Age differences in emotional experiences associated with helping and learning at work

Abstract

Drawing on socioemotional selectivity theory and goal theories of emotion, this study examined age differences in helping and learning activities at work and the emotional correlates of such activities. We hypothesized that older workers help colleagues more than younger workers and derive greater emotional benefits from helping; and that younger workers learn more often at work and derive greater emotional benefits from learning. Frequency of employees’ (N = 365; age 18–78 years) helping, learning, and emotional experience were monitored for five days using a modified day reconstruction method. We found that older workers engaged in helping more than younger workers and reported greater positive emotions from helping. Contrary to our hypothesis, younger and older workers engaged in learning activities at similar frequencies. However, in line with our hypothesis, learning was associated with more positive emotions for younger workers. Findings suggest thoughtful consideration of how to optimize work activities and practices that promote emotional well-being of both younger and older workers.

Introduction

The number of older Americans is increasing exponentially. In 2000, 13% of individuals in the labor force were aged 55 and older, compared to 24% in 2020 (Bureau of Labor Statistics, 2021). The increasing trend is expected to continue (Hurd & Rohwedder, 2014). As workforces grow increasingly age diverse, it is important to understand how and when workplace activities and practices align with activities and preferences of older and younger workers.

Although some evidence suggests that older workers are less motivated at work, these findings are based on study of relative disinterest in skill training and novel learning (Kanfer & Ackerman, 2004). In this study, we expanded examination of work motivations to include both helping and learning at work. Prosocial activities—defined as voluntary activities intended to benefit others (Eisenberg & Miller, 1987)—as well as novel learning experiences (e.g., performing new tasks) are ubiquitous aspects of work life. Although both prosocial activities and novel learning experiences are associated with emotional experience (Holman & Wall, 2002; Snippe et al., 2018), little is known about how age moderates these associations. Aiming to fill this gap, this study examined age differences in the frequency of learning and helping and the emotional correlates of these activities. We further explored whether associations between daily work activities and experienced emotions differed among younger and older workers.

Age Differences in Motivation and Workplace Activities

Life-span development theories of emotional experience may help explain age differences in work engagement. Socioemotional selectivity theory (SST; Carstensen, 2021; Carstensen et al., 1999) maintains that goals change as individuals come to perceive the future as limited. When future time is perceived as constrained, as is typical with older age, individuals focus more on emotionally meaningful experiences and less on knowledge acquisition.

Consistent with SST, evidence suggests that engaging in prosocial activities may serve as particularly meaningful experiences for older adults (Doerwald et al., 2021). Individuals often engage in prosocial activities at work, for example, by mentoring others, providing career advice, assisting colleagues, and providing emotional support to coworkers (Organ et al., 2006). Prosocial engagement differs by age: Older adults devote more time to informal helping than other age groups (Corporation for National and Community Service, 2013). A substantial body of research also suggests that older age is associated with prosocial dispositions at work, including greater engagement in interpersonal organizational citizenship behaviors, such as helping colleagues with work tasks (Ng & Feldman, 2008) and motivations to pass on knowledge and experiences (Stamov-Roßnagel & Biemann, 2012).

Reasoning theoretically from SST, younger workers may be more interested than older workers in novel learning experiences that provide them with opportunities for growth and development at work. Relative to older workers, younger workers favor work activities that allow them to acquire knowledge and accumulate skills to help advance their careers (Colquitt et al., 2000; Kooij et al., 2011). In the workplace, learning can occur formally through training programs or development courses to meet current and future work requirements (Jacobs & Park, 2009) or through informal means such as through completing a work task, receiving feedback from colleagues, or working collectively with colleagues. The current study focuses on informal learning as these activities occur more frequently in daily life.

Age Differences in Workplace Activities and Emotional Experience

Theories that link goal attainment and emotional experience postulate that emotional states arise as a function of progress towards reaching one’s goals (Emmons & Kaiser, 1996). Specifically, positive emotions are experienced when one reaches their goals, whereas negative emotional states arise when progress towards goals are blocked or slower than desired. If so, older workers may benefit emotionally from engaging in emotionally meaningful activities such as helping colleagues whereas younger workers may benefit emotionally when engaging in knowledge-related activities such as novel learning at work.

There is evidence for age differences in the emotional benefits derived from prosocial activities in daily life. In two national studies from the United States, older compared to younger adults reported less negative emotion on days when they volunteered or provided emotional support to others (Chi et al., 2021). Prosocial contributions in the workplace also appear to benefit emotional well-being, but less attention has been paid to potential age differences. Prosocial activities such as mentoring are associated with higher job satisfaction, greater organizational commitment, and higher assessments of job performance (Ghosh & Reio, 2013). In one daily diary study of eighty workers spanning a variety of occupations, helping at work was associated with more positive momentary emotions (Conway et al., 2009). Snippe and colleagues (2018) also assessed prosocial behavior and positive emotions three times per day for 30 days among community-dwelling Dutch workers and found that prosocial behavior predicted subsequent positive emotions. Koopman and colleagues (2016) also found that daily interpersonal organizational citizenship behaviors (e.g., helping a colleague) were associated with positive emotions. Experimental findings also point to the positive impact of prosocial activities in workers: In one 4-week experimental intervention with younger and middle-aged employees, individuals who were assigned to be “Givers” and asked to engage in specific prosocial activities to their colleagues (e.g., “bring someone a beverage”, “email a thank you note”) showed significant increases in life satisfaction and job satisfaction, as well as decreases in depressive symptoms, compared to individuals who were not assigned to prosocial activities (Chancellor et al., 2018). Although these studies suggest that workplace helping is associated with emotional experience¹, they did not explore age differences. It is unclear whether workplace helping may differentially benefit younger and older workers.

