Digital Dating: Online Profile Content of Older and Younger Adults

Abstract

Objectives:

Older adults are utilizing online dating websites in increasing numbers. Adults of different ages may share motivations for companionship and affection, but dating profiles may reveal differences in adults’ goals. Theories addressing age-related changes in motivation suggest that younger adults are likely to emphasize themselves, achievements, attractiveness, and sexuality. Older adults are likely to present themselves positively and emphasize their existing relationships and health.

Method:

We collected 4,000 dating profiles from two popular websites to examine age differences in self-presentations. We used stratified sampling to obtain a sample equally divided by gender, aged 18–95 years. We identified 12 themes in the profiles using Linguistic Inquiry and Word Count software (Pennebaker, Booth, & Francis, 2007).

Results:

Regression analyses revealed that older adults were more likely to use first-person plural pronouns (e.g., we, our) and words associated with health and positive emotions. Younger adults were more likely to use first-person singular pronouns (e.g., I, my) and words associated with work and achievement.

Discussion:

Findings suggest that younger adults enhance the “self” when seeking romantic partnership. In contrast, older adults are more positive in their profiles and focus more on connectedness and relationships to others.

Due to societal trends and deceased stigma, online dating is now one of the most common ways for adults of all ages to find a romantic partner (Rosenfeld & Thomas, 2012; Stephure, Boon, MacKinnon, & Deveau, 2009). Indeed, popular media report that more older adults are finding dates online (Watson, 2013) with one source claiming that adults aged 60 and older represent the largest growing segment of adults using online dating websites (Ellin, 2014). As such, online profiles provide a unique opportunity to examine similarities and differences in younger and older adults’ self-presentations when seeking a dating partner.

Self-presentation plays a central role in finding dating partners; potential partners use this information to decide whether to start a relationship (Derlega, Winstead, Wong, & Greenspan, 1987). Few studies have examined age differences in dating self-presentations (e.g., Alterovitz & Mendelsohn, 2009; 2013; McWilliams & Barrett, 2014) and have relied on content analysis and qualitative analysis. A systematic quantitative analysis of the language adults use in dating profiles may further illuminate motivations to date at different ages.

Adults of all ages may share certain motivations to date, including companionship and romance. However, adults of different ages have distinct goals, priorities, and motivations, which their self-presentations to potential dating partners may convey. Predictions in this study were derived from three key theories—sociocultural theories, evolutionary theory, and socioemotional selectivity theory. Collectively, these theories suggest that people of different ages may present themselves in distinct ways when seeking romantic partners.

Based on the structure of adulthood, young adults’ self-presentations may focus on life tasks aimed at gaining a foothold in adult roles, pursuing future achievements, and investing in potential procreation (and sexuality). In contrast, in late life, individuals may seek romantic partners who can become a part of their established social world. Rather than seeking new achievements, older adults’ self-presentations may be focused on financial and physical stability. This study explored how goals and motivations across the life span may be relevant to adults’ presentations in their profile content.

Self- Versus Other-Focus and Social Connections

Sociocultural theories suggest that younger and older adults may differ in their goals regarding themselves versus their connections to others (Hoppmann & Blanchard-Fields, 2010). In young adulthood, individuals engage in identity exploration and focus on self-relevant goals such as accepting responsibility for one’s self and making independent decisions (Arnett, 2000). In contrast, generativity concerns are heightened in midlife and later life, with a focus on the needs of the next generation (An & Cooney, 2006). Moreover, as they age, adults concentrate on connections to family, viewing themselves in a communal context (Fingerman, 2001). Based on a lifetime of investment, ties to family and friends are of heightened importance to older adults (Blieszner, 2006).

Individuals’ self-presentations may reflect age differences in motivations for the self and other. Indeed, the language people use when constructing their self-presentations may convey such differences. Prior research indicates that across a variety of contexts, young adults use more first-person singular pronouns, indicating a stronger focus on the self, whereas older adults use more first-person plural pronouns, indicating a focus on connections (Pennebaker & Stone, 2003; Schwartz et al., 2013).

