Enhancing Self-Efficacy During Community Service: Factors Influencing AmeriCorps’ Members’ Change in Self-Efficacy

Abstract

Our study uses a social cognitive theory lens to examine the development of AmeriCorps’ members’ context-specific self-efficacy (self-efficacy to perform community service). The analysis uses panel survey data from 189 Ohio State AmeriCorps members collected at the beginning and end of their service terms in 2017–2018 or 2018–2019. Using a random-effects model, the results indicate that generally self-efficacy increased from beginning to end of the service term. Perceptions of having performance accomplishments, having positive role models, and receiving useful feedback significantly increased self-efficacy to perform community service, while experiencing service stress diminished self-efficacy development.

Keywords

volunteers self-efficacy social cognitive theory random-effects models

Introduction

AmeriCorps programs provide us with a setting to understand the development of volunteers’ self-efficacy in a community-service context. AmeriCorps is a network of national service programs established through the National and Community Service Trust Act of 1993 that recruits “members” for terms of community service, the typical term being 1 year (Corporation of National & Community Service [CNCS], 2019a). In addition to addressing community needs, programs contribute to members’ professional and personal development. More than 75,000 members each year serve in AmeriCorps, with most performing service through state and local nonprofit organization hosts representing a range of missions (CNCS, 2019b). Members earn a small stipend and, in some cases, obtain an education award at the end of their service, and thus can be viewed as “stipended volunteers” (Tschirhart et al., 2001).

Our focus is on changes to individuals’ context-specific self-efficacy with the context being community service, more specifically, service through AmeriCorps state programs in Ohio. Context-specific self-efficacy is the conviction that one can successfully execute the specific behaviors required to produce one’s desired outcomes within the given context (Bandura, 1977; Stajkovic & Luthans, 1998). In contrast, general self-efficacy, sometimes referred to as generalized self-efficacy, is the overall belief that one can be successful. Insights from examining self-efficacy development of members in AmeriCorps state programs in Ohio lead to implications for studying other types of volunteers and practical guidance for community-service programs. If, as Luthans (2002) argued, and this study demonstrates, organizational practices affect context-specific self-efficacy, then knowledge about these effects may usefully guide volunteer management practice.

Benefits of Self-Efficacy Related to Volunteering in the Academic Literature

There is limited attention on both context-specific self-efficacy and general self-efficacy by volunteer management scholars. What exists suggests there is value in further investigating self-efficacy development. Throughout this article, when we refer to self-efficacy, we are referring to context-specific self-efficacy related to community-service volunteering unless otherwise noted.

Individuals with high-community-service-related self-efficacy may be the easiest for volunteer managers to recruit and retain (Harp et al., 2017; Haski-Leventhal et al., 2018; Ohmer, 2007; Wang & Wu, 2014). Harp et al. (2017) report that volunteers with a high level of context-specific self-efficacy for performing community service tend to seek out, proactively engage, and sustain themselves in challenging community-service activities. They believe they can make a meaningful difference in a community and look for opportunities to do so. Even more specifically, individuals with high self-efficacy related to environmental conservation tasks are more likely to have intentions to volunteer in ecological restoration projects than individuals with lower self-efficacy for this context (Strzelecka et al., 2018).

Self-efficacy may be an important mechanism through which volunteers actively involve, enjoy, and engage in their volunteering activities (M. Kim et al., 2007). Self-efficacy also may predict the level of initiative volunteers take to learn, teach, and engage in self-reflection (Ma & Tschirhart, 2018). Brown et al. (2012) note that volunteering is significantly related to individuals’ subjective well-being through the mediating effect of self-efficacy. They suggest that volunteers with higher self-efficacy levels may experience an enhanced perception of well-being.

Volunteering may result in greater willingness to perform certain tasks if it increases the context-specific self-efficacy related to them. Individuals like to feel that they are increasing their capacity to perform effectively. There is psychologically focused research on benefits of volunteering to individuals, such as enhanced self-esteem, guilt reduction, less loneliness, and improvement of subjective quality of life (Potočnik & Sonnentag, 2013) along with enhanced career-related competencies, new knowledge, and expanded social networks (Clary et al., 1996). We add to this literature by calling out increased self-efficacy as a potential benefit. Building context-specific self-efficacy may have an iterative effect, with each round increasing desire to pursue difficult tasks that can lead to even higher self-efficacy. Our findings suggest practical implications to help volunteer program managers balance organizational goals of effectiveness with volunteer goals of self-development, a challenge noted by Brudney and Meijs (2013).

Hypotheses and Controls

We believe social cognitive theory, with its focus on self-efficacy, is underutilized in volunteer management. “Efficacy beliefs are the foundation of human agency” (Bandura, 2001, p. 10). In social cognitive theory, individuals’ behaviors interact with their surrounding environments, with each reciprocally determining the other. Individuals are the products and producers (i.e., agents) of their environments (Bandura, 2001). If individuals do not believe they have any influence or control over what happens in a given context, they have little incentive to take any actions to change their situation and tend to quit when facing challenges (Bandura, 2001). Self-efficacy is, therefore, considered an essential component of individuals’ psychological capital, along with optimism, hope, and resilience (Dawkins et al., 2015).

Scholars note little effect to positive correlations between volunteering and self-efficacy (Krause et al., 1992; Ohmer, 2007), but explanations of this variance are limited. For richer explanations, we turn to Bandura’s (1977) theorizing of four primary sources for the development of context-specific self-efficacy: performance accomplishments, vicarious (observational) experiences, verbal persuasions, and physiological and psychological states. These sources of self-efficacy have been studied in various contexts, such as service learning, education, business, and military settings (Axtell & Parker, 2003; Eden & Kinnar, 1991; Freudenberg et al., 2011; van Dinther et al., 2011), but not in national service settings. While individuals’ general self-efficacy is, correctly, treated by scholars as a stable trait, context-specific self-efficacy is better treated as malleable (Bandura, 1997; Luthans, 2002; Ohmer, 2007; Primavera, 1999). If one conceptualizes general self-efficacy as a resource available to organizations, managers may be encouraged to recruit highly efficacious volunteers (Lindenmeier, 2008). Seeing context-specific self-efficacy as dynamic suggests a different strategy of volunteer managers’ fostering context-specific self-efficacy growth by influencing volunteer experiences and outcomes, no matter a volunteer’s initial level of it.

