Content uploaded by David M. Hansen
Author content
All content in this area was uploaded by David M. Hansen on Apr 06, 2018
Content may be subject to copyright.
1
Youth Program Adult Leader Supports and Development of Adolescents’ Capacity for Agency
David Hansen
University of Kansas
E. Whitney Moore
Wayne State University
Nadia Jessop
University of Kansas
Citation: Hansen, D. M., Moore, E. W., Jessop, N. (in press). Youth program adult leader
supports and development of adolescents’ capacity for agency. Journal of Research on
Adolescence.
2
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
Author Note
David M. Hansen, Educational Psychology, University of Kansas; E. Whitney Moore,
Department of Kinesiology, Health & Sport Studies, Wayne State University; Nadia Jessop,
Educational Psychology, University of Kansas.
This research was funded in part by a W.T. Grant Foundation grant. The authors’ express
gratitude to Jonathan Templin and Lesa R. Hoffman for lending their expert advice on the
analyses.
Address correspondence to David M. Hansen, Educational Psychology, University of Kansas,
1122 West Campus Road, JRP 642 Lawrence, KS 66045. E-mail: dhansen1@ku.edu
3
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
Youth Program Adult Leader Supports and Development of Adolescents’ Capacity for Agency
Adult life and work places a premium on adolescents who develop capacities for
exercising agency—capacities to self-direct/regulate one’s effort, attention, and behavior over
time to achieve goals. Increasingly, well-paying jobs require achieving goals, anticipating
outcomes, and overcoming obstacles to unstructured, open-ended problems (Levy & Murnane,
2013). With relatively few well-delineated pathways to adulthood in the United States
(Macmillan, 2005; Mortimer, Oesterle, & Krüger, 2005; Settersten & Gannon, 2005), adolescents
need capacities for exercising agency to find their own pathways (Schwartz, Zamboanga, Meca,
& Ritchie, 2012). Furthermore, adolescents need capacities for exercising agency to address
current and future personal life problems, the resolution of which has implications for well-being
and mental health (Wehmeyer, Shogren, Little, & Lopez, 2017).
Many organized youth programs, such as leadership, arts, or civic action programs
provide opportunities for adolescents to develop a capacity for exercising agency. For example,
some youth programs aim to foster adolescents’ skills for setting and achieving personal or group
goals through work on self-selected and/or self-directed projects. Such self-directed projects
include planning and running an event to address a community need (Larson & Hansen, 2005) or
working on a production (Heath, 1998). These projects are thought to develop adolescents’
agency related capacities for a) thinking strategically about how to accomplish work (e.g., make
flexible plans with contingencies; Larson & Angus, 2011; Larson & Hansen, 2005), b) assuming
responsibility for meeting expectations and obligations (e.g., personal, group, and program;
Salusky et al., 2014; Wood, Larson, & Brown, 2009), and c) becoming intrinsically motivated by
the work and its challenges (Larson, Hansen, & Walker, 2005; Pearce & Larson, 2006).
4
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
Youth program leaders play a key role in assisting adolescents in the process of building
capacities for agency. Larson and colleagues proposed that two types of adult supports contribute
to the development of adolescents’ capacity for agency (Larson & Angus, 2011; Larson &
Hansen, 2005; Larson, Lampkins-Uthando, & Armstrong, 2014). First, adult leaders can provide
directive assistance (e.g., establishing work norms, setting deadlines and benchmarks) to help
adolescents structure, control, and steer their own work. Second, adult leaders can provide
autonomy support to ensure that youth retain control (e.g., decision-making) over the work. The
purpose of this study was to evaluate the relations of youth program leader’s directive assistance
and autonomy support on adolescents’ capacity for exercising agency.
Agency Development
Although there are numerous nuanced definitions of human agency within psychology,
most focus on individuals’ sense of empowerment and belief in their ability to achieve a desired
goal or outcome (e.g., Wehmeyer et al., 2017). A core tenant of human agency theory is that
individuals seek to engage in self-determined, agentic actions—to exercise volition and control
over their actions (Bandura, 2006; Ryan & Deci, 2000). Thus, agentic individuals “are able to
decide for themselves which options are ‘right’ for them, to sort through these options largely on
their own, to ‘change course’ when their original plans are blocked, and to follow their efforts
through to completion” (Schwartz, Donnellan, Ravert, Luyckx, & Zamboanga, 2012, p. 341).
One’s sense of agency emerges from repeated experiences of engaging in self-directed actions,
that is, it emerges from exercising agency (Little, Hawley, Henrich, & Marsland, 2002). The
preponderance of research on human agency concerns agency-related beliefs or attitudes rather
than specific capacities or skills needed for achieving a desired goal or outcome. A focus on
5
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
capacities for exercising agency is needed to understand how we might create intentional
opportunities and conditions for adolescents to develop their capacity for agency.
Some scholars argue that adolescence offers an enhanced opportunity to develop
capacities for agency because of the concomitant rapid expansion and integration of cognitive
(e.g., metacognitive) and affective (e.g., motivational) regulatory capacities (Hansen & Jessop,
2017; Larson & Angus, 2011). With the advent of puberty, adolescents experience an extended
period of rapid brain development for the apparent ontogenetic aim of building capacities for the
conscious self-regulation and coordination of cognition, affect, and behavior (Keating, 2004;
Luna & Sweeney, 2004). Adult life and work in contemporary society increasingly demands
these agency-related capacities (Larson, 2000). Developing a capacity for agency, however, is
not a foregone conclusion of neurological maturation; it requires volitional engagement in
activities that demand the exercise of agency-related capacities. Thus, without specific
experiences that promote the development of adolescents’ capacity for agency, this capacity is
less likely to flourish (Hansen & Jessop, 2017).
Youth Programs as a Setting for Learning Capacities for Exercising Agency
Youth programs can provide adolescents with opportunities to develop a capacity for
agency (Eccles & Gootman, 2002; Heath, 1998; Mahoney, Larson, Eccles, & Lord, 2005).
Leadership and arts programs in particular are thought to provide foundational conditions for
learning capacities for exercising agency (Larson & Angus, 2011). For example, the National
Future Farmers of America (FFA), a salient youth program in rural communities, has made
adolescent participants’ development of capacities for exercising agency a major focus. FFA
participants can choose projects they want to work on, or create and implement their own
projects (Larson & Hansen, 2005; Larson et al., 2005). A key feature of such projects is that they
6
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
typically occur over extended periods (e.g., weeks or months), and thus push adolescents to
extend their reasoning and planning further into the future. Such experiences contribute to
developing the adolescents’ anticipatory or ‘prospective’ thinking capacities (Heath, 1998;
Larson, Lampkins-Uthando, et al., 2014).
Voluntary work on projects in youth programs can also provide opportunities for
adolescents to interact with complex human systems (e.g., businesses) that have their own
dynamics and rules of operation (e.g., Catch 22’s; Larson & Hansen, 2005). Larson and
colleagues propose that such interactions promote development of adolescents’ strategic thinking
capacities; for example, developing flexible heuristics (versus inflexible plans), creating schemes
for how others’ perspectives and motives affect plans, and engaging in pragmatic means-ends
reasoning about achieving goals (Larson, 2007; Larson & Hansen, 2005; Larson, Lampkins-
Uthando, et al., 2014).