We also examined gender differences in these patterns. Elsewhere, Groom and Pennebaker (2005) examined online dating profiles of younger adults and found that women were more likely to use pronouns (especially first-person singular) than were men. Moreover, older women also may be a key in maintaining family ties or serving as a “kinkeeper” (Brown & DeRycke, 2010), and this may extend to their presentations to potential dating partners.

Sexuality, Reproduction, and Health

Evolutionary theories posit that gender roles have evolved, such that men and women select for different qualities when choosing a partner (Buss, 1989). Studies using evolutionary theory highlight different mate selection strategies for men and women, reflecting gender differences in reproduction and parental investment. Accordingly, women prefer partners with higher status, whereas men prefer partners who are more attractive (signaling fertility and higher reproductive capability; Buss, 1989, 2003; Shackelford, Schmitt, & Buss, 2005). Although most studies examining evolutionary theory and mate selection have been limited to young adults of reproductive age (e.g., Eastwick & Finkel, 2008; Shackelford, et al., 2005), research suggests that similar to younger men, older men value physical attractiveness and sexuality in a partner more than their female counterparts (Calasanti & Kiecolt, 2007; Montenegro, 2003).

As an extension of the premise of attractiveness, adults’ self-presentations may mention their health. Based on evolutionary theory, older adults may seek to attract a mate by emphasizing their health or vitality and their ability to assist younger generations (Coupland, 2000; McWilliams & Barrett, 2014). Health also is more self-relevant for older adults, as they are more likely to suffer health problems than younger adults. Older adults also may be concerned with presenting their involvement in exercise and physical activity, so as not to appear frail. Older women in particular may value the health of a partner because they are reluctant to take on the caregiver role to an ailing partner (Carr, 2004; Dickson, Hughes, & Walker, 2005). As such, older men may be most likely to describe their health when presenting themselves to a potential romantic partner.

Work, Status, and Achievement

Achievements and status also may be key themes in self-presentations. Through early midlife, individuals focus on attaining roles necessary for a successful adult life by pursuing an education, building a career, seeking marriage, and becoming a parent (Cherlin, 2004; Furstenberg, 2010). As such, younger adults may be concerned with presenting themselves in terms of their careers and accomplishments. Conversely, the strength of achievement motivations appears to decline with age (Kanfer & Ackerman, 2004), suggesting that older adults may be less likely to present their achievements.

Self-presentations indicating status may not show age differences, however. Evolutionary theory suggests that adults are likely to highlight attributes indicative of higher mate value. For men especially, mate value is conceptualized in terms of status, reflecting an ability to provide for future offspring (Buss, 1989, 2003). A few studies have found that similar to younger women, older women emphasize the income of their desired partner (Calasanti & Kiecolt, 2007; McIntosh, Locker, Briley, Ryan, & Scott, 2011). Alterovitz and Mendelsohn (2009) used content analysis to examine 600 dating profiles from people aged 20 to more than 75 years. Older women emphasized desiring status in a partner, whereas older men offered more status-related information about themselves. Furthermore, research suggests that older women are particularly concerned with financial independence; therefore, older men may be more inclined to address money when seeking a partner in order to assuage concerns related to financial dependency (Dickson et al., 2005).

Socioemotional Motivations

We also considered socioemotional motivations in dating profiles. As people age, they prioritize emotionally meaningful goals, maximizing positive emotional experiences and minimizing unfulfilling experiences in their relationships (Charles & Carstensen, 2010). The “positivity effect” in older age is an extension of these premises, supported by findings that older adults focus on positive emotional features of experiences (Reed, Chan & Mikels, 2014). In contrast, younger adults may be more willing to accept and grapple with negative experiences in relationships (Birditt & Fingerman, 2003; Fingerman, Hay, & Birditt, 2004).

