries, and/or field observations to collect descriptions of interpersonal perception in vivo. Third, although we assessed the targets and locations of 40 different interpersonal perceptions in Study 1, we assessed only 12 of these in Study 3; thus, it is possible that age differences exist for the other types of interpersonal perceptions. However, such age differences might reflect categorization differences rather than differences in experience; indeed, results from Study 2 suggest some variability in the ways in which adults of all ages categorize some interpersonal perceptions. Fourth, our target and location assessments were independent from our assessments of frequency and importance; future research may consider whether frequency and importance vary for interpersonal perceptions directed toward specific targets and locations. Finally, the quality and timing categorizations shared considerable overlap; thus, it is possible that our findings relating to quality reflect differences between states and traits rather than between emotional and social perceptions. Future work focused on perceptions that are lay categorized in only one way may help to disentangle these effects.

In sum, our findings from three adult lifespan studies suggest that there are both age-related differences and similarities in the environmental features involved in making judgements about others. The studies presented here were the first to describe the interpersonal perceptions that adults across the lifespan make in their everyday lives. In so doing, these studies contributed detailed information regarding the who, what, and where of interpersonal perception across adulthood. Such information is critical to better understand the process by which interpersonal perceptions are made across the lifespan in everyday life and whether aspects of this process can be leveraged to reduce age differences.

Funding

This work funded by a postdoctoral training grant awarded to V. L. Castro (NIA F32-AG048687).

Conflicts of Interest

The authors report no conflicts of interest.