Youth programs can also provide opportunities for adolescents to develop higher-order
motivational capacities related to agency. Adolescents’ voluntary work on projects in youth
programs, particularly projects they initiate and self-direct, can heighten a sense of ownership
and responsibility, which helps sustain adolescents’ engagement in a project and see it through to
completion (Salusky et al., 2014; Wood et al., 2009). Scholars suggest this sustained volitional
engagement promotes the development of higher-order motivational capacities, including a
capacity to derive enjoyment and motivation from surmounting the challenges of the work
(Blumenfeld, Kempler, & Krajcik, 2006; Heath, 1999; Moore & Hansen, 2012; Pearce & Larson,
2006).
Adult Leader’s Supports and Adolescents’ Agency Development
7
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
Adult youth program leaders can play an instrumental role in supporting the development
of adolescents’ capacity for agency (Halpern, 2005; Kirshner, 2008; Larson & Angus, 2011).
(Peer relationships can also play an important role in the learning process, but this dimension
was beyond the scope and data of this study). Larson and Angus (2010) proposed two types of
supports adult program leaders can provide that contribute to adolescents’ developing a capacity
for exercising agency. First, adult youth program leaders can provide ‘directive assistance’ by
creating and maintaining appropriate structures that connect project activities to the program’s
culture, traditions, and norms (Halpern, 2005; Kirshner, 2008; Larson & Angus, 2011). Directive
assistance is thought to promote adolescents’ capacity for agency by giving them access to work
expectations (e.g., deadlines, accountability) and specific tasks required to accomplish their work
(Durlak & Weissberg, 2007; Pearce & Larson, 2006). Over time a “culture of accountability”
within a program can promote adolescents’ internalization of a sense of personal responsibility
for the outcomes of their work (Larson, Griffith, et al., 2014; Wood et al., 2009). Larson and
Angus (2011) further proposed that directive assistance promotes adolescents’ development of
motivation-related capacities for exercising agency. That is, adolescents learn they can mobilize
(i.e., regulate) their effort and purposively engage in the challenges of the work, which helps
them see a project through to completion and receive “confirmatory feedback” that they are
capable of meeting a priori demands of the project.
Second, adult youth program leaders can provide facilitative assistance through
intentionally promoting adolescents’ autonomy and control over their own work (Larson &
Angus, 2011; Larson & Hansen, 2005). Although Larson and Angus (2011) use the term
facilitative assistance, we use the term autonomy support here since support for autonomy was
integral to their conceptualization and it reflects the current study’s operationalization. Larson
8
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
and Angus (Larson & Angus, 2011) proposed that autonomy support promotes adolescents’
learning capacities for thinking strategically about how to accomplish work, including complex
means-end, pragmatic, and anticipatory reasoning, as well as contingency planning and task
prioritization (Larson & Hansen, 2005). A rich research literature from self-determination theory
indicates that choice and autonomy over one’s behavior and actions in a setting are essential for
gaining a sense of agency (Ryan & Deci, 2000). Based on the qualitative research of Larson and
colleagues, as well as self-determination theory, both autonomy support and directive assistance
should promote adolescents’ developing a capacity for exercising agency. Although the literature
suggest the type of support may target different capacities, there is not yet enough research to
hypothesize specific relations.
This Study
The purpose of this study was to evaluate the hypothesized relations of adult youth
program leaders’ directive assistance and autonomy support with the development of
adolescents’ capacity for exercising agency over the course of two years (two measurement
occasions). We addressed two hypotheses. First, directive assistance and autonomy support by
the adult leader in a program will positively correlate with adolescent participants’ capacity for
exercising agency within each measurement occasion. Second, we hypothesize that directive
assistance and autonomy support at Time 1 will positively predict adolescents’ capacity for
exercising agency at Time 2. Prior research has primarily been qualitative and thus does not
suggest the relative magnitude of these relations.
Methods
Sample
9
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
The sample for this study is rural high school students participating in local Chapters of
the National FFA over the course of two consecutive years. The FFA was a salient context in
which to examine adolescents’ development of their capacity for exercising agency. The stated
mission of the FFA is to make “a positive difference in the lives of students by developing their
potential for premier leadership, personal growth and career success through agricultural
education” (emphasis added; "FFA Mission & Motto," 2017). The three precepts in FFA’s
mission (italicized in mission statement) emphasize many components that reflect capacities for
exercising agency. FFA defines the premiere leadership component as “influence.” Included
within this precept is the development of capacities for Action (i.e., skills and competencies for
achieving outcomes), Vision (i.e., having a clear vision of what the future should be), and
Continuous Improvement (i.e., pursuit of learning and growth). Next, FFA defines the personal
growth component as “the positive evolution of the whole person” (p. 7). Included within this
precept is the development of capacities for Professional Growth (i.e., cultivating awareness and
application of skills for career success) and Mental Growth (i.e., developing applied and
effective reasoning, thinking, and coping skills). Last, FFA defines the career success component
as “qualities, attributes and skills” (p. 7) for future career success and an ability to be an
effective, contributing member of society. Included within this precept is the development of
capacities for Decision Making (i.e., “ability to analyze a situation and execute an appropriate
course of action” (p. 7)) and Flexibility/Adaptability (i.e., capacities for and will (drive) to
change).
The FFA is a common youth program located in rural high schools with linkages to
schools’ agricultural education courses. Each Chapter has at least one adult advisor who is also
the agricultural teacher during the school day. Chapters follow the National FFA curriculum and
10
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
structure, which provides a consistent emphasis across Chapters on the FFA mission. A particular
Chapter can participate in a range of FFA activities, such as social and agriculture-related events,
community service projects, and supervised agricultural experiences (owning and operating
agriculture-related business), plus state and national activities (e.g., conventions) and events
(e.g., skill competitions). FFA’s strong emphasis on developing adolescents’ capacities related to
exercising agency provided an appropriate setting for this study.
Sampling
Purposive sampling of FFA Chapters. Three criteria were used to select 10 FFA
Chapters from an initial pool of 58 Chapters located within a two-hour drive (approximately 150
miles) of a Midwest university; the distance was necessitated by logistic and funding constraints.
For the first criterion, performance data collected by the Facilitating Coordination in Agricultural
Education board on a Chapter’s “quality indicators” and “program standards” was used to
compute a sum score for each Chapter. We included 28 performance indicators: eight for
classroom instruction (e.g., teacher possesses minimum of 2,000 hours of work experience), four
for FFA participants supervised agricultural experiences, and 16 for FFA activities (e.g.,
conducted Agricultural Expos). One point was given for each indicator met by a Chapter for a
potential range of 0-28 points. Chapters with less than 15 points were excluded from the
selection pool because we reasoned they would not provide sufficient opportunities for
adolescents to engage in the FFA curriculum and thus limit the potential to develop a capacity for
agency.
Chapters were also considered for selection only if they had at least one advisor who had
been teaching for three or more years in the current school. We reasoned there could be
significant year-to-year change in a Chapter’s involvement in the FFA curriculum for advisors
11
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
who had recently transitioned to a new school/Chapter. Finally, only Chapters with at least 40
student members were considered for selection to ensure reasonable representation of the overall
experience in a Chapter. After applying these three criteria to the initial pool, there were 16
chapters with at least 15 points. Based on a median split, we divided Chapters into two groups:
eight with a score of 20 or more points and eight with 15-19 points. This split was deemed
necessary to ensure representation of the range of programs resulting from the random selection
of Chapters. From each of these two groups of eight, we randomly selected five chapters for a
final sample of 10 Chapters.
Adolescents. The adolescent sample consisted of 441 high school students (66.2% male)
from 10 FFA Chapters. The mean age of participants at Time 1 was 16.03 (range = 14 to 19). At
Time 1 there were 161 (36.7%) freshman, 164 (37.2%) sophomores, and 116 (26.3%) juniors.