Thus, older adults may be more likely to present themselves in positive terms than younger adults. This is not to say that we expected young adults to present themselves negatively, but rather, older adults may be more likely to focus on positive aspects of what they bring to a dating relationship, particularly positive emotions.

Overview of the Current Study

The current study involved a systematic analysis of the language used by adults of different ages in the text of online dating profiles. We utilized the Linguistic Inquiry and Word Count (LIWC) software, which calculates the proportion of words fitting different linguistic categories in a sample of text (Pennebaker, Booth, & Francis, 2007). Data from online dating profiles offered an ecologically valid way to examine age differences in self-presentations.

We also considered geographic location and ethnic differences. To examine a national sample of dating profiles, we drew samples from five major metropolitan areas (encompassing urban, suburban, and rural outlying areas) from across the United States. Prior research has not addressed geographic or ethnic differences in dating motivations, but they were considered in the current study.

In summary, we expected the following patterns regarding age differences in online profiles. Compared with younger adults, older adults will use a higher proportion of words in the following categories: first-person plural pronouns, family, friends, health, and positive emotion. Compared with younger adults, older adults will use a lower proportion of words in the following categories: first-person singular pronouns, attractiveness, sexuality, work, achievement, money, and negative emotion. We did not have strong age-by-gender predictions, but considered possible gender differences in each of these patterns.

Method

Source of Online Dating Profiles

The sample of dating profiles was drawn from two major dating websites. We identified these websites using search engines (e.g., Google, Bing, Yahoo, Ask.com) with the key words “online dating” as well as reports from Experian Hitwise (a consumer behavior firm) and Google Zeitgeist (which provides most frequent search queries in a given year). Selection criteria limited websites to the United States and excluded websites that catered to a “niche” audience (i.e., older adults, sexual minorities, religious denomination, extramarital affairs, “speed dating,” “hookups,” or relationships of an exclusively sexual nature). We also limited the study to dating websites that allow users to search for potential partners (rather than assigning a limited array of partners; e.g., eHarmony.com, Chemistry.com). After exclusions, two popular websites remained.

There was no charge for creating a profile on either website, but one of the websites charged to connect with a potential dating partner. Users completed an optional free response section (i.e., “About Me” or “In My Own Words”) in which they wrote anything they chose. The instructions to create the free response section differed among the websites. The first website instructed users to write a short description of who they are and what they are looking for, whereas the second website informed users that the free response description would constitute a “first impression” for potential partners. In this study, the number of words in the free response ranged from 30 to 1,256 (M = 146.18, SD = 128.40). We did not collect profiles that contained fewer than 30 words; 220 potential profiles from a random sampling (described in Participants) were excluded due to responses with fewer than 30 words.

Participants

The study included 4,000 profiles, 2,000 sampled from each of the online dating websites using random quota sampling without replacement. Within each website, we collected 1,000 profiles from heterosexual males and 1,000 profiles from heterosexual females. Users search for profiles via geographic location, age, and gender filters.

To ensure a geographic dispersion of profiles, we selected equal numbers of profiles from five major metropolitan areas including urban, suburban, and rural areas: Los Angeles, Denver, Chicago, Atlanta, and New York City. We randomly selected zip codes from each of the five areas to search for profiles. Within each zip code, for each gender, we then randomly selected profiles among four age groups: early young adulthood (aged 18–29 years; n = 1,000; M = 25.28, SD = 3.17), late young adulthood/early midlife (aged 30–49 years; n = 1,000; M = 36.63, SD = 5.61), late midlife (aged 50–64 years; n = 1,000; M = 55.02, SD = 3.99), and older adults (aged more than 65 years; n = 1,000; M = 69.02, SD = 4.29). We used these stratifications to assure a full age range of dating profiles in sampling. Because the older adults group could incorporate up to 30 years, we treated age as a continuous variable rather than as a grouping variable in analyses.