We extend previous scholarship by improving causal inferences and hypothesizing mechanisms of context-specific self-efficacy development. The use of individual-level panel datasets enables our analyses to move from correlations, as found in past studies, to causal inferences. Our longitudinal study examines positive and negative influences on the development of community-service self-efficacy, filling gaps in prior research. Ohmer (2007), using cross-sectional data, reports that through serving disadvantaged neighborhoods, volunteers improve their sense of control over personal and community decisions and gain service competency, but he is unable to provide causal evidence for these claims. Meyer et al. (2019) did not find that volunteering affects students’ general self-efficacy development. Individuals’ general self-efficacy is inherently, and by definition, not likely to yield perceptible changes within a short period of volunteering. As Bandura (2006) suggests, “The ‘one measure fits all’ approach usually has limited explanatory and predictive value because most of the items in an all-purpose test may have little or no relevance to the domain of functioning” (p. 307). Addressing this limitation, Meyer et al. (2019) suggest that future research should use context-specific self-efficacy measures as our study does.

We draw from Bandura’s (1977) work along with the findings from a range of scholarship in nonvolunteering contexts to operationalize four factors that may affect self-efficacy development in a volunteering context: sense of performance accomplishment, positive role models, useful performance feedback, and stress in the performance context. Bandura (1977) emphasizes enactive mastery experience as the most critical and durable source of self-efficacy. Individuals’ self-appraisals of what they can do in the future are based on their perception of their performance accomplishments in the past, especially when they attribute their successful experiences to their own ability and efforts (Bandura, 1977). Empirical findings in a range of contexts reinforce the importance of sense of performance accomplishment to self-efficacy development (e.g., van Dinther et al., 2011). Having firsthand and successful experiences create long-lasting self-confidence and perceptions of control (Bandura, 1977). In our study context, Ohio AmeriCorps may encourage self-efficacy development through programs in which volunteers (“members” in the AmeriCorps lingo) accumulate successful service experiences in a community. Therefore, we propose the following hypotheses:

Hypothesis 1: Member community-service self-efficacy level increases through their participation in their AmeriCorps program.

Hypothesis 2: The more members have an increasing sense of performance accomplishment while in their service program, the greater their increase in community-service self-efficacy.

Individuals form judgments of their abilities by observing, relating to and emulating role models (Bandura, 1997). In an organizational context, supervisors may become reference points for employees and thus contribute to the employees’ self-efficacy development (Bandura, 1997). In a military setting, an Israel Defense Forces study reveals that young soldiers gain self-efficacy by having their senior officers tell them, “I was in your shoes” (Eden & Kinnar, 1991). Other studies show that self-efficacy develops by observing senior professionals (e.g., Freudenberg et al., 2011; Subramaniam & Freudenberg, 2007; Tucker & McCarthy, 2001; van Dinther et al., 2011). In our study context, all Ohio State AmeriCorps programs provide program coordinators, but there is variation in the extent to which members perceive them and others in the program as role models. Given that individuals use social comparisons and can persuade themselves that they can improve their own performance if someone like them has improved and achieved desired outcomes, we propose the following hypothesis:

Hypothesis 3: The more members experience positive role models in the course of their service program, the greater their increase in community-service self-efficacy.

Verbal persuasion through others’ performance feedback is another mechanism for developing context-specific self-efficacy (Bandura, 1977; van Dinther et al., 2011). Performance feedback from others may be a weak source of self-efficacy because it is not a direct experience of success or failure. Also, personal assessment of one’s performance as a success or failure can conflict with what others say. If not perceived as being accurate or useful, a volunteer may dismiss the feedback. In addition, the duration of verbal persuasion is usually short (Bandura, 1997). Still, positive feedback from others is shown to be effective in improving self-efficacy of counselors, students, and employees (Daniels & Larson, 2001; Sanders et al., 2016; Schunk & Swartz, 1993). Because of its readiness to implement as a management tool, feedback has been widely applied to boost self-efficacy. This leads to the next hypothesis:

Hypothesis 4: The more members receive useful performance feedback from others during the course of their service program, the greater their increase in community-service self-efficacy.

Physical and emotional stress may hamper individuals’ self-efficacy by invoking feelings of vulnerability and jeopardizing their self-appraisals of their capabilities (Bandura, 1997). Work stress is negatively associated with self-efficacy across nine different cultures (Perrewé et al., 2002). Previous research found that self-efficacy mediates between work stress and burnout (Yu et al., 2015). L. E. Kim and Burić (2019) claim that teachers’ burnout decreases their self-efficacy. Van Dinther et al. (2011) link stress in the learning environment to less self-efficacy development in students. Long service hours, modest compensation, and inadequate training during service are among the factors that contribute to AmeriCorps’ members’ service stress (Tschirhart & Ma, 2019). This leads to our last hypothesis:

Hypothesis 5: The more members report increased service-related stress over the course of their service program, the less likely they are to have an increase in community-service self-efficacy.

Including demographic differences in a study of self-efficacy is helpful when the context is a national service program like AmeriCorps. As a reviewer for this article suggested, personal past experiences and cultural capital such as having a “privileged” background differ among members of service programs and can result in different experiences and stressors. For example, some AmeriCorps members struggle with having a stipend that leaves them in poverty (Ceresola, 2015, 2018). Due to their demographics, members may experience biases held by or toward service recipients. Members may find that others in their identity circle disagree with their choice to serve in a national service program, like some Arabs in the Israeli service program (Yanay-Ventura et al., 2020). Also, previous research contends that women and ethnic minorities have relatively lower levels of self-efficacy (Block, 1983; Gurin et al., 1978). Self-efficacy may be associated with individuals’ age, although the effect is unclear (Trouillet et al., 2009). Education may influence sense of mastery and control, and therefore improve self-efficacy (Mirowsky & Ross, 1983). Gender, ethnicity, age, and education also may affect one’s cultural capital and others’ feedback, and thus might ultimately affect self-efficacy. Due to these possibilities, we include gender, ethnicity, age, and education as control variables in the analysis.