Seniors were not included here because they graduated by Time 2. Thus, there were longitudinal
data for three grade groups. The majority of students, 66.8%, lived in a “rural area (on a farm in
the country or not on a farm but in the country),” 32.1% lived “in a small town or city (less than
10,000 people),” and the remainder (1.1%) reported living in a “medium size city (between
10,000 and 200,000 people).” The sample was 84.1% White, 1.8% Native American, 0.5%
Hispanic, 4.3% reported multiple ethnicities or “other,” and 9.3% chose not to self-report.
Adult advisors. There were 11 adult advisors (one school had two advisors). Ten of the
11 advisors were male and all were White. The average number of years these advisors had been
teaching was 28.8, with a range of 5 to 37 years.
Procedures
The research team staff administered questionnaires to students and advisors during the
spring of 2006, Time 1 (T1), and again in the spring of 2007, Time 2 (T2). We followed ethical
12
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
research procedures approved by the university’s institutional review board. Parents received
information regarding the study approximately two weeks before the scheduled administration of
the questionnaires, and they could indicate if they did not want their child to participate in the
study; none chose to do so. On the days of the administration, adolescents indicated their assent
and completed the study questionnaire; they did not receive compensation. Advisors also gave
their consent and completed a questionnaire in which they rated each student’s agency-related
capacities. Advisors were compensated $50 (plus $1 for each additional student if there were
more than 40) for the time required to complete the measure on each student in their Chapter.
Measures
Adolescent-reported directive assistance. Three items developed for this study assessed
directive assistance focusing on advisor’s expectations about adolescent’s work on projects in
FFA. The items were, “Advisor’s place high importance on finishing the projects we start,”
“Advisor’s place high standards on youth in FFA,” and “Advisor’s will be disappointed with you
if you do not finish what you said you would do” (reverse coded). Items were rated on a 5-point
Likert scale from 1 = “Strongly disagree” to 5 = “Strongly agree.” The three items were used as
indicators of a latent directive assistance variable. Internal reliability for the scale was = 0.70
(T1) and 0.81 (T2). The latent means for directive assistance ranged from 3.28 to 3.56.
Adolescent-reported autonomy support. We operationalized autonomy support using
the six-item short form of the Learning Climate Questionnaire (LCQ; Williams, Grow,
Freedman, Ryan, & Deci, 1996). The LCQ is a self-report measure that asks participants to rate
the autonomy supportive conditions within a given setting. Items on the measure were worded in
reference to the advisor’s level of autonomy support perceived by the participant. A sample item
from the scale is, “I feel that my advisor provides me choices and options.” Items were on a 5-
13
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
point Likert scale from 1 = “Strongly disagree” to 5 = “Strongly agree.” In previous studies, the
LCQ has demonstrated strong internal consistency (e.g., > .90; Black & Deci, 2000).
Although the LCQ measure has a rich history in research, initial longitudinal
confirmatory factor analyses (CFA; described below) of the study’s conceptual model indicated
problems with model fit associated with the autonomy support scale. Subsequent confirmatory
factor analysis with only the LCQ measure failed to confirm the latent construct using all six
items. Based on model fit indices (CFI ≥ .90; RMSEA ≤ .08), factor loadings, and modification
indices, we dropped two items because they failed to load on the latent factor and no reasonable
modification (e.g., correlated residuals) led to adequate model fit (i.e., CFI > .90). The two
dropped items were, “I feel understood by my advisor” and “My advisor conveyed confidence in
my ability to do well.” The remaining four items focused on the advisor’s actions related to
adolescents’ retaining control for ideas, plans, and work. The four items were used as indicators
of a latent autonomy support variable. For the present study, Cronbach’s internal reliability
alphas were 0.93 (T1) and .96 (T2). The latent means for autonomy support ranged from 3.56 to
3.81.
Adolescents’ self-reported capacity for exercising agency. Adolescents’ self-reported
capacity for agency was calculated using three scales: engagement with challenge, strategic
planning scale, and responsibility and dependability. First, the engagement with challenge scale
(EwC) included six items that assess the linkages between adolescents’ intrinsic motivation and
the challenges that occur from working toward a goal or a project (Moore & Hansen, 2012).
Conceptually, engagement with challenge concerns the pairing of challenges and enjoyment
(intrinsic motivation). Thus, items in the scale focus on this pairing, rather than on separately
assessing challenge and intrinsic motivation constructs. We recognize this pairing may give the
14
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
appearance of double-barreled questions. However, the scale authors argue that in this instance
separately assessing challenge and motivation would undermine the validity of the construct and
increase, rather than reduce ambiguity. A sample item from the scale is: What we do in this
program is both difficult and enjoyable.” Students indicated their agreement with each statement
using a 6-point Likert scale from 1 = “Strongly disagree” to 6 = “Strongly agree.” Second, the
strategic planning scale (SPS) was developed for this study to assess students’ capacity to make
and execute plans. Students rated how much they engaged in three planning strategies when
working on projects in FFA: planning ahead, planning when to do tasks, and making back-up
plans. The items were rated on a 5-point Likert scale with descriptive anchors of a strategy
associated with the two ends and middle of the response scale. For example, in response to “How
much have you planned ahead?” the anchors were 1 = “Little, I figured things out as I went
along,” 3 = “Some, I made a couple of specific plans before starting—figured out the rest as I
went along,” and 5 = “A lot, I developed a pretty complete plan of what to do before starting.”
Third, personal responsibility and dependability within the FFA was assessed with four items
developed for this study. For two of the items, adolescents indicated the description that best
represented “how responsible” and “how dependable” they act in FFA on a 7-point Likert scale
from 1 = “Very irresponsible/ undependable” to 7 = “Very responsible/dependable.” For the
other two items, adolescents indicated their agreement on a 5-point Likert scale from 1 =
“Strongly Disagree” to 5 = “Strongly Agree” to the following: “I have had a lot of
responsibilities to do in FFA” and “I have a lot of obligations that I need to complete in FFA.”
All items were first converted into a 10-point scale (percent of maximum score) in order
to provide a common metric for both the advisor’s ratings of each students’ capacities (described
next) and the adolescent-reported ratings (Little, 2013). Once in the new metric, we created a
15
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
mean score for each of the three domains of capacities for exercising agency. Creating three
mean scores rather than using all indicators from each scale was preferable since there was a
different number of items in each scale (e.g., six items for the EwC and three for the SPS). The
three mean composites were used as indicators of a latent variable of students’ self-reported
capacity for agency. Internal reliability for the three mean indicator scores of students self-
reported agency was = .75 (T1) and .76 (T2). The latent means for adolescents’ self-reported
agency ranged from 5.73 to 6.15.
Advisor’s ratings of each adolescents’ capacity for agency. An adult advisor of a
Chapter rated each adolescent’s capacity for agency across the same three domains as the
adolescent-reported capacity for agency measure: engagement with challenges, planning, and
responsibility. The following definitions were given for each of the three domains: “Engagement
with Challenge refers to how motivated and engaged each student has been in challenging
activities of FFA;” “Planning and Executing the plan requires skill proficiency in many areas.