From each profile, we extracted: gender, age, ethnicity, and the “About Me” or “In My Own Words” free response section. To assure anonymity of profile writers, we did not obtain additional demographic information (e.g., education, religious preferences, income) that could serve as identifying information. The sampling method is illustrated in Supplementary Appendix A.

The sample ranged in age from 18 to 95 years. An independent t-test revealed no difference in mean age for women (M = 46.46, SD = 17.42) and men (M = 46.52, SD = 17.31). The breakdown of ethnicity in the sample was 70% White/Caucasian, 11% Black/African American, 7% Hispanic/Latino, 2% Asian, and 10% mixed race/other.

Profile content

We used the LIWC software to analyze the content of the profiles. This software calculates the frequency and proportions of specific categories of words within a text file. The LIWC program compares each word of a text file with an internal dictionary of more than 4,500 words assigned to word categories. This study drew on 11 established LIWC categories: first-person singular pronouns, first-person plural pronouns, friends, family, work, achievement, money, health, sexuality, positive emotion, and negative emotion. Table 1 contains example words in each of the LIWC category (for additional information regarding these codes, see LIWC, 2007).

Table 1.

Mean Percentage of Responses Fitting Each Linguistic Inquiry and Word Count (LIWC) Category by Age

LIWC category	Total sample aged 18–95 years (n = 4,000)	Young adulthood aged 18–29 years (n = 1,000)	Early midlife aged 30–49 years (n = 1,000)	Late midlife aged 50–64 years (n = 1,000)	Late life aged 65 and older (n = 1,000)
First-person plural (we, us, our)	0.34 (0.78)	0.19 (0.54)	0.33 (0.77)	0.41 (0.80)	0.44 (0.92)
Family (son, husband, aunt)	0.57 (1.01)	0.51 (0.95)	0.61 (1.03)	0.50 (0.92)	0.65 (1.13)
Friends (buddy, pal, neighbor)	0.62 (0.97)	0.51(0.90)	0.64 (1.02)	0.62 (0.92)	0.69 (1.00)
Health (ache, doctor, exercise)	0.91 (1.14)	0.72 (1.05)	0.87 (1.09)	1.02 (1.20)	1.03 (1.18)
Positive emotion (love, sweet, nice)	10.44 (4.72)	9.09 (4.34)	10.13 (4.60)	11.26 (4.87)	11.30 (4.69)
First-person singular (I, me, mine)	9.01 (3.64)	10.55 (3.44)	9.27 (3.44)	8.39 (3.47)	7.82 (3.63)
Work (job, majors, boss)	1.87 (1.90)	2.15 (2.08)	1.80 (1.83)	1.62 (1.70)	1.89 (1.94)
Achievement (earn, hero, win)	1.80 (1.58)	1.94 (1.70)	1.95 (1.64)	1.76 (1.56)	1.56 (1.39)
Money (audit, cash, owe)	0.51 (0.87)	0.45 (0.81)	0.52 (0.89)	0.49 (0.85)	0.58 (0.94)
Attractiveness (hot, beautiful, cute)	0.38 (0.71)	0.38 (0.73)	0.38 (0.75)	0.39 (0.69)	0.36 (0.66)
Sexual (arouse, horny, sex)	1.46 (1.70)	1.55 (1.70)	1.42 (1.62)	1.51 (1.79)	1.37 (1.70)
Negative emotion (hurt, ugly, nasty)	0.81 (1.13)	1.07 (1.30)	0.91 (1.19)	0.69 (1.02)	0.59 (0.94)

We also created a category of words for attractiveness not available in established LIWC categories. We followed procedures for construction of LIWC categories (Tausczik & Pennebaker, 2010) by generating a comprehensive list of words from dictionaries, thesauruses, questionnaires from previous research, and words generated by the research team. Then, we selected 25 words most representative of attractiveness based on appearance in thesauruses and participant responses (e.g., cute, good-looking, handsome, hot). The attractiveness category was almost entirely distinct from the sexual category, with only one overlapping word (sexy). Examples for the attractiveness category are also found in Table 1; for the complete list of words in the attractiveness category, see Supplementary Table 1.