The associations between workplace learning and emotional experience are less clear. Learning opportunities provide individuals with the ability to develop new skills and grow their knowledge base, which ultimately contributes positively to well-being (Holman & Wall, 2002). Zaniboni and colleagues (2013) found that task variety (i.e., performing novel and different tasks) led to less burnout (e.g., work frustration) for younger workers relative to older workers. Yet, it can also elicit negative emotions. Rausch and colleagues (2015) found that learning, particularly from feedback from others, is associated with both negative (e.g., nervousness and worry) and positive (e.g., motivated and glad) emotional states. Based on younger workers’ preferences for learning and knowledge acquisition (e.g., Schulz & Roßnagel, 2010), we may expect that younger workers have greater emotional benefits from novel learning compared to older workers. The links between age and novel learning may also occur through mechanisms such as perceived time horizons at work. In an age-diverse sample of working adults, Kooij and Zacher (2016) found that work time horizons mediate the relationship between age and attitudes towards learning and development at work. That is, due to perceiving their work futures as more limited, older workers held a less positive attitude about learning and development. Studying the day-to-day associations between novel learning with emotional experience is an important step towards understanding the mechanisms involved.

In summary, the existing body of evidence suggests that prosocial activities –in particular, helping – and learning are associated with emotional experiences, yet less is known about how these associations differ by age in workers spanning a variety of occupations.

Current Study

This study examined age differences in individuals’ contributions and activities at work. Based on SST, we hypothesized that older workers would engage in more helping (an operational definition of prosociality) and less novel learning at work compared to younger workers. Consistent with theories surrounding goal attainment and emotions (Emmons & Kaiser, 1999), we also hypothesized that engagement in work activities differentially benefit the emotional experience (operationalized as experience of positive and negative emotions) of younger and older workers. Specifically, we hypothesized that older workers experience relatively greater emotional benefits from helping than younger workers, and that younger workers experience relatively greater emotional benefits from novel learning than older workers. Following past research, we controlled for factors known to be associated with helping, learning, and emotional experience at work, including gender, race, education, and supervisory/managerial responsibilities at work (Bommer et al., 2003; Conway et al., 2009; Snippe et al., 2018).

To assess workplace activities and emotion associations in a daily life context without obstructing participants during work, we utilized the Day Reconstruction Method (Kahneman et al., 2004), an assessment strategy where participants are asked to recall their work activities and emotional experiences at the end of each day. In the present study, participants were asked to reconstruct the beginning, the middle, and the end of their workdays (operationalized as three work periods similar in length). As part of the DRM, participants responded to questions regarding what they were doing each work period and how they felt during that part of the day (work-period emotional experience). Participants then reflected on how they felt overall at work at the end of each workday (daily emotional experience). Analyses explored how engagement in helping and learning activities was related to emotional experience at two time-scales, across work-periods and across days, with expectation that the same pattern of associations would manifest at both time-scales.

Methods

Transparency and Openness

Prior to data collection, the study’s hypotheses, design, and planned sample size were preregistered on the Open Science Framework (Chi, 2023) as part of a broader project that investigated a variety of hypotheses. The present paper addresses hypotheses 1, 3, and 4 in the preregistration. Following past research indicating that the associations between workplace activities (e.g., helping) and same-day well-being are relatively small (e.g., Sonnentag & Grant, 2012), a priori power analyses calculated (using G*Power 3.1.9.7; Faul et al., 2007) with Cohen’s f² = 0.03, β = 80%, and α = 5% suggested a sample size of 368 participants (see OSF documentation). Data, research materials, and analytic code are publicly available at the Open Science Framework (Chi, 2023).

Design

Data were collected through the Prolific platform in 2021 from a nationally representative and age-stratified sample of working adults. Eligibility criteria included that participants were (a) age 18 years or older and employed full-time in the United States, (b) had worked for their current employer/company in their current position for at least 1 year (given the questions of interest on social interactions and helping at work), and (c) be in a job where they interact with other employees (e.g., colleagues or assistants). The study was approved by Stanford University’s Institutional Review Board (Study Title: Workplace Activities and Well-Being; Protocol No.: 60781).

After providing informed consent, participants completed an intake survey assessing demographics and workplace characteristics. To accommodate variation in work hours and schedules, participants were asked to indicate five workdays in the upcoming weeks that would be representative of their typical workday. Participants were asked to complete a survey at the end of each of these five workdays, reporting on their work activities and emotional experience for the workday. The daily surveys utilized the DRM by asking participants to reconstruct the workday into a beginning, middle, and end (formalized as three work periods similar in length). Participants then completed items about what they were doing during each work period, who they were with, and how they felt during that part of the day, and finally a measure about how they felt about their workday overall.