Other controls in our analysis relate to the AmeriCorps service context. Ohio State AmeriCorps programs encompass three service areas: 12 education-related programs are “improving learning and strengthening schools,” four economic development programs are “building skills and increasing impact,” and seven health-related programs are “promoting health, wellness, and environmental sustainability” (ServeOhio, 2019). The hours of members range from full-time basis (1,700 hr) to minimum-time basis (300 hr) (ServeOhio, 2019). Members may serve another term in the same or a different program. In our data, we have 2 years of responses. Members’ program focus, time commitment, returning status, and program year might affect their self-efficacy or be correlated with the error term. Therefore, these factors are control variables.

Method

Data Collection

As university researchers, we conducted this study with the cooperation of ServeOhio, which oversees state AmeriCorps programs in Ohio. More detail on the study administration is available in reports provided to ServeOhio (Tschirhart & Ma, 2019, 2020). The study has university IRB (Institutional Review Board) approval. Data collection is through online Qualtrics surveys. Members got a link to the survey at entry to their program (within 3 months of term start) and at exit (within 3 months of term end). Four surveys are used: 2017–2018 entry, 2017–2018 exit, 2018–2019 entry, and 2018–2019 exit. ServeOhio sent a link to the survey via email to 762 (year 2017–2018) and 852 (year 2018–2019) members, a total of 1,614 members. While the survey distribution used a ServeOhio emailed link and noted ServeOhio endorsement, the survey had an Ohio State University logo and only the university researchers had access to the data. The cover note promised anonymity to respondents. Participation was voluntary. The survey completion time counted as part of member’s logged service time. Members had access to a computer to use to complete the survey at their service site. Also, the survey was mobile friendly so it could be answered using a cell phone. Members who wanted a hard copy of the online survey received one. Therefore, we do not see serious limitations from use of an online survey. The surveys took at least 15 min to complete and included many questions not used in this particular study. The survey length likely affected response rates. ServeOhio staff approved the questions. The questions for influences on self-efficacy are from published research, and the demographics and program-related questions are from AmeriCorps.

For this study, we use a small subset of the data we collected because we are using a panel approach requiring matching of entry to exit surveys and only 189 respondents gave us the extra information to do this match and responded to all needed questions in both the exit and entry surveys (11.7% response rate for this study, compared with 35%–46% for each individual survey). The respondents who participated in the panel study by providing extra identification do not significantly differ in their demographics from the full sample of our respondents. Response rates for each survey are in Table 1.

Table 1.

Response Rates.

Response rates	Entry survey	Exit survey	Matched surveys in each year by same members’ names	Response rate for our subsample
2017–2018 members	301 responses 301/762 = 39.50%	350 responses 350/762 = 45.93%	81 responses 81/762= 10.63%	189/1,614 = 11.71%
2018–2019 members	297 responses 297/852 = 34.86%	351 responses 351/852 = 41.20%	108 responses 108/852=12.68%	189/1,614 = 11.71%

While we wish the response rate for the matched responses was higher, we still believe the sample is adequate. There are reasons for our limited ability to match entry and exit responses within our larger sample. Some individuals dropped out before an exit survey for their cohort was available or they entered their cohort after the promotional push for the entry survey and so missed one of the two surveys needed for matching. It is likely that some respondents filled out both surveys but did not provide their names for both. In retrospect, we could have emphasized and provided more detail on the value of providing a name, which may have increased our sample for this study.

Measures

See the appendix for summary statistics and comparison of our sample with the national sample and our full sample from which the panel data are drawn. All the variables directly related to the hypotheses use the same scale: 1 = no extent, 2 = little extent, 3 = some extent, 4 = much extent, and 5 = great extent.

Community-service self-efficacy

When self-efficacy is context-specific, it requires a customized measure (Bandura, 2006). Following Bandura’s guidance, community-service self-efficacy at entry and at exit are measured by adding responses to four questions which ask the extent to which members agree with the following statements: “I have significant influence over what happens in my service program”; “I have a great deal of control over what happens in my service program”; “I am self-assured about my capabilities to perform my service activities”; and “I am confident about my ability to do my service activities.” The four items yield a factor score with Cronbach’s alpha of .80, indicating adequate internal consistency. Community-service self-efficacy at each time point is the total of the four responses. The minimum score is 4, occurring with a response of “1 = no extent” for all four questions; the maximum score is 20, for a response of “5 = great extent” for all four questions. Each respondent’s entry and exit self-efficacy scores are in the analysis to capture change from start to end of the program.

Performance accomplishment extent

Few respondents report “no extent” so that response is combined with “little extent” resulting in four rather than five categories. We transformed the categorical variable into four dummy variables. Each is coded 1 if the response is in that category, 0 otherwise. At exit, the question is as follows: “I helped the community served by this program.” The entry question is as follows: “I wanted to help the community served by this program.” The two questions do not match because, at entry, the respondents have little information on to what extent they are helping the community, but they can report how much intention they have to help. (The “wanted to help” question is in a long list of potential intentions.) Before individuals perform a behavior, they first form a proximal goal or intention to perform it based partially on a prediction of their likely performance success (Bandura, 1977). Therefore, the intention/goal before the task performance is a proxy for starting level of context-specific self-efficacy. Another analysis option is to treat performance accomplishment as a control variable measured only at exit. We chose to maintain the longitudinal data structure for this article’s analysis by using the behavior goal question for the entry data and the matching behavior performance accomplishment as the exit data. When using the alternative method with only the exit score in the analysis, performance accomplishment is significantly related to self-efficacy development at the p < .001 level. This alternative analysis is available from the authors.

Positive role models extent

We use four dummy variables from one question: “I have positive role models in my AmeriCorps service environment.” No extent, which had few responses, is combined with little extent. Each category forms a dummy variable coded as 1 if the response belongs to that category, 0 otherwise.

Useful performance feedback extent

Again, we use four dummy variables from one question: “I am getting useful feedback on my performance as an AmeriCorps member.” Given that only a few members report no extent, we combine no and little extent into one variable. Each category forms a dummy variable coded as 1 if the response belongs to that category, 0 otherwise.