Analytic Strategy

We first examined descriptive statistics for the proportions of words people of different ages used in their dating profiles. We also generated illustrative figures portraying the most common words.

We then turned to hypothesis testing using ordinary least squares regression. The outcome variables in this study were the proportion of words fitting each of the 12 categories in the LIWC analyses. The LIWC categories were all positively skewed due to the number of zero values (i.e., participant did not use any words in the category). We ran each analysis with a square-root transformation (used to address non-normality in prior studies utilizing the LIWC; DeAndrea, Shaw & Levine, 2010; Hirsh & Peterson, 2009). The pattern of findings was similar after applying the transformations. For ease of interpretation, findings are presented using the untransformed LIWC category data. The independent variable was age, treated as a continuous variable. We also included gender.

Initially, we ran the regressions including the Age × Gender interaction term. One significant interaction was found in the category of positive emotion, such that women had higher mean proportions of positive emotion words than men at all ages, with women showing a slightly steeper linear increase with age than men. Thus, we did not include the interaction term for Age × Gender in the models reported here.

We examined potential differences by website, geographic region, and ethnicity using t-tests and analysis of variance (ANOVA) for the LIWC category percentages. For the two websites, six of the twelve t-tests were significant in the following categories: first-person singular [t(3998) = −5.61, p < .001], sexual [t(3998) = −3.90, p < .001], work [t(3998) = −6.31, p < .001], money [t(3998) = −2.42, p < .001], attractiveness (t(3998) = −0.93, p = .001) and positive emotion [t(3998) = 8.921, p < .001]. These differences likely reflect disparities in instructions provided by the websites and the fact that one website charges daters to contact a potential romantic partner and the other allows contacts for free. Due to the number of significant t-tests, we adjusted for the effect of website in our analyses by treating it as a dummy coded covariate.

Regarding ethnicity, 6 of 12 ANOVAs were significant, in the following LIWC categories: first-person singular F(4, 3995) = 4.87, p = .001; family F(4, 3995) = 2.76, p = .026; negative emotion F(4, 3995) = 5.25, p < .001; health F(4, 3995) = 3.37, p = .009; sexual F(4, 3995) = 4.39, p = .002; and work F(4, 3995) = 2.66, p = .031. (see Supplementary Table 2 for means, standard deviations, and contrasts between ethnic groups). Contrasts revealed significant differences between White and all other ethnic groups in four of the six significant ANOVAs. For this reason, we included ethnicity as a dummy-coded covariate in analyses (0 = White, 1 = All other ethnic groups).

Of the 12 ANOVA tests related to geographic region, only two were significant (family and positive emotion). Because the differences were not theoretically meaningful, we did not consider geographic region in subsequent analyses.

Results

Descriptive Statistics and Illustrations of Commonly Used Words

Frequency of word use is evident in descriptive statistics (see Table 1) and via word-clouds. The word-cloud technique illustrates the most commonly used words across the entire sample and in each of the age groups. The word-cloud program automatically excludes certain words, including articles (a, and, the) and prepositions (to, with, on). The remaining content words are scaled in size relative to their frequency, creating an intuitive portrait of the most prevalent content words across the sample (Wordle, 2014).

Figure 1 shows the 20 most common content words used in the entire sample. As can be seen, the most frequently used words were love (appearing in 67% of profiles), like (appearing in 62% of profiles), looking (appearing in 55% of profiles), and someone (appearing in 50% of profiles). Thus, the most common words were similar across age groups.

Figure 1.

Twenty most common content words across the entire sample.