Participants

Data were provided by 527 participants. Following the preregistered plan, the analysis sample included participants who completed at least four of the five daily surveys (N = 388; 74% compliance), and excluded participants who indicated that the study period was not at all typical of their everyday life, work life, and how much they typically help or learn at work (N = 23) – as measured via items asking how much the study period was considered typical of their everyday life (0 = not at all, 4 = very much; M = 2.91), work life (M = 3.26), and how much they typically help (M = 2.97) and learn (M = 2.75) at work.

The final sample included 365 participants (59% male; 78% White, 11% African American), ages 18 and 78 years (M = 42.43, SD = 11.94), who provided valid data for an average of 14.43 (SD = 1.18, range = 12 to 15) work periods. Participants worked in a range of jobs (e.g., analyst, teacher, cashier) and industries (e.g., business, healthcare, service) in the United States. Most of the sample (76%) had completed a four-year college degree and had supervisory responsibilities at work (59%); 24% reported being in a customer facing/front line employee role (e.g., retail or banking). Overall, the sample matches the age distributions of recent U.S. labor force surveys (Bureau of Labor Statistics, 2020) and may be considered representative of that population.

Measures

Work period helping.

On each of the five daily surveys, participants were asked whether they engaged in helping behaviors during each of three work periods: “Did you help or support coworkers, supervisors, and/or junior colleagues at work? Some examples of helping include but are not limited to: providing advice or information, listening to others’ problems, and helping someone complete a work task.” Responses, yes (=1) or no (=0) indicated that participants helped or supported colleagues during 40% (SD = 27%) of work periods. Participants’ reports about the recipient(s) and type of help they provided for each work period indicated that helping was directed towards peers (61%), junior colleagues (45%), and supervisors (30%); and included providing information (69%), advice (43%), and emotional support (14%).

Work period learning activities.

At each of the daily surveys, participants were also asked whether they learned something novel during each of the three work periods: “Did you learn something new at work? Some examples of learning activities include but are not limited to: working alone or with others to develop new ideas, performing new tasks, and receiving feedback from colleagues.” Responses yes (=1) or no (=0) indicated that participants learned something novel during 22% (SD = 25%) of work periods. Participants indicated that they mostly learned through acquiring new information such as searching the internet (50%), developing new ideas or solutions to problems (39%), and performing new tasks (34%).

Work period emotional experience.

Emotional experience during each work period was assessed using 4 negatively-valenced emotion items (sad, bored, angry, anxious) and 4 positively-valenced emotion items (excited, calm, happy, curious). Responses provided on 5-point scales (0 = not at all, 1 = a little, 2 = somewhat, 3 = very, 4 = extremely) were averaged to compute one negative emotion (within-person: ω_h = .49, R_c = .48; between-person: ω_h = .74, R_KF = .98) and one positive emotion (within-person: ω_h = .59, R_c = .57; between-person: ω_h = .86, R_KF = .99) score for each work-period. On average, participants’ level of negative emotion at work was relatively low (M = 1.57, SD = 0.51; 0 to 4 scale) and level of positive emotion at work was relatively high (M = 2.94, SD = 0.77; 0 to 4 scale).

Daily helping.

Frequency of engagement in helping was calculated as the number of work periods that day that included helping activities (scores ranged from 0 to 3). On average, participants engaged in helping activities for 1.62 (SD = 0.97) work periods per day.

Daily learning activities.

Frequency of engagement in learning activities was calculated as the number of work periods that day that included learning something new (scores ranged from 0 to 3). On average, participants learned something new during 0.83 (SD = 1.01) work periods per day.

Daily emotional experience at work.

Participants’ daily emotional experience was assessed at the end of each workday using 18 emotion items adapted from Carstensen et al. (2011) that capture a broad range of positively and negatively valenced emotions experienced across the adult life span. Specifically, participants reported how much they felt 9 negatively-valenced emotions (sad, bored, angry, frustrated, anxious, embarrassed, stressed, concerned, lonely) and 9 positively-valenced emotions (excited, proud, calm, appreciative, pleasant, satisfied, happy, curious, close to others) during that workday. Responses provided on 5-point scales (0 = not at all, 4 = extremely) were averaged across the 9 relevant items to obtain daily negative emotion (within-person: ω_h = .70, R_c = .68; between-person: ω_h = .86, R_KF = .97) and daily positive emotion (within-person: ω_h = .79, R_c = .78; between-person: ω_h = .96, R_KF = .99) scores. Across days, participants reported relatively low levels of daily negative emotion (M = 1.68, SD = 0.49) and high levels of daily positive emotion (M = 2.85, SD = 0.79).

Age.

Age was assessed with a single item in the intake survey, “What is your age (in years)?” (M = 42.43, SD = 11.94).

Covariates.

Analyses included several demographic variables and work characteristics that are known to be related to emotional experiences (e.g., Conway et al., 2009; Snippe et al., 2018), including gender (female, male), race (White, non-White), and education (college graduate, non-college graduate). We included an additional control variable that was not preregistered, regarding whether participants had supervisory/managerial responsibilities at work (following evidence that employees with different levels of seniority may have different opportunities to engage in helping and learning experiences; Bommer et al., 2003). Follow-up analyses made use of an adapted version of the future time perspective scale for the work context (Zacher & Frese, 2009) in which participants were asked about occupation-related time horizons (e.g., “Many opportunities await me in my occupational future”). Responses for the 10 items were averaged to obtain a single score (Cronbach’s α = .93), with higher scores suggesting more expansive time horizons (M = 4.25, SD = 1.31).