This includes: Advanced Planning, Scheduling, Creating Back-up Plans, Not Procrastinating, and
Monitoring and Making Adjustments;” and “Responsibility refers to being someone who can be
counted on to fulfill obligations.” Advisor’s rated each adolescent compared to other adolescents
of the same age/grade using a 10-point scale where 1 = ‘0-10%,’ 2 = ‘11-20%,’ 3 = ‘21-30%,’ 4 =
‘31-40%,’ 5 = ‘41-50%,’ 6 = ‘51-60%,’ 7 = ‘61-70%,’ 8 = ‘71-80%,’ 9 = ‘81-90%,’ and 10 = ‘91-
100%.’ The three advisor-rated items were used as indicators of a variable of adolescents’ latent
capacity for agency. The internal consistency was 0.97 at both T1 and T2. The latent means for
advisor’s ratings of each adolescent’s capacity for agency ranged from 5.29 to 5.95.
---------------------------------------
Place Table 1 approximately here
16
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
----------------------------------------
Plan of Analysis
Preliminary analyses. We first conducted analyses to test for univariate and multivariate
normality of study variable, as well as the distribution of responses on measures to identify
potential scale issues (e.g., ceiling effects).
Longitudinal confirmatory factor analyses (CFA). Prior to evaluating the proposed
hypotheses, we conducted a longitudinal CFA of the study’s conceptual model (Figure 1) to
ensure we were measuring the same latent constructs across the time and grade groups and to
evaluate the homogeneity of parameters across both time and grade groups. The longitudinal
CFA proceeded in the following order: evaluation of measurement invariance (i.e., configural,
metric, scalar, and residual variance invariance) and evaluation of structural homogeneity (i.e.,
latent variances, covariances, and means).
We evaluated the quality of a given model’s fit using CFI and RSMEA values. The
criteria for acceptable fit were: CFI values of at least .90, and RMSEA values of .08 or less and
also within the 90% confidence interval (Kline, 2015). To compare the relative fit of two nested
models, we used the −2LL rescaled difference test (-2LL). Therefore, a p > .01 indicated the
more restrictive model (i.e., parameters constrained to be equal) maintained acceptable model fit
compared to the less restrictive model (i.e., freely estimated parameters); conversely, a p ≤.01
indicated a model failed to maintain acceptable model. When a model failed to maintain fit, we
used modification indices to identify parameters that were not equatable across groups and/or
time. The criteria for identifying a single unequatable parameter using a modification index was
2 > 6, p < .01 (Kline, 2015).
---------------------------------------
17
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
Place Figure 1 approximately here
----------------------------------------
Test of hypotheses. Each regression path was tested separately for significance using the
same nested model testing procedure (i.e., comparing a model with the path’s parameter
constrained to ‘0’ with a model in which the parameter was freely estimated) and significance
criteria described above (i.e., -2LL, with df = 1 and p ≤ .01). We used the nested model testing
procedure rather than the Wald test because it is not affected by sample size; the Wald test uses
the standard error of the estimated parameter coefficient to determine significance (Kline, 2015).
Results
Preliminary Analyses
We conducted univariate and multivariate tests of normality, since substantial departures
from normality create serious problems for interpretations based on the Maximum Likelihood
(ML) estimator (Byrne, 2010; DeCarlo, 1997). Mardia’s (1970) omnibus test of skew and
kurtosis with p < .01 indicated the data were not univariate or multivariate normal. Thus, we
used the robust maximum likelihood estimator (MLR) in all subsequent analyses, which were
conducted in Mplus, Version 6 (Muthén & Muthén, 2010). In addition, although the data were
nested (students within programs), and thus violated assumptions of independence, sample size
did not permit multi-level modeling. To account for the effect of the nesting of the data on the
results, we used the “cluster” function in Mplus, which adjusts for the intraclass correlation with
a scaling correction factor. Table 2 displays the standardized factor loadings and standard errors
for the latent factors in the study’s model.
---------------------------------------
Place Table 2 approximately here
18
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
----------------------------------------
Longitudinal CFA
The longitudinal CFA model proceeded in the following order: evaluation of
measurement invariance (Table 3, Panel A) and evaluation of structural homogeneity (Table 3,
Panel B).
Measurement invariance. The configural measurement model was specified (i.e., all
parameters freely estimated) with the second loading for each factor fixed to ‘1’ and its
corresponding intercept fixed to ‘0’ for all grade groups and both time points to identify the
model. The configural model demonstrated acceptable fit, CFI = .93, RSMEA = .06, 90% CI
[.06-.070]. However, modification indices indicated there were sources of misfit due to
correlated residuals (i.e., 2 > 6). We deemed it important to address these sources of misfit
since ignoring them could result in the error variances associated with the correlated residuals
being inappropriately included in the latent parameters, which can result in unstable parameter
estimates across models. We applied the following conceptual rule when deciding to allow
correlated residuals in order to avoid inflating model fit solely for statistical reasons: the
indicated correlated residuals had to be conceptually related, and preferably one residual had to
be from the adolescent and the other from the advisor. After applying this rule, we allowed three
correlated residuals, tested in sequential models starting with the largest modification index (See
Table 3, Models 1.b-1.d). Fit statistics for of the final configural model (Model 1.d) was CFI = .
94, RSMEA = .06, 90% CI [.054-.067].
---------------------------------------
Place Table 3 approximately here
----------------------------------------
19
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
The metric invariance model (i.e., all factor loadings held equal across time and groups)
was next evaluated with the 11th grade group’s Time 1 factor variances fixed to ‘1’ as the
reference group and time period. The metric invariance model did not maintain acceptable fit
compared to the final configural model, −2LL(45) = 101.55, p < .01. We evaluated subsequent
metric invariance models (Table 3, Models 2.b-2.e), freely estimating individual factor loadings,
starting with the loading associated with the largest modification index. Based upon nested
model comparisons, there were four factor loadings freely estimated because they were not
equatable: two T1 autonomy support items for the 9th grade group (Models 2.b-2.c), one T1
adolescent-reported planning item for the 11th grade group (Model 2.d), and one T1 directive
assistance item for the 10th grade group (Model 2.e). All other factor loadings were equatable
across groups and time. Since there were only two measurement time points, if a parameter was
not equatable across groups it was also not equitable across time. The final, partial metric
invariance model (2.e) maintained acceptable fit compared to the configural model, −2LL(41) =
48.41, p =.20.
The scalar invariance model (i.e., all indicator intercepts held equal across time and
groups) was next evaluated, with the 11th grade group’s Time 1 latent means fixed to ‘0’ as the
reference group (Table 3, Model 3.a). The scalar invariance model maintained acceptable fit
compared to the partial metric invariance model, −2LL(41) = 46.68, p =.25. However,
modification indices indicated that 10th grade group’s intercept for the first item of T2’s
autonomy support measure could not be constrained to be equal to the other grade groups. After
freeing this one intercept, the final, partial scalar invariance model maintained acceptable fit
compared to the partial metric invariance model, −2LL(40) = 41.00, p = .43, and was a
20
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
significant improvement in fit over the initial scalar invariance model, −2LL(1) = 20.34, p < .01.
No additional modifications indicated.
Fit of the residual invariance model (i.e., all indicator residuals held equal across time
and groups) was next evaluated (Table 3, Model 4.a). The residual invariance model maintained
marginally acceptable fit compared to the final scalar model, −2LL(59) = 77.61, p = .05.
Modification indices indicated two residual variances could not be constrained to be equal: T2’s
residual variance for the 9th grade group’s autonomy support item four and T2’s residual variance
for 11th grade group’s advisor-report of students’ responsibility item. The final residual variance
model (4.c) maintained acceptable fit compared to the partial scalar invariance model, −2LL(57)
= 66.34, p = .19, and significantly improved fit compared to the previous residual invariance
model (2.b), -2LL(2) = 11.27, p < .01.