Service stress extent

Service stress is captured with four dummy variables using the question: “My AmeriCorps service activities are stressful.” Given few responses of no extent, we combine it with little extent. Each category forms a dummy variable coded as 1 if response belongs to that category, 0 otherwise.

Demographic controls: Age, gender, ethnicity, and education level

Age ranges from 18 to 76 with a mean of 30.62 and standard deviation of 13.28. Gender is 1 for women and 0 if the respondent indicates being men or belonging to another category; in this study, we have four respondents reporting transgender or unwilling to disclose gender. White/Caucasian is 1, otherwise 0 for other race/ethnicities and combinations. Education level is 1 if have master’s or PhD degree, 0 if have no more than an undergraduate degree. Only six respondents are without a college degree. Our sample is older, has higher education attainment, and is slightly Whiter compared with the 2018 national baseline, but is similar to our full sample.

Program controls: Program focus, returning status, time commitment, and program year

Program focus is measured based on ServeOhio classification (ServeOhio, 2019): community education, health, economic development, or environment. Respectively, 59%, 16.4%, 24%, and 6% of respondents are in each program focus. Returning status is 1 if returned for an additional term, otherwise 0. Most respondents, 87.3%, are in the first year of service, and a minority, 12.7%, are in their same program for a second year. Time commitment is 1 if full-time (serving a total of 1,700 hr). For respondents serving less than 1,700 hr, the code is 0. Most, 80.95%, are full-time, 19.05% less than full-time. Program year captures if respondent is in the 2017–2018 cohort (coded as 1) or 2018–2019 cohort (coded as 0). The sample is 57.1% for 2018–2019 and 42.9% for 2017–2018.

Time in service

This variable captures whether the data are from an entry or exit survey. Entry survey responses have a code of 0 and exit responses have a 1. Thus, half of each member’s responses are 0 and half are 1.

Empirical Strategy and Analysis

If members’ service experience has a positive impact on their self-efficacy, they should have a higher level of self-efficacy at program exit compared with program entry. Panel data enable us to estimate the within- and between-member self-efficacy changes over a service year. Two types of regression models are available for panel data: fixed-effects and random-effects (Gujarati, 2015). We use the random-effects model so we can incorporate member demographics (Wooldridge, 2016) and because it is more economical and efficient in parameters estimated (Gujarati, 2015). Hausman (1978) developed a statistical test to discern if a random-effects model can be used without significantly biasing the results. For this study, the Hausman test is nonsignificant (p = .14 > .05) so a random-effects model yields statistically similar results as a fixed-effects model. We can provide readers with fixed-effects model results, Hausman test result, Breusch–Pagan Lagrange multiplier test result which concludes that random-effects models are more appropriate in our case than ordinary least squares (OLS) models, and model equations.

Results

As expected at program exit, everything else constant and, on average, community-service self-efficacy increases from the starting self-efficacy score (0.933 increase in points on average equivalent to a 30.5% standard deviation increase). This finding supports Hypothesis 1. Doing community service increases community-service self-efficacy for most individuals in our sample. Figure 1 provides a picture of the change in community-service self-efficacy for each respondent. The first chart shows the members with no change or a decrease in community-service self-efficacy. The lines represent decreased self-efficacy. The second chart shows members with an increased self-efficacy score from entry to exit, the lines representing the increase. They are more lines showing increases than decreases. The X-axis represents member ID and Y-axis represents community-service self-efficacy. Dots indicate self-efficacy at entry and triangles at exit. No change in self-efficacy is in the first graph as a dot inside a circle inside a triangle.

Figure 1.

Volunteers’ change in self-efficacy from entry to exit. SE = self-efficacy.

Table 2 presents results for the random-effects model using panel data. You can see the significance level for each variable in the last column labeled p > |z|. The coefficient indicates whether the relationship of the variable to change in community-service self-efficacy is positive or negative, just like an OLS regression coefficient would.

Table 2.

Effects on Change in Community-Service Self-Efficacy by a Random-Effects Model.

Predictors (changes from entry to exit for time variant variables)		Coefficient	SE	z	p >\|z\|
Time in service (entry)		Referent
Time in service (exit)		0.933***	0.215	4.34	.000
Accomplishment (no + little extent)		Referent
Accomplishment (some extent)		0.839	0.916	0.92	.360
Accomplishment (much extent)		1.489*	0.875	1.70	.089
Accomplishment (great extent)		2.368***	0.869	2.73	.006
Role model (no + little extent)		Referent
Role model (some extent)		0.690	0.793	0.87	.384
Role model (much extent)		1.591**	0.725	2.20	.028
Role model (great extent)		2.629***	0.748	3.51	.000
Feedback (no + little extent)		Referent
Feedback (some extent)		0.559	0.457	1.22	.221
Feedback (much extent)		1.247***	0.431	2.89	.004
Feedback (great extent)		1.762***	0.505	3.49	.000
Service stress (no + little extent)		Referent
Service stress (some extent)		−0.402	0.339	−1.19	.236
Service stress (much extent)		−0.957***	0.338	−2.83	.005
Service stress (great extent)		−2.223***	0.427	−5.19	.000
Age		0.015	0.017	0.88	.378
Ethnicity (White)		−0.082	0.306	0.27	.789
Gender (women)		−0.600*	0.339	−1.77	.077
Education level (graduate)		1.185***	0.358	3.31	.001
Program focus (education)		Referent
Program focus (health)		0.651*	0.356	1.83	.068
Program focus (economic)		0.144	0.422	0.34	.733
Returning status (returned member)		0.332	0.450	0.74	.460
Time commitment (full-time)		−1.081**	0.422	−2.56	.010
Panel year (2018–2019)		−0.046	0.884	−0.15	.884
R² (within-member)	.2686	Wald χ²		255.67
R² (between-member)	.5273	Prob > χ²		.0000
R² (overall)	.4560
corr (u_i, X) = 0 (assumed)		Rho		0.286
Sigma_u	1.233	Number of observations		378
Sigma_e	1.946	Number of members		189

p < .1. **p < .05. ***p < .01.