Data Analysis

Person-level relations among the key study variables and covariates were examined using bivariate correlations. The hypotheses that older workers have higher rates of helping and lower rates of novel learning compared to younger workers were tested using person-level data in two linear regressions where average daily frequency of helping or average daily frequency of learning activities were regressed on age and the covariates (gender, race, education, and supervisory status).

The hypotheses that older workers experience relatively greater emotional benefits from helping than younger workers, and that younger workers experience relatively greater emotional benefits from novel learning than older workers were tested using multilevel models (Snijders & Bosker, 2012) that accommodated the nested structures of the data. Emotion outcomes were measured separately at both the work-period level and daily level in a 3-level structure (work-periods nested within days nested within persons). However, the 3-level variance decompositions indicated the proportion of variance located at the day level was extremely small, .001 and .002 for work-period negative and positive emotion, respectively. Thus, the work-period level emotion outcomes and the day-level emotion outcomes were each analyzed in parallel using four 2-level models. For example, when examining the association between work-period positive emotion and work-period helping, the model was specified as

$$\mathit{\text{Work period positive emotio}}{n}_{ti}={\beta }_{0i}+{\beta }_{1i}({\mathit{\text{ Work period helping}}}_{ti})+e_{ti}$$ (1)

where the positive emotion experienced by individual i during work period t, work period positive emotion_ti, was modeled as a function of a person-specific intercept, β_0i, indicating baseline level of emotion; a person-specific slope, β_1i, indicating the within-person association between helping and positive emotion; and residual error, e_ti, that is assumed to be normally distributed. Between-person differences in the person-specific coefficients were modelled as

$$\begin{gathered} {\beta }_{0i}={\gamma }_{00}+{\gamma }_{01}\left({\mathit{\text{Person average of helping}}}_{c_i}\right)+{\gamma }_{02}\left({\mathit{\text{Age}}}_{c_i}\right)+{\gamma }_{03}\left({\mathit{\text{Gender}}}_{c_i}\right)+{\gamma }_{04}\left({\mathit{\text{Race}}}_{c_i}\right)+{\gamma }_{05}\left({\mathit{\text{Education}}}_{c_i}\right)+{\gamma }_{06}\left({\mathit{\text{Supervisory Status}}}_{c_i}\right)+u_{0i} \\ {\beta }_{1i}={\gamma }_{10}+{\gamma }_{11}\left({\mathit{\text{Age}}}_{c_i}\right)+u_{1i} \end{gathered}$$ (2)

where the γ₀₀ to γ₀₉ indicate the prototypical level of emotion and how differences in that level are related to differences in person average of helping, age, gender, race, education, and supervisory status; and where γ₁₀ and γ₁₁ indicate the prototypical within-person association between helping and positive emotion, and how age moderated the within-person link between helping and positive emotion. The us indicate unexplained between-person differences in the intercept and the within-person association of helping and work-period positive emotion. Age, person average of helping, and covariates (gender, race, education, and supervisory status) were centered at the sample mean. Work-period helping was treated as a categorical variable. Models for the association between learning and emotion followed the same configuration. Because the list of emotions assessed at the end of workdays was more extensive than the subset assessed during work-periods, we reran the analyses using only those emotions that were common to both daily and work-period assessments; namely, sad, bored, angry, anxious, excited, calm, happy, and curious. In these analyses, frustrated, embarrassed, stressed, concerned, lonely, proud, appreciative, pleasant, satisfied, and close to others were omitted.

Analyses were conducted using R version 4.1.1 (R Core Team, 2021) using the base package for the correlations and linear regressions, and the lme4 package (Bates et al., 2015) with restricted maximum likelihood estimation for the multilevel models. In one set of models, the random effects were trimmed to obtain effective model convergence (with results very similar across different random effect configurations).

Results

Age Differences in Work Activities and Emotional Experiences

Bivariate correlations are presented in Table 1. Older age was associated with more helping (r = .12) and less negative emotion at work (r = −.32). Supervisors/managers reported both more novel learning (t_362.206 = 5.95, p < .001) and helping (t₃₆₃ = 3.75, p < .001) compared to non-supervisors/managers. Age had a moderate negative correlation with occupational future time perspective (r = −.45). In line with our hypothesis, older age was significantly associated with more helping at work (b = 0.003, p = .03), after controlling for gender, race, education, and supervisory status. Counter to expectations, age was not associated with amount of novel learning at work (b = −.001, p = .31), i.e., older workers reported comparable levels of learning. Results were very similar in follow-up analyses where occupational future time perspective was included as an additional potential explanatory factor.