Structural homogeneity. Panel B of Table 3 presents results of the evaluation of
structural homogeneity. Fit of the homogeneity of the latent variances (i.e., same latent factor
variances held equal to ‘1’ across time and groups) was first evaluated (Table 3, Model 5.a). The
latent factor variances model failed to maintain acceptable fit compared to the partial residual
invariance model (4.c). −2LL(20) = 38.54, p = .01. Modification indices indicated that the 9th
grade group’s T1 directive assistance variance was not equatable across groups or time. With
this parameter freed, the latent variances model demonstrated acceptable fit compared to the
partial residual invariance model (4.c), −2LL(19) = 27.88, p = .09. No additional modifications
indicated. No additional modifications indicated.
Next, fit of the homogeneity of the covariances was tested. First, the within time (cross-
sectional) latent covariances were held equal across grade groups and time (6.a). This model
maintained acceptable fit compared to final latent variances model, −2LL(31) = 40.11, p = .13.
21
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
Second, the homogeneity of the latent cross-lag coefficients model (7.a) maintained acceptable
fit compared to latent covariance model, −2LL(23) = 17.60, p = .78. Third, the homogeneity of
the latent autocorrelations model (8.a) failed to maintain acceptable fit compared to the
homogeneity cross-lag model, -2LL(8) = 26.45, p < .01. The only autocorrelation that needed to
be freed was for the 10th grade group’s directive assistance, -2LL(7) = 16.58, p = .02.
Finally, fit of the homogeneity of latent means model was evaluated, with factor
variances fixed to ‘1’ as they were in the prior model (8.b) and all factor means fixed to ‘0’. The
homogeneity of latent means model maintained marginally acceptable fit compared to the final
autocorrelations model, -2LL(20) = 29.12, p = .09. Modification indices indicated that the 11th
grade groups Time 1 mean for directive assistance was not equatable across grade groups or
time. After freeing this mean, this latent means model (9.b) maintained acceptable fit compared
to the autocorrelations model (8.b), -2LL(19) = 23.35, p = .22. Although model 9.b met our fit
criteria, we subsequently freed the 9th grade groups Time 1 mean for advisor reports of youth
agency for two reasons: modification indices indicated the mean was close to being unequatable
(2 = 5.30) and keeping it equated in the subsequent regression model caused parameter
instability that resulted in non-equivalent fit. After freeing this mean, the final latent means
model (9c) maintained acceptable fit compared to the autocorrelations model (8b), -2LL(18) =
22.89, p = .20.
Evaluation of Study Hypotheses
To test the hypotheses, we converted all cross-lag paths from covariances to regressions
(10.a, Table 3) in a base regression model (Figure 1) that had all cross-sectional correlations and
regressions (cross-lag and autoregressions) estimated, and had the required equivalent fit to the
22
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
final latent means model (9.c). All subsequent models were compared to this base regression
model for evaluating specific latent relationships using the nested model approach (Table 4).
---------------------------------------
Place Table 4 approximately here
----------------------------------------
Cross-sectional correlations among latent factors. Nested model -2LL difference tests
were conducted to evaluated the significance of each latent correlation by comparing the fit of
the model with the correlation estimated to the model when the correlation was set to ‘0.’ These
six tests indicated all the latent correlations were significant and positive (Table 4, Panel A). As
hypothesized, both directive assistance and autonomy support were positively correlated with
adolescent-reported capacity for agency (r = .56 and r = 56, p < .01, respectively) and advisor-
reports of adolescents’ capacity for agency (r = .25 and r = .23p < .01, respectively). Directive
assistance and autonomy support were moderately correlated (r = .55, p < .01) with each other,
as were adolescent-reported and advisor-reports of adolescents’ capacity for agency (r
= .51, p
< .01).
Longitudinal paths. The same nested model difference test was used to evaluate all
longitudinal paths. All autoregressive paths were statistically significant and positive (Table 4,
Panel B). We next evaluated the cross-lag regression paths for significance corresponding to our
hypotheses that both directive assistance and autonomy support at T1 would positively predict
both T2 adolescent-reported capacity for agency and T2 advisor-reports of adolescents’ capacity
for agency (Table 4, Panel C). Of the four hypothesized paths, only T1 autonomy support
significantly predicted T2 adolescent-reported capacity for agency (
= .21, p < .01). This
23
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
regression path explained approximately 4% of the variance in T2 adolescent-reported agency,
after controlling for T1 agency.
Finally, tests of the remaining cross-lagged paths indicated that two of the eight paths
were significant (Table 4, Panel C). T1 directive assistance negatively predicted T2 autonomy
support (
= -.19, p < .01), explaining approximately 5% of the variance in T2 autonomy
support. T1 autonomy support positively predicted T2 directive assistance (
= .27, p < .01),
explaining approximately 8% of the variance in T2 directive support.
The final pruned regression model had three significant paths: autonomy support
positively predicted both future adolescent-reported agency and directive support, and directive
support negatively predicted future adolescent-reported autonomy support.
Discussion
This study evaluated hypotheses that youth program adult advisors’ directive assistance
and autonomy support would predict adolescents’ capacity for agency. Results of this study
provided partial support for the hypotheses. Within each time point (cross-sectional),
adolescents’ perceptions of their advisor’s directive assistance and autonomy support were
positively and moderately correlated with adolescents’ capacity for agency. Longitudinally, only
autonomy support at T1 predicted adolescents’ self-reported capacity for agency at T2. Overall,
findings suggest directive assistance and autonomy support may both be needed for immediate
exercise of agency, while only autonomy support seems to promote the over-time development
of adolescents’ capacity for agency.
Advisor’s Supports and Adolescents’ Capacity for Agency
The pattern of cross-sectional and longitudinal findings in this study suggests directive
assistance and autonomy support may relate to a capacity for agency in different ways.
24
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
Directive assistance. The cross-sectional, but not longitudinal relationships, between
directive assistance and adolescents’ capacity for agency might reflect the functional immediacy
of directive support’s role—to provide a motivational context for adolescents to engage in their
work and exercise agency. Advisor’s directive support, then, perhaps provided “external”
motivational prompts (e.g., deadlines) that helped adolescents mobilize their effort to complete
projects. This finding is consistent with a qualitative study by Larson and Angus (2011) who
reported that youth program advisor’s directive assistance helped adolescents marshal their effort
to finish projects, which was important in order for youth to “obtain the validating feedback that
success provided” (p. 298).
The results of this study left the role of directive assistance for building adolescents’
capacity for agency over-time ambiguous. Directive assistance at Time 1 might have failed to
predict adolescents’ capacities for agency at Time 2 because of our level of analysis. We
analyzed a capacity for agency as a whole, rather than as its component parts. One of those
component parts was personal responsibility for one’s work. Larson and colleagues suggest that
the directives and norms in a program for youths’ work (e.g., standards for accountability for
work) promote adolescents’ adoption and internalization of those norms, including personal
responsibility (Larson, Griffith, et al., 2014; Wood et al., 2009). In this study, directive assistance
focused on adolescents’ perceptions of the norms their advisors have for their work but not if
these norms had become internalized as a sense of responsibility. Alternatively, T1 directive
assistance’s failure to predict T2 capacity for agency might indicate that directive support is not
integral to fostering capacities for agency. Qualitative research suggests adult program leaders
often face a delicate balancing act between exerting too little and too much control and direction
25
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
over youths’ activities (Walker & Larson, 2006). Future research is needed to more fully explore
the role of directive assistance and its relation to adolescents’ exercise of agency.