Overall, we find support for our hypotheses on influences (“sources” using Bandura’s term) on community-service self-efficacy development. Briefly, higher level jumps up in performance accomplishment, having positive role models, and getting useful performance feedback positively influence community-service-self efficacy development, while higher level jumps up in service-related stress negatively affects it. Although not explicitly hypothesized, it is consistent with the results that a decrease from entry to exit in the extent of performance accomplishment, role models, and feedback, on average, leads to a decrease in community-service self-efficacy. What we see is a positive relationship, so more of an increase in the source leads to more self-efficacy, more decrease of the source to less self-efficacy. However, our support for our hypotheses is limited in that we cannot significantly predict increase in community-service self-efficacy for those reporting no to little extent of performance accomplishment, role models, feedback, and stress at entry, and also report some extent at exit. This jump up at the lowest levels is insignificant. We discuss the results in more depth below.

For Hypothesis 2, we see that members reporting “no to little extent” of performance accomplishment at entry and moving to “some extent” at exit do not significantly gain in their community-service self-efficacy from entry to exit. However, in general, members who changed their response from any category lower than “much extent” to a category of “much extent,” or from any category lower than “great extent” to “great extent” of service accomplishment significantly increase community-service self-efficacy, by 1.489 and 2.368 points, respectively, equivalent to 48.68% and 77.41% standard deviation increases, compared with members that did not increase their response categories.

Support for Hypothesis 3 is similar. Compared with members reporting “no to little extent” of having positive role models at the beginning of their services and who did not change their response category at exit, members who changed from “no to little extent” to “some extent” do not significantly gain in their community-service self-efficacy from entry to exit. However, on average, members who changed their response related to extent of role models from any category lower than “much extent” at entry to “much extent” at exit, or from any category lower than “great extent” at entry to “great extent” at exit significantly increase service self-efficacy, by 1.591 and 2.629 points, respectively, equivalent to 52% and 85.94% standard deviation increases. Particularly at the high end of the scale, the greater the extent that members report more extent of positive role models at exit than at entry, the more the increase in self-efficacy from entry to exit.

Results for Hypothesis 4 are similar. Compared with members reporting “no to little extent” of receiving useful performance feedback at entry and did not change their responses at exit, members who changed to reporting “some extent” at exit do not significantly gain in their community-service self-efficacy. However, on average, members who changed their response on useful feedback from any category lower than “much extent” to “much extent,” and any category lower than “great extent” to “great extent” significantly increase service self-efficacy from entry to exit, by 1.247 and 1.762 points, respectively, equivalent to 40.76% and 57.6% standard deviation increases. Any increase to the level of much or great extent of positive feedback from entry to exit is significantly linked to increase in self-efficacy.

Hypothesis 5 follows the same pattern. Compared with members reporting “no to little extent” of service stress or who did not report increased stress, members reporting “some extent” of stress at exit do not significantly decrease their community-service self-efficacy from what they reported at entry to what they report at exit. However, on average, members who changed their response about stress from any category lower than “much extent” at entry to “much extent” at exit, and any category lower than “great extent” at entry to “great extent” at exit show a significant decrease in self-efficacy from entry report to exit report, by 0.957 and 2.223 points, respectively, equivalent to 31.28% and 72.67% standard deviation decreases. In general, high levels of service stress increasing from entry to exit negatively affect self-efficacy.

Gender, education level, program focus, and time commitment significantly relate to community-service self-efficacy change. Everything else equal, women increase self-efficacy on average 0.6 points less than males. Members with a graduate degree have on average 1.185 points higher self-efficacy at exit than members with a college degree or less. Members in health-related programs gain 0.651 points more self-efficacy compared with members in education-related programs. Full-time members have a 1.081-point decrease in their self-efficacy on average at the end of service, compared with part-time members, which deserves further study on the optimal hours for community-service self-efficacy development. On the contrary, age, being White, being a returning member, and service year are insignificant.

Theoretically, our results are consistent with social cognitive theory and Bandura’s approach to self-efficacy. Still, there may be useful improvements to our analysis, such as addressing possible interaction or moderating effects. We did examine Bandura’s (1997) proposition that feedback may be a moderator of performance accomplishments in addition to being an independent contributing source to self-efficacy. Our analysis, available upon request, did not reveal this dual role. Also, as a check on the theorized relationship, discussed earlier, between context-specific self-efficacy and intention to perform the context-relevant specific task, we correlated the self-efficacy score at exit with the intention to perform similar service at exit and found a positive and significant relationship (Pearson’s r = .163, p = .03, N = 189). Intention is measured with a yes/no question, with 48.9% of the sample saying yes and 51.1% saying no will not perform similar service in the future.

Discussion and Implications

Managers and others should not assume that context-specific self-efficacy remains stable over time or only increases. Our findings suggest that managers of volunteer programs wishing to improve volunteers’ community-service self-efficacy should facilitate performance accomplishment, provide role models, and give useful performance feedback. Reducing service-related stress experienced by volunteers may reduce decreases in self-efficacy. Thus, volunteer managers could simultaneously support organizational goals of effectiveness and volunteer goals of self-efficacy development. As volunteers’ context-specific self-efficacy grows with useful feedback, positive role models, and actual successes, they are more likely to increase their belief in their abilities and seek to perform difficult tasks they are mastering, rather than avoiding them. As long as the tasks help the accomplishment of the organization’s goals and mission, a managers’ investment in supporting the volunteers’ growth in context-specific self-efficacy is likely to be worthwhile.

Positive role models may have the largest impact on members’ community-service self-efficacy development, as they do in this study, followed by performance accomplishment, useful feedback, and stress levels. Therefore, we suggest encouraging managers and others to see themselves as potential role models for members to emulate. In the AmeriCorps setting, more research can build understanding of who members see as their role models and why. Some role models’ performance outcomes may be more visible and salient to members than others.

Coming back to the same program may not help members gain more self-efficacy through repeating successful performance of previously learned tasks, but it likely does not decrease it. We find no significant differences in community-service self-efficacy development between first- and second-year members. Perhaps second-year members served as role models for those in their first year and performed new tasks such as orientation and training. In doing so, their own self-efficacy may have increased. The potential of conscious role modeling as a builder of self-efficacy deserves attention by scholars.