Table 1:

Sample-Level Means, Standard Deviations, and Correlations of Central Study Variables

Variable	M	SD	1	2	3	4	5	6	7	8	9	10
1. Age	42.43	11.94
2. Gender	0.59	0.49	−.10 [−.20, .00]
3. Education	0.76	0.43	−.05 [−.15, .05]	.15** [.05, .25]
4. Race	0.78	0.41	.25** [.15, .34]	−.08 [−.19, .02]	−.08 [−.18, .02]
5. Frontline worker	0.24	0.43	−.11* [−.21, −.01]	−.05 [−.15, .06]	−.06 [−.16, .05]	−.11* [−.21, −.01]
6. Supervisor	0.59	0.49	.01 [−.09, .11]	.15** [.05, .25]	.24** [.14, .33]	−.10* [−.20, −.00]	.10* [.00, .20]
7. Helping	0.40	0.27	.12* [.01, .22]	.01 [−.10, .11]	.04 [−.06, .14]	.02 [−.09, .12]	.09 [−.02, .19]	.19** [.09, .29]
8. Novel Learning	0.22	0.25	−.05 [−.16, .05]	.11* [.01, .21]	.26** [.16, .35]	−.14** [−.23, −.03]	.04 [−.07, .14]	.28** [.18, .37]	.42** [.34, .51]
9. Work positive emotions	2.94	0.77	.01 [−.09, .11]	.13* [.03, .23]	.25** [.15, .34]	−.16** [−.26, −.06]	.06 [−.04, .16]	.26** [.16, .35]	.18** [.07, .27]	.41** [.33, .50]
10. Work negative emotions	1.57	0.51	−.32** [−.41, −.22]	.00 [−.10, .10]	.06 [−.04, .16]	−.09 [−.19, .01]	.05 [−.06, .15]	.02 [−.09, .12]	−.04 [−.14, .06]	−.01 [−.11, .09]	−.33** [−.42, −.23]
11. Occupational future time perspective	4.25	1.31	−.45** [−.53, −.36]	.10 [−.01, .20]	.10 [−.01, .20]	−.17** [−.27, −.07]	.01 [−.09, .12]	.05 [−.05, .15]	−.05 [−.15, .05]	.12* [.02, .22]	.36** [.26, .44]	−.08 [−.18, .03]

Note. N = 365. Values in square brackets indicate the 95% confidence interval for each correlation.

Gender was coded 0 = female and other gender, 1 = men. Education was coded 0 = did not complete college education, 1 = completed college education. Race was coded as 0 = non-White, 1 = White. Frontline worker was coded as 0 = not employed in a customer facing/front line employee role (e.g., retail or banking), 1 = employed in a customer facing/front line employee role. Supervisor was coded as 0 = does not have supervisory responsibilities at work, 1 = has supervisor responsibilities at work. The momentary measures of helping, learning, positive emotions at work, and negative emotions at work were averaged to obtain person-level means.

^*indicates p < .05.

^**indicates p < .01.

Work-Period Helping and Emotional Experiences

Results from the multilevel models examining whether helping contributes to emotional experience during the same work period, and whether those associations were moderated by age are shown in Table 2. Helping was associated with greater positive emotion (γ₁₀ = 0.13, p <.001) during the same work period. In line with hypotheses, age moderated the extent to which helping was associated with work-period positive emotion (γ₁₁ = 0.003, p = .04). As shown in Figure 1, older workers experienced more positive emotion during work periods in which they helped (simple slope: b = 0.16, p < .001), relative to younger workers (simple slope: b = 0.10, p < .001). As expected, helping was associated with less negative emotion (γ₁₀ = −0.03, p = .04) during the same work period, but contrary to hypotheses, age did not moderate the association between helping and negative emotion (γ₁₁ = −0.002, p = .07).

Table 2:

Results from Multilevel Models Examining Age Differences in the Association between Work-Period Helping and Emotional Experience

	Work-Period Positive Emotion	Work-Period Negative Emotion
	Parameter Estimate [95% CI]
Fixed effects
Intercept, γ00	2.891** [2.82, 2.97]	1.583** [1.53, 1.63]
Work-Period Helping, γ10	0.130** [0.10, 0.16]	−0.028* [−0.05, 0.00]
Age, γ02	0.002 [0.00, 0.01]	−0.013** [−0.02, −0.01]
Gender, γ03	0.105 [−0.05, 0.26]	−0.043 [−0 .15, 0.06]
Race, γ04	−0.254** [−0.44, −0.07]	−0.008 [−0 .13, 0.12]
Education, γ05	0.332** [0.15, 0.51]	0.058 [−0.06, 0.18]
Supervisor, γ06	0.259** [0.10, 0.42]	0.015 [−0. 09, 0.12]
Person Average of Helping, γ01	0.254 [−0.03, 0.54]	0.015 [−0. 18, 0.21]
Age X Work-Period Helping, γ11	0.003* [0.0001, 0.01]	−0.002 [−0.004, 0.0001]
Random effects, SDa
Intercept, σuo	0.70 [0.65, 0.76]	0.48 [0.44, 0.51]
Residual, σe	0.70 [0.48, 0.49]	0.63 [0.39, 0.41]
ICC	0.68	0.59

Note. 5263 work-periods nested within 365 participants. The covariates, Gender coded 0 = female and other gender, 1 = men; Education coded 0 = did not complete college education, 1 = completed college education; Race coded 0 = non-White, 1 = White; Supervisor coded 0 = does not have supervisory responsibilities at work, 1 = has supervisor responsibilities at work, were sample-mean centered. ICC reflects the proportion of variance on the person level. Person average of helping represents the proportion of work periods when a person engaged in helping over the study period.

^aRandom effect for work-period helping was trimmed to obtain effective convergence.

^*p < .05,

^**p < .01

Figure 1.