Autonomy support. In this study, autonomy support was related to adolescents’ capacity
for agency at each time point and over time (self-report only). Research from self-determination
theory has consistently found that support for autonomy (e.g., choice and opportunity for self-
direction) promotes a greater sense of perceived competence and self-efficacy for an activity
(Ryan & Deci, 2000). In the present study, autonomy support correlated with both reports (self-
report and advisor) of adolescents’ capacity for agency within each time point (cross-sectional),
suggesting that, like directive assistance, autonomy support was related to adolescents’
immediate exercise of agency. Unlike directive assistance, however, autonomy support appeared
to also foster adolescents’ capacity for agency over time.
Based on their qualitative research, Larson and Angus (2011) proposed that adult youth
program leaders’ support for adolescents’ control over their work (e.g., freedom to make
decisions and experiment) promotes their learning strategic thinking skills. Larson and Angus
(2011) described strategic thinking skills as “Thinking that involves the inference of system
processes as a means to anticipate events and formulate courses of action to achieve goals in the
program” (p. 282). Results of the current study provides some support for this proposition as
strategic planning was one component of our operationalization of adolescents’ capacity for
agency. However, the present findings suggest that the role of autonomy support may not be
limited to fostering strategic thinking. Autonomy support may also play a role in fostering other
capacities needed for exercising agency. Self-determination theory research has found that
support for autonomy facilitates intrinsic motivation, an inclination to seek out novelty and
challenge, and experience them as rewarding (Ryan & Deci, 2000). In this study, we
26
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
operationalized the motivational component of a capacity for agency as enjoyment of meeting
and overcoming challenges that occur while working toward a goal or project. Thus, the current
findings are consistent with self-determination theory research.
Strengths and Limitations
Our operationalization of adolescents’ capacity for agency differed in important ways
from other lines of research on agency. Much of the research on agency focuses on the outcomes
of individuals’ agency-related beliefs or perceptions of their ability (Bandura, 1982; Shogren,
Little, & Wehmeyer, 2017). In this study, we focused on three capacities for exercising agency
because, theoretically, they transfer across different settings and because youth program leaders
can intentionally structure opportunities to build these capacities. We presume that building these
capacities would also facilitate adolescents’ agency-related beliefs, but we did not assess agency-
related beliefs, which could have provided validity evidence for our agency construct. It would
be important to assess both capacities and beliefs in a future study. We also think there is
important measure development work to be done on these three, and perhaps additional, agency-
related capacities.
There could be at least two reasons that directive assistance and autonomy support at
Time 1 failed to predict T2 advisor’s reports of adolescents’ capacity for agency. First, to ease
time demands, we asked advisors to rate each youth in their program (40+ youth) with three
items representing the three capacities, rather than with the same 13 items adolescents rated,
which resulted in non-equivalent measures of agency capacity between youth and advisors.
Despite non-equivalent measures, however, there was a moderately strong cross-sectional
correlation between advisor-reported and adolescent-reported capacity for agency, r = .54.
Second, we may have inadvertently reduced the likelihood of finding change in advisor’s reports
27
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
of adolescents’ capacity agency because advisors used an age-norm rating scale (i.e., agency in
relation to peers of same age) and not a criterion-referenced scale (e.g., change in capacity
relative to individual’s past). Thus, an individual’s change in the capacity for agency would need
to be exceptional relative to peers to indicate growth. Partial evidence for this explanation can be
seen in the significant cross-sectional correlations between both directive support and autonomy
support and advisor’s ratings.
We sampled youth programs from the same organization that followed a common youth
development curriculum, which helped reduce that likelihood that the pattern of findings were
due to curricular or organizational differences. Although advantageous for present purposes, the
purposive sampling leaves unanswered questions about the variation and impact of these
foundational conditions across the range of youth programs (e.g., sports, arts, academic clubs)
and the settings (e.g., urban). For example, how does participation in youth programs that, of
necessity, limit choice and control over tasks and activities (e.g., organized youth sport practice)
affect learning different capacities for the exercise of agency? Addressing this and related types
of questions in future research could provide valuable insights into how to promote the
development of capacities for exercising agency across the spectrum of youth programs.
Finally, we did not examine adolescents’ patterns of involvement in projects and Chapter
activities, which would be important in subsequent research to evaluate their relation to a
capacity for exercising agency. Quantitative research has found correlations between indicators
of “dosage” (e.g., number of hours participating in program) and developmental outcomes
(Fredricks & Eccles, 2006). Qualitative research also suggests that being able to work on projects
over extended periods of time is a key factor that promotes adolescents’ development of different
capacities for agency (Larson & Angus, 2011; Larson & Hansen, 2005). Future research, then,
28
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
could assess different indicators of involvement in projects and activities, and how these
indicators interact with the type of support advisors provide to facilitate the development of a
capacity for agency.
Conclusion
Overall, both directive assistance and autonomy support appeared related to a capacity
for exercising agency at a given time, but only autonomy support appeared to help youth build a
capacity for agency over time. This pattern suggests that the types of support adolescents receive
could differentially affect their development of a capacity for exercising agency. For long-term
development of agency, providing adolescents with autonomy support might be the more
effective strategy. Given the salience of a capacity for agency in adulthood, it will become
increasingly important for educators and society to understanding how to support its
development.
29
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
References
Bandura, A. (1982). Self-efficacy mechanism in human agency. American psychologist, 37(2),
122.
Bandura, A. (2006). Toward a psychology of human agency. Perspectives on psychological
science, 1(2), 164-180.
Black, A. E., & Deci, E. L. (2000). The effects of instructors' autonomy support and students'
autonomous motivation on learning organic chemistry: A self‐determination theory
perspective. Science education, 84(6), 740-756.
Blumenfeld, P. C., Kempler, T. M., & Krajcik, J. S. (2006). Motivation and cognitive
engagement in learning environments: na.
Byrne, B. (2010). Multivariate applications series. Structural equation modeling with AMOS:
Basic concepts, applications, and programming (2nd ed.). New York: Routledge/Taylor &
Francis Group.
DeCarlo, L. T. (1997). On the meaning and use of kurtosis. Psychological methods, 2(3), 292.
Durlak, J. A., & Weissberg, R. P. (2007). The Impact of After-School Programs that Promote
Personal and Social Skills. Collaborative for Academic, Social, and Emotional Learning
(NJ1).
Eccles, J., & Gootman, J. (2002). Community programs to promote youth development: National
Academies Press.
FFA Mission & Motto. (2017). Retrieved from https://www.ffa.org/about/who-we-are/mission-
motto
Fredricks, J. A., & Eccles, J. S. (2006). Is extracurricular participation associated with beneficial
outcomes? Concurrent and longitudinal relations. Developmental psychology, 42(4), 698.
30
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
Halpern, R. (2005). Instrumental relationships: A potential relational model for inner‐city youth
programs. Journal of Community Psychology, 33(1), 11-20.
Hansen, D. M., & Jessop, N. (2017). A Context for Self-Determination and Agency: Adolescent
Developmental Theories Development of Self-Determination Through the Life-Course
(pp. 27-46): Springer.
Heath, S. B. (1998). Working through language. In S. M. Hoyle & C. T. Adger (Eds.), Kids Talk:
Strategic Language Use in Later Childhood (pp. 217-240). New York: Oxford University
Press.
Heath, S. B. (1999). Dimensions of language development: Lessons from older children. In A. S.
Masten (Ed.), Cultural Processes in Child Development: The Minnesota Symposium on
Child Psychology (Vol. 29, pp. 59-75). Mahwah, NJ: Erlbaum.
Keating, D. (2004). Cognitive and brain development. InR. M. Lerner & L. Steinberg (Eds.),
Handbook of adolescentpsychology (pp. 45-84): New York: Wiley. KeatingCognitive and
brain development45Handbook of adolescent psychology2004.