We found that full-time members experience a decrease in self-efficacy over the service term compared with part-time members. Thus, time commitment in a volunteer program deserves further study. There may be a saturation or nonlinear effect related to program duration that we cannot capture in this study. Too little or too much time performing the same tasks may be detrimental to context-specific self-efficacy development. Too much time in service may lead to burnout if volunteers address the same issues over and over again, like tutoring a student struggling with math, combating health problems systemic in a community, or continual need to clean trash from the same stream. However, our finding may have less to do with repeatedly performing the same tasks to address the same problems and more to do with something not captured in our study that significantly differentiates part-time from full-time members. For example, perhaps the decrease is due to financial stress caused by reliance on the AmeriCorps stipend for living expenses by full-timers. Part-timers have more time for paid supplemental work than full-timers.

Age and ethnicity appear irrelevant to community-service self-efficacy development. However, we did not look at whether minority/majority status within a program makes a difference to self-efficacy development. Perhaps those in the minority on these demographics within their service cohort are different in their self-efficacy development from those in the majority. While we know the general composition of our subsamples for each program, we do not know if this is similar to the full population of members for each program.

Gender and education demographic characteristics are significant predictors. In general, there are greater increases in self-efficacy for those who do not identify as women. Not surprisingly, education was a predictor of self-efficacy development, with greater increases for those who have more education. These findings warrant a deeper examination to reveal the underlying mechanisms. For example, females and those with less education may be less likely to report high levels of self-efficacy due to socialization or other systemic influences.

Type of program stands out in our study, though only at the p = .068 level. We found that health-focused AmeriCorps programs are associated with more self-efficacy development than programs with other foci. This may relate to the nature of the service activities performed, different types of accountability expectations, unmeasured feedback opportunities (such as ability to obtain a certification as a community health worker but no certification opportunities in the other program areas), or other factors. We did not examine possible interaction effects of our predictors of self-efficacy with program type, but they and other types of interactions may exist. Further study may reveal additional insights related to program differences.

There are limits to the generalizability of this study. For example, our sample is older than the national baseline, and almost all had college degrees. This diverges from the national baseline statistics for AmeriCorps programs. Future research could focus on members with a college degree or less, who likely will be younger than many of our respondents. The sample representativeness for other demographic variables (gender and ethnicity) is more in line with the AmeriCorps national survey data (CNCS Office of Research and Evaluation, 2018). Age and education may affect the types of tasks assigned to a volunteer with possible consequences for availability of role models, useful feedback, and experiences of successful performance accomplishment.

We recognize the challenges of measuring self-efficacy, like other researchers using this construct. This study may have a social desirability bias, because only a few members report having no to little extent of self-efficacy at the beginning or end of service. Certain types of individuals may be especially reluctant to admit feelings of inefficacy. Furthermore, our model captures change in self-efficacy. In our analysis, those who consistently reported low self-efficacy are treated the same as those consistently reporting high self-efficacy. We may gain additional insights by using level of self-efficacy reported at exit, rather than change in level. Also, respondents at the upper end of the self-efficacy scale at the time of the first survey cannot go any higher for the second survey. Those at the low end of the scale at entry and still there at exit may have dynamics at play that have little to do with our self-efficacy model. Given insignificant results for members in the low-end categories of our self-efficacy influencers, we should not rule out threshold effects (must meet a minimum for effect to start), but nonlinear dynamics are undiscoverable with our measurement and analysis technique. Larger samples would allow parsing out the bottom, middle, and upper ends of the response spectrum. Also, despite our measure having adequate internal consistency, it could be evaluated in comparison with a similar Community-Service Self-Efficacy Scale (Reeb et al., 1998).

In sum, our study is useful to both scholars and practitioners. We use Bandura’s social cognitive theory to find predictors of community-service self-efficacy development in the understudied volunteering context. Our results are consistent with hypotheses but only at the higher end of the scales for the theorized influences. This suggests possible threshold or nonlinear dynamics for future study. We are able to model causal effects and go beyond past treatments of community-service self-efficacy as a stable personality trait. We followed Bandura’s guidance to develop our measure of context-specific self-efficacy, and it is available for other scholars’ use. We contribute to practice by showing that volunteers’ access to positive role models, experience of performance accomplishment, receipt of useful feedback, and lack of service-related stress likely increase their community-service self-efficacy. We also have findings to support a warning to volunteer managers that community-service self-efficacy can decrease. Based on our findings, it is worse, for example, to take away high levels of useful feedback, than to have only offered medium levels of feedback but to have done so consistently through a volunteer term. With this knowledge, managers can design and implement programs to foster and not undermine self-efficacy development. Managers might then reap the reward of volunteers seeking out and mastering challenging tasks rather than avoiding them.

Footnotes

Appendix

Appendix.

Summary Statistics.

Statistics are average of two time points, N = 189	M	SD	Minimum	Maximum	Ohio State AmeriCorps Programs Full Sample	National AmeriCorps 2018 Baseline
Self-efficacy index	15.07	3.059	4	20
Time in service	0.50	0.501	0	1
Performance accomplishment
No to little extent	0.213	0.145	0	1
Some extent	0.104	0.306	0	1
Much extent	0.315	0.465	0	1
Great extent	0.560	0.497	0	1
Positive role model
No to little extent	0.0344	0.182	0	1
Some extent	0.0873	0.283	0	1
Much extent	0.376	0.485	0	1
Great extent	0.503	0.501	0	1
Useful feedback
No to little extent	0.127	0.333	0	1
Some extent	0.169	0.376	0	1
Much extent	0.434	0.496	0	1
Great extent	0.270	0.444	0	1
Service stress
No to little extent	0.347	0.477	0	1
Some extent	0.230	0.421	0	1
Much extent	0.291	0.455	0	1
Great extent	0.132	0.339	0	1
Program focus (education)	0.593	0.492	0	1
Program focus (health)	0.243	0.430	0	1
Program focus (economic)	0.164	0.371	0	1
Returning status (returned)	0.13	0.333	0	1
Time commitment (full-time)	0.81	0.393	0	1
Panel year	0.43	0.496	0	1
Age	30.62	13.284	18	76
Percentages
Age <20	2%				1%	2%
Age 20–29	72%				67%	69%
Age 30–39	11%				11%	7%
Age 40–49	3%				5%	3%
Age 50–59	4%				7%	2%
Age 60–69	7%				8%	2%
Age >70	2%				2%	1%
Ethnicity
White	68%				56%	54%
African/Black	20%				19%	22%
Other	17%				15%	24%
Gender
Female	77%				76%	75%
Male or Other	23%				24%	25%
Education level
Graduate	21%				19%	6%
Undergraduate	77%				77%	45%
High school or below	2%				4%	49%

Declaration of Conflicting Interests

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

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Yinglin Ma

Author Biographies

Yinglin Ma is a doctoral candidate at John Glenn College of Public Affairs at the Ohio State University. Her dissertation focuses on nonprofit volunteers’ self-efficacy and learning.