Age moderated the association between helping and work-period positive emotions such that older workers had higher increases in positive emotions during the same work period when they helped someone, relative to younger workers. For illustrative purposes, model implied simple slopes are depicted for adults at age 30 (−1 SD), 42 (mean), and 54 (+1SD) years. Age was used as a continuous variable in all models.

b = unstandardized coefficient of simple slope. SE of simple slopes are indicated in parentheses.

*** p < .001

Work-Period Learning and Emotional Experiences

As shown in Table 3, novel learning at work was associated with greater positive emotion (γ₁₀ = 0.29, p <.001) during the same work period. Contrary to hypotheses, age did not moderate the association between novel learning and positive emotion (γ₁₁ = −0.001, p = .66). Similarly, although novel learning at work was associated with less negative emotion (γ₁₀ = −0.07, p <.001) during the same work period, age did not moderate the association between learning and negative emotion (γ₁₁ = 0.001, p = .40).

Table 3:

Results from Multilevel Models Examining Age Differences in the Association between Work-Period Novel Learning and Emotional Experience

	Work-Period Positive Emotion	Work-Period Negative Emotion
	Parameter Estimate [95% CI]
Fixed effects
Intercept, γ00	2.880** [2.81, 2.95]	1.585** [1.53, 1.64]
Work-Period Learning, γ10	0.291** [0.24, 0.35]	−0.066** [−0.10, −0.03]
Age, γ02	0.004 [0.00, 0.01]	−0.014** [−0.02, −0.01]
Gender, γ03	0.061 [−0.08, 0.20]	−0.041 [−0 .14, 0.06]
Race, γ04	−0.214* [−0.39, −0.04]	−0.014 [−0 .14, 0.11]
Education, γ05	0.212* [0.04, 0.38]	0.064 [−0.06, 0.19]
Supervisor, γ06	0.188* [0.04, 0.34]	0.012 [−0. 09, 0.12]
Person Average of Learning, γ01	0.740** [0.44, 1.04]	−0.010 [−0.22, 0.20]
Age X Work-Period Learning, γ11	−0.001 [−0.01, 0.003]	0.001 [−0.002, 0.004]
Random effects, SD
Intercept, σuo	0.69 [0.63, 0.74]	0.49 [0.45, 0.53]
Work-Period Learning, σu1	0.31 [0.24, 0.36]	0.17 [0.11, 0.22]
Correlation between intercept and work-period learning, σuo, u1	−0.41 [−0.57, −0.24]	−0.56 [−0.76, −0.36]
Residual, σe	0.68 [0.46, 0.48]	0.63 [0.39, 0.40]
ICC	0.68	0.59

Note. 5262 work-periods nested within 365 participants. The covariates, Gender coded 0 = female and other gender, 1 = men; Education coded 0 = did not complete college education, 1 = completed college education; Race coded 0 = non-White, 1 = White; Supervisor coded 0 = does not have supervisory responsibilities at work, 1 = has supervisor responsibilities at work, were sample-mean centered. ICC reflects the proportion of variance on the person level. Person average of learning represents the proportion of work periods when a person engaged in novel learning over the study period.

^*p < .05,

^**p < .01

Daily Helping and Emotional Experiences

Results from the multilevel models examining whether frequency of helping on a given day was related to daily emotional experience, and whether those associations were moderated by age are shown in Table 4. Although helping during more work periods was associated with more same-day positive emotion (γ₁₀ = 0.07, p <.001), age did not moderate the association between daily helping and positive emotion (γ₁₁ = 0.001, p = .44). Contrary to our hypotheses, helping was not associated with same-day negative emotion (γ₁₀ = 0.002, p = .84), nor did age moderate the association (γ₁₁ = −0.001, p = .61). In follow-up analyses where the daily emotion composites were composed of the same limited set of items that were available at the work-period level, the findings were similar. Helping during more work periods was associated with more daily positive emotion (γ₁₀ = 0.04, p = .01) and age did not moderate the association (γ₁₁ = 0.002, p = .23). Helping was not associated with daily negative emotion (γ₁₀ = 0.002, p = .84) and age did not moderate the association (γ₁₁ = −0.0005, p = .61).

Table 4:

Results from Multilevel Models Examining Age Differences in the Association between Daily Helping and Emotional Experience

	Daily Positive Emotion	Daily Negative Emotion
	Parameter Estimate [95% CI]
Fixed effects
Intercept, γ00	2.766** [2.68, 2.85]	1.676** [1.62, 1.73]
Daily Helping, γ10	0.073** [0.04, 0.10]	0.002 [−0.02, 0.03]
Age, γ02	0.004 [0.00, 0.01]	−0.009** [−0.01, 0.00]
Gender, γ03	0.143 [−0.01, 0.30]	−0.073 [−0.17, 0.03]
Race, γ04	−0.278** [−0.47, −0.09]	−0.007 [−0.13, 0.11]
Education, γ05	0.316** [0.14, 0.50]	0.083 [−0.03, 0.20]
Supervisor, γ06	0.312** [0.15, 0.47]	0.067 [−0.03, 0.17]
Person Average of Helping, γ01	0.022 [−0.06, 0.10]	0.027 [−0.03, 0.08]
Age X Daily Helping, γ11	0.001 [−0.002, 0.003]	−0.001 [−0.002, 0.001]
Random effects, SD
Intercept, σuo	0.76 [0.69, 0.83]	0.47 [0.42, 0.52]
Daily Helping, σu1	0.11 [0.05, 0.16]	0.04 [0.003, 0.09]
Correlation between intercept and daily helping, σuo, u1	−0.53 [−0.83, −0.28]	−0.73 [−1.00, −0.15]
Residual, σe	0.63 [0.38, 0.41]	0.57 [0.31, 0.34]
ICC	0.76	0.65