Kirshner, B. (2008). Guided participation in three youth activism organizations: Facilitation,
apprenticeship, and joint work. The Journal of the Learning Sciences, 17(1), 60-101.
Kline, R. B. (2015). Principles and practice of structural equation modeling: Guilford
publications.
Larson, R. (2000). Toward a psychology of positive youth development. American psychologist,
55(1), 170.
Larson, R. (2007). From “I” to “We”: Development of the capacity for teamwork in youth
programs. Approaches to positive youth development, 277-292.
31
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
Larson, R., & Angus, R. (2010). 4 Pursuing Paradox: The Role ofAdults in Creating
Empowering Settings for Youth. Empowering settings and voices for social change, 65.
Larson, R., & Angus, R. (2011). Adolescents’ development of skills for agency in youth
programs: Learning to think strategically. Child Development, 82(1), 277-294.
Larson, R., Griffith, A., Wu, J., Raffaelli, M., Sugimura, N., & Guzman, M. (2014). How
adolescents develop responsibility: What can be learned from youth programs. Journal of
Research on Adolescence, 24(3), 417-430.
Larson, R., & Hansen, D. (2005). The development of strategic thinking: Learning to impact
human systems in a youth activism program. Human Development, 48(6), 327-349.
Larson, R., Hansen, D., & Walker, K. (2005). Everybody’s gotta give: Development of initiative
and teamwork within a youth program. Organized activities as contexts of development:
Extracurricular activities, after-school and community programs, 159-183.
Larson, R., Lampkins-Uthando, S., & Armstrong, J. (2014). Adolescents' development of new
skills for prospective cognition: Learning to anticipate, plan, and think strategically.
Journal of Cognitive Education and Psychology, 13(2), 232-244.
Levy, F., & Murnane, R. J. (2013). Dancing with robots: Human skills for computerized work.
Washington, DC: Third Way NEXT.
Little, T. D. (2013). Longitudinal structural equation modeling: Guilford Press.
Little, T. D., Hawley, P. H., Henrich, C. C., & Marsland, K. W. (2002). 17: Three Views of the
Agentic Self: A Developmental Synthesis. Handbook of self-determination research, 389.
Luna, B., & Sweeney, J. A. (2004). The emergence of collaborative brain function: FMRI studies
of the development of response inhibition. Annals of the New York Academy of Sciences,
1021(1), 296-309.
32
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
Macmillan, R. (2005). The structure of the life course: Classic issues and current controversies.
Advances in life course research, 9, 3-24.
Mahoney, J., Larson, R., Eccles, J., & Lord, H. (2005). Organized activities as developmental
contexts for children and adolescents. Organized activities as contexts of development:
Extracurricular activities, after-school and community programs, 3-22.
Mardia, K. V. (1970). Measures of multivariate skewness and kurtosis with applications.
Biometrika, 57(3), 519-530.
Moore, E. W., & Hansen, D. (2012). Construct-Validity of the Engagement with Challenge
Measure for Adolescents: Structural-and Criterion-Validity Evidence. Psychology, 3(10),
923.
Mortimer, J. T., Oesterle, S., & Krüger, H. (2005). Age norms, institutional structures, and the
timing of markers of transition to adulthood. Advances in life course research, 9, 175-
203.
Muthén, L. K., & Muthén, B. O. (2010). Mplus User's Guide: Statistical Analysis with Latent
Variables: User'ss Guide: Muthén & Muthén.
Pearce, N., & Larson, R. (2006). How teens become engaged in youth development programs:
The process of motivational change in a civic activism organization. Applied
Developmental Science, 10(3), 121-131.
Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic
motivation, social development, and well-being. American psychologist, 55(1), 68.
Salusky, I., Larson, R. W., Griffith, A., Wu, J., Raffaelli, M., Sugimura, N., & Guzman, M.
(2014). How adolescents develop responsibility: What can be learned from youth
programs. Journal of Research on Adolescence, 24(3), 417-430.
33
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
Schwartz, S. J., Donnellan, M., Ravert, R., Luyckx, K., & Zamboanga, B. L. (2012). Identity
development, personality, and well-being in adolescence and emerging adulthood:
Theory, research, and recent advances.
Schwartz, S. J., Zamboanga, B. L., Meca, A., & Ritchie, R. A. (2012). Identity around the world:
An overview. New Directions for Child and Adolescent Development, 2012(138), 1-18.
Settersten, R. A., & Gannon, L. (2005). Structure, agency, and the space between: on the
challenges and contradictions of a blended view of the life course. Advances in life
course research, 10(05), 35-55.
Shogren, K. A., Little, T. D., & Wehmeyer, M. L. (2017). Human Agentic Theories and the
Development of Self-Determination Development of Self-Determination Through the
Life-Course (pp. 17-26): Springer.
Walker, K. C., & Larson, R. W. (2006). Dilemmas of youth work: Balancing the professional and
personal. New Directions for Student Leadership, 2006(112), 109-118.
Wehmeyer, M. L., Shogren, K. A., Little, T. D., & Lopez, S. J. (2017). Development of Self-
Determination Through the Life-Course: Springer.
Williams, G. C., Grow, V. M., Freedman, Z. R., Ryan, R. M., & Deci, E. L. (1996). Motivational
predictors of weight loss and weight-loss maintenance. Journal of personality and social
psychology, 70(1), 115.
Wood, D., Larson, R., & Brown, J. (2009). How adolescents come to see themselves as more
responsible through participation in youth programs. Child Development, 80(1), 295-309.
34
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
Table 1. Latent Factor Means and Standard Errors (SE) across Grade Groups and Time
Scale
9th Grade Group 10th Grade Group 11th Grade Group
Time 1 Time 2 Time 1 Time 2 Time 1 Time 2
Directive Assistance (range 1-5) 3.56
(0.07)
3.48
(0.14)
3.37
(0.10)
3.28
(0.15)
3.75
(0.04)
3.51
(0.18)
Autonomy Support (range 1-5) 3.79
(0.18)
3.64
(0.19)
3.81
(0.18)
3.59
(0.24)
3.76
(0.19)
3.63
(0.24)
Adolescent-Reported Capacity for
Agency (range 1-10) 6.14
(0.10)
6.15
(0.14)
5.77
(0.25)
5.78
(0.26)
5.92
(0.22)
5.73
(0.15)
Advisor-Reports of Adolescent’s Capacity
for Agency (range 1-10) 5.29
(0.45)
5.47
(0.35)
5.34
(0.39)
5.76
(0.59)
5.93
(0.39)
5.95
(0.46)
35
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
Table 2. Standardized Factor Loadings and Standard Errors for Study Constructs by Grade
Groups and Time
9th Grade
Group 10th Grade Group 11th Grade
Group
Time 1 Time 2 Time 1 Time 2 Time 1 Time 2
Directive Assistance
λ
SE
λ
SE
λ
SE
λ
SE
λ
SE
λ
SE
Item 1 .58
(.05)
.72
(.04)
.82
(.09)
.67
(.05)
.60
(.06)
.67
(.05)
Item 2 .78
(.05)
.88
(.03)
.87
(.04)
.87
(.04)
.80
(.04)
.88
(.06)
Item 3 .57
(.04)
.71
(.04)
.66
(.04)
.66
(.04)
.59
(.04)
.59
(.06)
Autonomy Support*
Item 1 .78
(.02)
.87
(.04)
.83
(.04)
.75
(.07)
.84
(.04)
.75
(.07)
Item 2 .71
(.06)
.91
(.02)
.88
(.03)
.91
(.01)
.82
(.03)
.85
(.05)
Item 3 .91
(.02)
.93
(.01)
.91
(.02)
.93
(.02)
.91
(.02)
.95
(.02)
Item 4 .92
(.01)
.88
(.04)
.91
(.01)
.93
(.01)
.92
(.01)
.91
(.02)
Adolescent-Reported Capacity for Agency
Engagement with Challenge1.80
(.03)
.80
(.03)
.77
(.03)
.80
(.03)
.80
(.03)
.75
(.06)
Strategic Planning1.59
(.04)
.59
(.04)
.59
(.04)
.59
(.04)
.41
(.12)
.62
(.08)
Responsibility & Dependability1.78
(.03)
.78
(.03)
.78
(.03)
.78
(.03)
.78
(.03)
.80
(.06)
Advisor-Reports of Adolescent’s Capacity for Agency
Engagement with Challenge .96
(.01)
.95
(.01)
.96
(.01)
.95
(.01)
.96
(.01)
.93
(.04)
Strategic Planning .93
(.02)
.93
(.02)
.93
(.02)
.93
(.02)
.93
(.02)
.93
(.03)
Responsibility & Dependability .97
(.01)
.97
(.01)
.97
(.01)
.97
(.01)
.97
(.01)
.94
(.02)
Note. Loadings taken from final CFA model. *Two of the six items were dropped because they
failed to load on factor. 1Based on mean scores.