Figure 2 shows the next 30 most common content words in the youngest and oldest age groups. By removing the first 20 common content words across the sample, we illustrate heterogeneity in the dating profiles. In the next 30 words for the youngest age group, high percentage words included get (36% of profiles in the youngest age group), go (33% of profiles in the youngest age group), and work (28% of profiles in the youngest age group). In contrast, the oldest age group had higher percentages of words such as travel (31% of profiles in the oldest age group), great (24% of profiles in the oldest age group), and relationship (19% of profiles in the oldest age group).

Figure 2.

Next 30 most common words in the youngest and oldest age groups (after subtracting the 20 most common words from Figure 1).

Hypothesis Testing of Age Differences in Language in Dating Profiles

To test hypotheses, the percentage of words from the dating profile that fit each LIWC category served as the dependent variables in regressions. We examined age and gender as independent variables as well as adjusting for website and ethnicity.

Hypothesis 1: Older age will be associated with a higher percentage of words in the following categories: first-person plural pronouns, family, friends, health, and positive emotion.

Findings largely supported Hypothesis 1 (see Table 2). Four of the five regressions revealed a significant main effect for age, such that as the age of the profile writer increased, the percentage of words in the category increased in the following categories: first-person plural, friends, health, and positive emotion. We found no significant age effect for the proportion of words in the family category.

Table 2.

Regression Analysis Predicting Percentage of Words in Linguistic Inquiry and Word Count (LIWC) Categories (Hypothesis 1)

	LIWC category: first-person plural		LIWC category: family		LIWC category: friends		LIWC category: health		LIWC category: positive emotion
	B	SE	B	SE	B	SE	B	SE	B	SE
Variables
Intercept	0.33***	0.02	0.72***	0.03	0.76***	0.03	0.97***	0.03	11.68***	0.13
Age	0.01***	0.00	0.00	0.00	0.00***	0.00	0.01***	0.00	0.05***	0.00
Gendera	0.06*	0.02	−0.27***	0.03	−0.19***	0.03	−0.11**	0.04	−1.42***	0.14
Covariates
Websiteb	−0.06*	0.02	0.02	0.03	−0.08**	0.03	−0.09*	0.04	−1.33***	0.14
Ethnicityc	0.05	0.03	−0.07*	0.04	−0.03	0.03	0.13**	0.04	0.45*	0.16
R 2	.02		.02		.02		.02		.08
F	19.19***		20.30***		15.80***		18.23***		84.47***

^aGender: 0 (female) and 1 (male). ^bWebsite: The two websites were dictomously coded as 1 and 0. ^cEthnicity: 0 (White) and 1 (Ethnic or racial minority).

*p < .05. **p < .01. ***p < .001.

Hypothesis 2: Younger age will be associated with a higher percentage of words in the following categories: first-person singular pronouns, work, achievement, money, attractiveness, sexuality, and negative emotion.

We found mixed support for Hypothesis 2 (see Table 3). Four of the seven regressions revealed a pattern consistent with hypotheses, such that as the age of the profile writer increased, the percentage of words in the category decreased. Younger adults showed higher percentages of words in the first-person singular, work, achievement, and negative emotion categories. The model for the category of money showed a significant main effect of age in the opposite direction of predictions, such that as age increased, so did the percentage of words in the money category. The models for attractiveness and sexuality categories did not show significant effects of age.

Table 3.

Regression Analysis Predicting Percentage of Words in Linguistic Inquiry and Word Count (LIWC) Categories (Hypothesis 2)