Note. 1754 days nested within 365 participants. The covariates, Gender coded 0 = female and other gender, 1 = men; Education coded 0 = did not complete college education, 1 = completed college education; Race coded 0 = non-White, 1 = White; Supervisor coded 0 = does not have supervisory responsibilities at work, 1 = has supervisor responsibilities at work, were sample-mean centered. ICC reflects the proportion of variance on the person level. Person average of helping represents the person-averages of the daily frequency of helping over the study period.

^*p < .05,

^**p < .01

Daily Learning and Emotional Experiences

As shown in Table 5, engaging in novel learning during more work periods was associated with more same-day positive emotion (γ₁₀ = 0.07, p <.001). In line with hypotheses, age moderated the association between learning and daily positive emotion (γ₁₁ = −0.003, p = .02). As shown in Figure 2, learning was associated with more daily positive emotion for younger workers (simple slope: b = 0.11, p < .001), but not for older workers (simple slope: b = 0.03, p = .29). Novel learning was not associated with same-day negative emotion (γ₁₀ = 0.005, p = .71), nor did age moderate this association (γ₁₁ = 0.001, p = .31). Results from follow-up analyses where the daily emotion composite variables were calculated on the more limited set of items that were also available at the work-period level were consistent with the main analysis in that engaging in novel learning during more work periods was associated with more daily positive emotion (γ₁₀ = 0.05, p = .01) and novel learning was not associated with daily negative emotion (γ₁₀ = −0.017, p = .28). However, in this analysis, age did not moderate the association between learning and positive emotion (γ₁₁ = −0.002, p = .31).

Table 5:

Results from Multilevel Models Examining Age Differences in the Association between Daily Learning and Emotional Experience

	Daily Positive Emotion	Daily Negative Emotion
	Parameter Estimate [95% CI]
Fixed effects
Intercept, γ00	2.809** [2.73, 2.89]	1.674** [1.62, 1.73]
Daily Learning, γ10	0.069** [0.03, 0.11]	0.006 [−0.02, 0.03]
Age, γ02	0.008* [0.00, 0.01]	−0.010** [−0.01, −0.01]
Gender, γ03	0.101 [−0.05, 0.25]	−0.063 [−0.16, 0.04]
Race, γ04	−0.204* [−0.38, −0.03]	−0.015 [−0.14, 0.11]
Education, γ05	0.249* [0.07, 0.43]	0.083 [−0.04, 0.20]
Supervisor, γ06	0.295** [0.14, 0.45]	0.071 [−0.03, 0.17]
Person Average of Learning, γ01	0.131** [0.06, 0.21]	0.019 [−0.03, 0.07]
Age X Daily Learning, γ11	−0.003* [−0.01, −0.001]	0.001 [−0.001, 0.003]
Random effects, SD
Intercept, σuo	0.72 [0.66, 0.78]	0.47 [0.42, 0.50]
Daily Learning, σu1	0.13 [0.05, 0.18]	0.09 [0.01, 0.13]
Correlation between intercept and daily learning, σuo, u1	−0.68 [−1.00, −0.45]	−0.47 [−1.00, −0.17]
Residual, σe	0.63 [0.38, 0.42]	0.57 [0.31, 0.34]
ICC	0.74	0.66

Note. 1754 days nested within 365 participants. The covariates, Gender coded 0 = female and other gender, 1 = men; Education coded 0 = did not complete college education, 1 = completed college education; Race coded 0 = non-White, 1 = White; Supervisor coded 0 = does not have supervisory responsibilities at work, 1 = has supervisor responsibilities at work, were sample-mean centered. ICC reflects the proportion of variance on the person level. Person average of learning represents the person-averages of the daily frequency of novel learning over the study period.

^*p < .05,

^**p < .01

Figure 2.

Age moderated the association between learning and daily positive emotions such that younger workers experienced more positive emotions on days when they learned something novel at work. Older workers did not experience more positive emotions on days when they learned more. For illustrative purposes, model implied simple slopes are depicted for adults at ages 30 (−1 SD), 42 (mean), and 54 (+1SD) years. Age was used as a continuous variable in all models.

b = unstandardized coefficient of simple slope. SE of simple slopes are indicated in parentheses.

*** p < .001

Discussion

The results of our analysis suggest that older workers help colleagues more often than younger workers. Older workers experience more positive emotions when helping at the workperiod level. Contrary to our hypotheses, there was no evidence of age differences in how often workers engaged in learning at work, but age did moderate the association between learning and emotion at the daily level. Younger workers reported more positive emotion on days when they learned more than did older workers. According to goal theories of emotion, positive emotions are experienced when goals are approached and negative emotions are experienced when they are blocked (Emmons & Kaiser, 1996). In this context, the pattern of findings suggests that helping is more valued by older workers and learning is more valued by younger workers.