Mary Tschirhart is the Stephen Joel Trachtenberg Professor in Public Service and the director of the Trachtenberg School of Public Policy & Public Administration at the George Washington University. Her interests include human resource management for nonprofit organizations especially those with volunteers and members.

References

Axtell

C. M.

Parker

S. K.

(2003). Promoting role breadth self-efficacy through involvement, work redesign and training. Human Relations, 56(1), 113–131. https://doi.org/10.1177/0018726703056001452

Bandura

(1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84(2), 191–215. https://doi.org/10.1037/0033-295X.84.2.191

Bandura

(1997). Self-efficacy: The exercise of control. W. H. Freeman.

Bandura

(2001). Social cognitive theory: An agentic perspective. Annual Review of Psychology, 52(1), 1–26. https://doi.org/10.1146/annurev.psych.52.1.1

Bandura

(2006). Guide for constructing self-efficacy scales. In Pajares

Urdan

T. C.

(Eds.), Self-efficacy beliefs of adolescents (pp. 307–337). Information Age Publishing.

Block

J. H.

(1983). Differential premises arising from differential socialization of the sexes: Some conjectures. Child Development, 54(6), 1335–1354. https://doi.org/10.2307/1129799

Brown

K. M.

Hoye

Nicholson

(2012). Self-esteem, self-efficacy, and social connectedness as mediators of the relationship between volunteering and well-being. Journal of Social Service Research, 38(4), 468–483. https://doi.org/10.1080/01488376.2012.687706

Brudney

J. L.

Meijs

L. C. P. M

. (2013). Our common commons: Policies for sustaining volunteer energy. Nonprofit Policy Forum, 4(1). https://doi.org/10.1515/npf-2012-0004

Ceresola

(2015). Doing poor in AmeriCorps: How national service members deal with living below the poverty line. Qualitative Sociology Review, 11(4), 116–137.

10.

Ceresola

(2018). The influence of cultural capital on how AmeriCorps members interpret their service. VOLUNTAS: International Journal of Voluntary and Nonprofit Organizations, 29(1), 93–103. https://doi.org/10.1007/s11266-017-9858-9

11.

Clary

E. G.

Snyder

Stukas

A. A.

(1996). Volunteers’ motivations: Findings from a national survey. Nonprofit and Voluntary Sector Quarterly, 25(4), 485–505. https://doi.org/10.1177/0899764096254006

12.

Corporation of National & Community Service. (2019a). About CNCS: Legislation. https://www.nationalservice.gov/about/legislation

13.

Corporation of National & Community Service. (2019b). CNCS: AmeriCorps programs. https://www.nationalservice.gov/programs/americorps/americorps-programs

14.

Corporation of National & Community Service Office of Research and Evaluation. (2018). AmeriCorps state commissions 2015 – 2018 trend report.

15.

Daniels

J. A.

Larson

L. M.

(2001). The impact of performance feedback on counseling self-efficacy and counselor anxiety. Counselor Education and Supervision, 41(2), 120–130. https://doi.org/10.1002/j.1556-6978.2001.tb01276.x

16.

Dawkins

Martin

Scott

Sanderson

(2015). Advancing conceptualization and measurement of psychological capital as a collective construct. Human Relations, 68(6), 925–949. https://doi.org/10.1177/0018726714549645

17.

Eden

Kinnar

(1991). Modeling Galatea: Boosting self-efficacy to increase volunteering. Journal of Applied Psychology, 76(6), 770–780. https://doi.org/10.1037/0021-9010.76.6.770

18.

Freudenberg

Cameron

Brimble

(2011). The importance of self: Developing students’ self efficacy through work integrated learning. The International Journal of Learning: Annual Review, 17(10), 479–496. https://doi.org/10.18848/1447-9494/CGP/v17i10/58816

19.

Gujarati

D. N.

(2015). Econometrics by example (2nd ed.). Palgrave Macmillan.

20.

Gurin

Morrison

B. M.

(1978). Personal and ideological aspects of internal and external control. Social Psychology, 41(4), 275–296. https://doi.org/10.2307/3033581

21.

Harp

E. R.

Scherer

L. L.

Allen

J. A.

(2017). Volunteer engagement and retention: Their relationship to community service self-efficacy. Nonprofit and Voluntary Sector Quarterly, 46(2), 442–458. https://doi.org/10.1177/0899764016651335

22.

Haski-Leventhal

Meijs

L. C. P. M.

Lockstone-Binney

Holmes

Oppenheimer

(2018). Measuring volunteerability and the capacity to volunteer among non-volunteers: Implications for social policy: Measuring volunteerability and the capacity to volunteer among non-volunteers: Implications for social policy. Social Policy & Administration, 52(5), 1139–1167. https://doi.org/10.1111/spol.12342

23.

Hausman

J. A.

(1978). Specification tests in econometrics. Econometrica, 46(6), 1251–1271. https://doi.org/10.2307/1913827

24.

Kim

L. E.

Burić

(2019). Teacher self-efficacy and burnout: Determining the directions of prediction through an autoregressive cross-lagged panel model. Journal of Educational Psychology, 112(8), 1661–1676. https://doi.org/10.1037/edu0000424

25.

Kim

Chelladurai

Trail

G. T.

(2007). A model of volunteer retention in youth sport. Journal of Sport Management, 21(2), 151–171. https://doi.org/10.1123/jsm.21.2.151

26.