	LIWC category: first-person singular		LIWC category: work		LIWC category: achievement		LIWC category: money		LIWC category: attractiveness		LIWC category: sexual		LIWC category: negative emotion
	B	SE	B	SE	B	SE	B	SE	B	SE	B	SE	B	SE
Variables
Intercept	8.96***	0.10	1.58***	0.06	1.85***	0.05	0.45***	0.03	0.35***	0.02	1.58***	0.05	0.80***	0.03
Age	−0.06***	0.00	−0.01***	0.00	−0.01***	0.00	0.00*	0.00	0.00	0.00	−0.00	0.00	−0.01***	0.00
Gendera	−0.54***	0.11	0.14*	0.06	0.06	0.05	0.02	0.03	0.02	0.02	−0.52***	0.05	0.06	0.04
Covariates
Websiteb	0.65***	0.11	0.38***	0.06	−0.16**	0.05	0.70*	0.03	0.02	0.02	0.21***	0.05	−0.09*	0.04
Ethnicityc	−0.03	0.12	0.08	0.07	0.03	0.06	0.40	0.03	0.03	0.03	0.13*	0.06	0.10*	0.04
R 2	.10		.02		.01		.00		.00		.03		.03
F	105.59***		15.74***		14.12***		4.15*		0.70		30.68***		32.91***

^aGender: 0 (female) and 1 (male). ^bWebsite: The two websites were dictomously coded as 1 and 0. ^cEthnicity: 0 (White) and 1 (Ethnic or racial minority).

*p < .05. **p < .01. ***p < .001.

Regressions also revealed significant gender differences in the proportion of words in relevant LIWC categories. For example, women had a higher percentage of words in the first-person singular category, whereas men had a higher percentage of words in the first-person plural category. Men had higher proportions of words in the work category. Women had higher proportions of words in the categories of friends, family, health, sexuality, and positive emotion. No significant gender differences were found in the categories of achievement, money, attractiveness, or negative emotion.

Discussion

Older adults are increasingly involved in dating. Recent years have seen a proliferation of romantic relationships formed via the Internet, and older adults are increasingly turning online to find romantic partners. The ubiquity of dating websites as a means to find a relationship provides scholars with a unique opportunity to examine dating strategies and motivations in the context in which they actually occur. This study is the largest examination of age differences in dating profiles to date; we collected profile text from 4,000 adults across the United States.

The few studies comparing older and younger adults’ dating profiles using qualitative techniques have pointed out age differences in dating profiles (Alterovitz & Mendelsohn, 2009; 2013; McWilliams & Barrett, 2014). This study revealed similarities in the most widely used words in the dating profiles, with a focus on affection and affiliation (love, like) and companionship. Of course, this inherent homogeneity among dating profiles may reflect task demands that are highly scripted and constrained by the structure of the websites (Ellison, Heino, & Gibbs, 2006). Nonetheless, the questions were open ended, and findings suggest that adults of all ages may share predilections in seeking affection in romantic partnerships.

Age Differences in Dating Profiles

Though the current study revealed similarities in profile content across ages, these systematic analyses revealed age differences consistent with predictions across a variety of key content areas. Self-presentations reflected goals in a dating context and more broadly reflected the motivations of individuals at different life stages. Profiles revealed differences in how younger and older adults approach finding a mate.

Self- versus other-focus and social connections

Findings generally supported the hypotheses and were consistent with prior research on sociocultural motivations. Younger adults are focused on establishing themselves and their identities in an adult world (Arnett, 2000), and those goals translated to self-focused self-presentations. Younger adults’ higher percentages of words in the categories of work and achievement were consistent with an emphasis on self-presentation strategies related to personal goals and status.

With regard to a focus on self versus others, older adults may feel more connected to existing relationships and the needs of others (Blieszner, 2006). Findings were consistent with prior research that revealed that older adults use fewer self-references (Pennebaker & Stone, 2003; Schwartz et al., 2013). Usage of pronouns is commonly used to assess relational focus in writing (Tausczik & Pennebaker, 2010), and findings in this study regarding first-person plural pronouns were consistent with a shift toward a higher relational focus in late life.

Sexuality, reproduction and health

As in prior studies utilizing evolutionary theory, both younger and older adults mentioned their sexuality and physical attractiveness (Alterovitz & Mendelsohn, 2009; McWilliams and Barrett 2014). Research suggests that older adults who have romantic partners are sexually active (Waite, Laumann, Das, & Schumm, 2009), and findings from the current study suggest that older adults who seek dating partners are interested in physical attraction and sexuality. Gender differences in evolutionary motivations across life stages also were evident, with older and younger men mentioning their work, whereas older and younger women focusing more on sexuality.