Contrary to expectations, we did not find evidence that age was associated with engagement in novel learning opportunities. One prior study of employees in a mailing company also found no differences across age groups for participation in training and development opportunities that were required components of the job (Schulz & Roßnagel, 2010). However, our findings suggest that novel learning may be more aligned with the goals of younger workers since they experienced increases in daily positive emotions from learning, whereas older workers did not. Yet, comparable frequencies of learning were reported by younger and older workers. The lack of age differences in frequency of learning activities may indicate that even though older workers express less interest in novel training at work, they are learning on the job.

It is important to note that the age moderation differed across time scales. Age did not moderate the association between learning and positive emotional experience when examined at the work period time-scale but did moderate the association when examined at the daily time-scale. One explanation is that since learning is a process of sense-making and acquiring of new knowledge (Eraut, 2011), learning experiences take time to influence emotions. Thus, younger workers may derive greater benefits from learning after being able to reflect on their experiences as relevant to their goals and progress at work. Theoretically, this aligns with postulates set by SST as younger adults are more motivated to invest in exploratory goals that pay off in the future, whereas older adults engage in emotionally satisfying goals that have more immediate benefits (Carstensen, 2021). Note that age moderation was not observed in analyses based only the more limited set of emotions that were common to both work-periods and daily assessments. It is unclear whether this points to the importance of the additional emotions assessed only at the day level (proud, appreciative, pleasant, satisfied, close to others) or whether it reflects inconsistency as a function of time scale. Furthermore, we found that older workers reported greater increases in positive emotions from helping during the same work period, but not at the end of the day, suggesting that the immediate payoffs of this activity matters when examining age differences. An emerging literature suggests that older adults are more prosocial than younger adults (Carstensen & Chi, 2021; Chi et al., 2021; Cutler et al., 2021; Raposo et al., 2020). Findings from the current study extend such evidence to the work context.

Limitations and Future Directions

Although the present study focused on age differences in workplace activities, we cannot know whether differences reflect cohort or age. Previous studies suggest that attitudes towards learning new technology differ across cohorts (Morris & Venkatesh, 2000). Future studies should also examine how perceived time horizons at work influence preferences for helping and learning. Recent evidence suggests that experimental constraints on work time horizons eliminate age differences in helping (Shavit et al., 2022), yet findings from the present study suggest that age differences in helping remain even after controlling for time horizons.

There are a number of limitations of the study. Though typical of daily diary research, small effect sizes demand replication (e.g., Chi et al., 2021; Sonnentag & Grant, 2012) and larger data sets will be needed to more closely examine how time horizons at work shape age differences in workplace activities. The measures may also provide a limited view of how emotional experience unfolds at work. In particular, the somewhat low within-person reliability of the work-period items suggests that the items capture different aspects of emotional change. Future research might probe if and how specific emotions, valence, and arousal fluctuate in relation to work activities. Another limitation is the relatively young sample. Also, even though the age range of our sample is consistent with the U.S. labor force (Bureau of Labor Statistics (2020), there were fewer older than younger participants in our study sample. As more people work longer, it will be important to examine the generalizability of the findings to samples that skew older. Nonetheless, findings about the correlates between workplace activities and emotional experience are an important first step in identifying mechanisms that link helping and learning with emotional experience.

Of course, many questions remain. Far more research is needed to understand age differences in the range of work contexts, from high to low complexity jobs and those that involve physical labor to those that are more clerical in nature. Participants in the study were mostly college graduates and, relatedly, included a higher percentage of supervisors, though reported findings persisted after controlling for supervisory status in the analyses. It will also be important to isolate potential age differences as a function of occupation and experience in the stated positions, which could influence opportunities to help and learn. Future research is needed to isolate causal mechanisms. Quasi-experimental interventions involving random assignment to helping or learning (e.g., Chancellor et al., 2018) will more clearly distinguish persistent age differences from related benefits of engaging in learning and/or prosociality. Furthermore, our findings suggest that age differences in helping and learning with emotional experience differed across time scales, although we did not have specific hypotheses about these differences. Future studies will be necessary to examine the reliability of these findings and assess emotion outcomes at different time scales that capture younger and older workers’ investment in exploratory vs. emotionally satisfying goals.

We expect that activities that build on age differences in motivation, such as mentoring programs that utilize older employees and diverse work offerings for younger employees, will improve workforce performance. Because age-diverse work teams are becoming increasingly common, it will be important to consider potential complementarity of preferences and motivation of team members. Fasbender and colleagues (2021) found that younger workers high on development striving (e.g., importance of novel learning) were more likely to receive knowledge from older colleagues. Such findings highlight the importance of understanding motivation in both younger and older workers.

Overall, the findings demonstrate that helping and learning are differentially associated with emotional experience in younger and older workers spanning a broad range of occupations. As employers begin to plan for increasingly age-diverse workforces, we urge thoughtful consideration of work activities and practices that will enhance the complementary of motives held by younger and older workers.

Public Significance Statement.

This study suggests that older workers help colleagues more often than younger workers and experience more positive emotions while doing so. Although we observed no age differences in frequencies of learning activities, younger workers experienced more positive emotions on days when they learned more at work. These findings on age differences in motivation can inform the creation and management of work teams, especially the potential complementarity of age diversified work teams.