Krause

Herzog

A. R.

Baker

(1992). Providing support to others and well-being in later life. Journal of Gerontology, 47(5), P300–P311. https://doi.org/10.1093/geronj/47.5.P300

27.

Lindenmeier

(2008). Promoting volunteerism: Effects of self-efficacy, advertisement-induced emotional arousal, perceived costs of volunteering, and message framing. VOLUNTAS: International Journal of Voluntary and Nonprofit Organizations, 19(1), 43–65. https://doi.org/10.1007/s11266-008-9054-z

28.

Luthans

(2002). Positive organizational behavior: Developing and managing psychological strengths. Academy of Management Perspectives, 16(1), 57–72. https://doi.org/10.5465/ame.2002.6640181

29.

Tschirhart

(2018). Stipended volunteers’ self-efficacy and learning: An AmeriCorps Study [Working paper]. The Ohio State University.

30.

Meyer

Neumayr

Rameder

(2019). Students’ community service: Self-selection and the effects of participation. Nonprofit and Voluntary Sector Quarterly, 48(6), 1162–1185. https://doi.org/10.1177/0899764019848492

31.

Mirowsky

Ross

C. E.

(1983). Paranoia and the structure of powerlessness. American Sociological Review, 48(2), 228. https://doi.org/10.2307/2095107

32.

Ohmer

M. L.

(2007). Citizen participation in neighborhood organizations and its relationship to volunteers’ self- and collective efficacy and sense of community. Social Work Research, 31(2), 109–120. https://doi.org/10.1093/swr/31.2.109

33.

Perrewé

P. L.

Hochwarter

W. A.

Rossi

A. M.

Wallace

Maignan

Castro

S. L.

Ralston

D. A.

Westman

Vollmer

Tang

Wan

Van Deusen

C. A.

(2002). Are work stress relationships universal? A nine-region examination of role stressors, general self-efficacy, and burnout. Journal of International Management, 8(2), 163–187. https://doi.org/10.1016/S1075-4253(02)00052-2

34.

Potočnik

Sonnentag

(2013). A longitudinal study of well-being in older workers and retirees: The role of engaging in different types of activities. Journal of Occupational and Organizational Psychology, 86(4), 497–521. https://doi.org/10.1111/joop.12003

35.

Primavera

(1999). The unintended consequences of volunteerism: Positive outcomes for those who serve. Journal of Prevention & Intervention in the Community, 18(1–2), 125–140. https://doi.org/10.1300/J005v18n01_10

36.

Reeb

Katshyama

Sammon

Yoder

(1998). The community service self-efficacy scale: Evidence of reliability, construct validity, and pragmatic utility. Michigan Journal of Community Service Learning, 5, 48–57.

37.

Sanders

M. J.

Van Oss

McGeary

(2016). Analyzing reflections in service learning to promote personal growth and community self-efficacy. Journal of Experiential Education, 39(1), 73–88. https://doi.org/10.1177/1053825915608872

38.

Schunk

D. H.

Swartz

C. W.

(1993). Writing strategy instruction with gifted students: Effects of goals and feedback on self-efficacy and skills. Roeper Review, 15(4), 225–230. https://doi.org/10.1080/02783199309553512

39.

ServeOhio. (2019). Ohio AmeriCorps programs. https://www.serveohio.org/AmeriCorps/Ohio-AmeriCorps-Programs

40.

Stajkovic

A. D.

Luthans

(1998). Self-efficacy and work-related performance: A meta-analysis. Psychological Bulletin, 124(2), 240–261. https://doi.org/10.1037/0033-2909.124.2.240

41.

Strzelecka

Woosnam

K. M.

Nisbett

G. S.

(2018). Self-efficacy mechanism at work: The context of environmental volunteer travel. Journal of Sustainable Tourism, 26(11), 2002–2020. https://doi.org/10.1080/09669582.2018.1526297

42.

Subramaniam

Freudenberg

(2007). Preparing accounting students for success in the professional environment: Enhancing self-efficacy through a work integrated learning programme. Asia-pacific Journal of Cooperative Education, 8(1), 77–92.

43.

Trouillet

Gana

Lourel

Fort

(2009). Predictive value of age for coping: The role of self-efficacy, social support satisfaction and perceived stress. Aging & Mental Health, 13(3), 357–366. https://doi.org/10.1080/13607860802626223

44.

Tschirhart

(2019). ServeOhio AmeriCorps member survey results from 2017-2018 entry and exit surveys. The Ohio State University.

45.

Tschirhart

(2020). ServeOhio AmeriCorps member survey results from 2018-2019 entry and exit surveys. The Ohio State University.

46.

Tschirhart

Mesch

D. J.

Perry

J. L.

Miller

T. K.

Lee

(2001). Stipended volunteers: Their goals, experiences, satisfaction, and likelihood of future service. Nonprofit and Voluntary Sector Quarterly, 30(3), 422–443. https://doi.org/10.1177/0899764001303002

47.

Tucker

McCarthy

(2001). Presentation self-efficacy: Increasing communication skills through service-learning. Journal of Managerial Issues, 13(2), 227–244.

48.

van Dinther

Dochy

Segers

. (2011). Factors affecting students’ self-efficacy in higher education. Educational Research Review, 6(2), 95–108. https://doi.org/10.1016/j.edurev.2010.10.003

49.

Wang

(2014). Volunteers’ motivation, satisfaction, and management in large-scale events: An empirical test from the 2010 Shanghai World Expo. VOLUNTAS: International Journal of Voluntary and Nonprofit Organizations, 25(3), 754–771. https://doi.org/10.1007/s11266-013-9350-0

50.

Wooldridge

J. M.

(2016). Introductory econometrics: A modern approach (6th ed.). Cengage Learning.

51.

Yanay-Ventura

Issaq

Sharabi

(2020). Civic service and social class: The case of young Arab women in Israel. VOLUNTAS: International Journal of Voluntary and Nonprofit Organizations. https://doi.org/10.1007/s11266-020-00210-z

52.

Wang

Zhai

Dai

Yang

(2015). The effect of work stress on job burnout among teachers: The mediating role of self-efficacy. Social Indicators Research, 122(3), 701–708. https://doi.org/10.1007/s11205-014-0716-5