Work, status and achievement

Findings regarding achievements and status only partially supported the hypotheses. Surprisingly, older adults were more likely to mention money in their profiles than younger adults. Some research suggests that older women are particularly concerned with the income of potential dating partners, to avoid entering a relationship that becomes a financial strain (William et al., 2011) or losing their financial independence (Calasanti & Kiecolt, 2007). Older adults were also more likely to mention health, which likely reflects the greater relevance of health to the identities of older adults (Levy, 2009).

Socioemotional motivations

Reflecting broader changes in socioemotional motivations (Charles & Carstensen, 2010), older adults were more likely to mention positive emotion in their profiles, whereas younger adults showed greater proportions of negative emotion words than older adults. These findings highlight the predominance of positive emotion in older adulthood and are consistent with dozens of studies regarding the “positivity effect” in late life (e.g., Carstensen & Mikels, 2005; Mather & Carstensen, 2005; Reed et al., 2014).

Finally, gender differences revealed that women are more likely to focus on the self as well as themes related to positive emotion, friends, family, health, and sex. Men focused more on others and occupation, as evidenced by higher percentage of words in the first-person plural and work categories. These findings support the broader literature on gender differences in language and self-presentation, such that women are more likely to use positive emotion words and self-references (Newman, Groom, Handelman, & Pennebaker, 2008; Schwartz et al., 2013).

Limitations and Future Directions

The current study was limited by a dearth of information regarding the profile writers; we could not examine differences in dating profiles by education level, occupation, religion, or prior marital status. For example, the youngest profile writers may be in the process of completing their education and have less income than the older profile writers. Furthermore, we were unable to identify how prior marriage may alter dating self-presentations. Previous work using data from the internet has revealed that it is possible to reconstruct identities with even a limited number of background variables (Parry, 2011; Zimmer, 2010). Indeed, the data were obtained from public websites, but without participant consent. As such, ethical use of ecologically valid data constrained the number of background characteristics included in this study of age differences. Future research may consider potential associations involving additional variables by utilizing surveys or experimental designs and obtaining informed consent from the participants.

Differences between the two major websites suggest that distinct types of people may choose different dating websites. Yet, studies of online dating profiles typically sample from only a single website (e.g., Alterovitz & Mendelsohn, 2009, 2013; Ellison et al., 2006; Groom & Pennebaker, 2005). The websites provided different instructions for the open-ended sections and had different fee structures. Future research should consider sampling from a greater variety of websites to learn more about motivations for dating partners.

Furthermore, the literature has rarely addressed ethnic differences in dating motivations or behaviors. Thus, we had no a priori hypotheses regarding ethnic differences in our sample. Future research may benefit from a closer examination of ethnic differences in dating motivations and behaviors.

Finally, future research might examine whether the content of online dating profiles predicts outcomes, such as being contacted for dates, number of dates, or eventual relationship satisfaction or stability. Some research suggests that specific components of profiles may play a role in dating success. Individuals are more likely to respond to initial messages in an online dating context if the messages contain fewer self-references (Schöndienst & Dang-Xuan, 2011). Additionally, greater positive emotionality appears to predict more favorable evaluations of online dating profiles (Rosen, Cheever, Cummings, & Felt, 2008). This is encouraging news for older adults, as the current study revealed a general picture of older adults as more positive and less self-focused in their dating profiles than younger adults.

In summary, similarities and differences are evident in the self-presentation strategies older and younger adults use when seeking new relationships in an online context. As the ubiquity of seeking dates online increases, older adults are utilizing this context as a means to find romantic partners. Adults of all ages may share the desires of affiliation and love when seeking a partner. However, subtle differences in these online self-presentations may reflect the changing motivations that accompany aging more broadly.