36
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
Table 3. Longitudinal CFA Results
Model
#
Free
Par.
-2LL
CFI
RSMEA
Value Scale
Factor Value Lower
CI
Upper
CI
p-
value
Panel A. Measurement Invariance
1.a Configural 357 -11,941.49 1.59 0.93 0.06 0.06 0.07 <.01
1.b Configural (11th T1 A-RS w/ T1 DIR-4) 358 -11,935.68 1.57 0.94 0.06 0.06 0.07 <.01
1.c Configural (11th T1 Y-EC w/ T1 A-EC) 359 -11,930.42 1.57 0.94 0.06 0.06 0.07 <.01
1.d Configural (10th T1 Y-RS w/ T2 Y-EC) 360 -11,924.62 1.56 0.94 0.06 0.05 0.07 <.01
2.a Metric 315 -11,976.46 1.64 0.93 0.06 0.06 0.07 <.01
2.b Metric (9th T1 AS-2) 316 -11,965.66 1.63 0.93 0.06 0.05 0.07 <.01
2.c Metric (9th T1 AS-1) 317 -11,957.23 1.63 0.94 0.06 0.05 0.07 0.01
2.d Metric (11th T1 Y-PN) 318 -11,953.10 1.63 0.94 0.06 0.05 0.07 0.01
2.e Metric (10th T1 DS-1) 319 -11,949.76 1.63 0.94 0.06 0.05 0.07 0.01
3.a Scalar 278 -11,978.08 1.69 0.94 0.06 0.05 0.06 0.03
3.b Scalar (10th T2 AS-1) 279 -11,975.10 1.68 0.94 0.06 0.05 0.06 0.03
4.a Residual Variance 220 -12,047.35 1.63 0.94 0.06 0.05 0.06 0.06
4.b Residual Variance (11th T2 A-RS) 221 -12,041.99 1.64 0.94 0.06 0.05 0.06 0.07
4.c Residual Variance (9th T2 AS-4) 222 -12,036.78 1.63 0.94 0.06 0.05 0.06 0.09
Panel B. Structural Homogeneity
5.a Factor Variance 202 -12,056.86 1.70 0.94 0.06 0.05 0.06 0.06
5.b Factor Variance (9th T1 DIR) 203 -12,052.17 1.69 0.94 0.06 0.05 0.06 0.09
6.a Factor Covariance 172 -12,075.34 1.78 0.94 0.06 0.05 0.06 0.10
7.a Factor Cross Lags 149 -12,083.85 1.91 0.94 0.05 0.05 0.06 0.11
8.a Factor Autocorrelations 141 -12,093.50 1.98 0.94 0.06 0.05 0.06 0.11
8.b Factor Autocorrelations (10th DIR) 142 -12,090.27 1.97 0.94 0.05 0.05 0.06 0.13
9.a Factor Means 122 -12,112.77 2.04 0.94 0.05 0.05 0.06 0.12
9.b Factor Means (11th T1 DIR)
9.c Factor Means (9th T1 A-AG)
123 -12,108.53 2.03 0.94 0.05 0.05 0.06 0.14
124 -12,105.60 2.06 0.94 0.05 0.05 0.06 0.14
10.a Base Regression Model 124 -12,105.12 2.04 0.94 0.05 0.05 0.06 0.15
Note. AS = youth reported autonomy support; DS = youth reported directive assistance; A-RS = Advisor
reports youth responsibility; A-EC = Advisor reports youth engagement with challenge; Y-EC = Youth
reported engagement with challenge; Y-PN = youth reported planning; A-AG = Advisor reports of youth
agency
37
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
Table 4. Standardized Parameter Estimates and Tests of Pathways among Latent Factors
-2LL
A. Cross-Sectional Correlations SE Value Scale
Factor p
DS <—> Y-AG 0.56 .07 -12146.38 2.08 < .01
DS <—> A-AG 0.25 .03 -12115.00 2.06 < .01
AS <—> Y-AG 0.56 .02 -12173.38 2.09 < .01
AS <—> A-AG 0.23 .06 -12118.84 2.04 < .01
DS <—> AS 0.55 .06 -12153.17 2.09 < .01
Y-AG <—> A-AG 0.51 .08 -12166.54 2.02 < .01
B. Autoregressive Paths
DS (9th and 11th grade groups) 0.18 .16 -12106.76 2.04 < .01
DS —10th only10.33 .13 -12108.96 2.05 <.01
AS 0.71 .03 -12138.80 2.08 < .01
Y-AG 0.44 .13 -12113.27 2.04 < .01
A-AG 0.54 .13 -12134.90 2.03 < .01
C. Cross-Lagged Regressive Paths
T1 DS T2 Y-AG -0.12 .06 -12106.29 2.05 0.06
T1 DS T2 A-AG 0.05 .06 -12105.38 2.04 0.45
T1 AS T2 Y-AG 0.21 .08 -12107.91 2.05 < .01
T1 AS T2 A-AG 0.14 .09 -12106.82 2.04 0.07
T1 DS T2 AS -0.19 .07 -12108.94 2.04 < .01
T1 AS T2 DS 0.27 .05 -12110.15 2.06 < .01
T1 Y-AG T2 DS 0.04 .18 -12106.74 2.03 0.22
T1 A-AG T2 DS 0.08 .08 -12107.12 2.04 0.08
T1 Y-AG T2 AS -0.09 .08 -12107.19 2.05 0.02
T1 A-AG T2 AS 0.07 .09 -12107.03 2.04 0.12
T1 A-AG T2 Y-AG 0.12 .08 -12106.87 2.04 0.05
T1 Y-AG T2 A-AG -0.05 .07 -12107.91 2.04 0.05
Note. Parameters tested using −2LL rescaled difference test with df = 1 (p < .01); 10.a base regression
model in Table 3 is reference model. AS = youth reported autonomy support; DS = youth reported
directive support; Y-AG = youth reported agency; A-AG = Advisor reports of youth agency. 1Non-
equivalent covariance required testing this path separately. Parameters in a panel above the dashed line
represent the a priori hypothesized paths.
38
PROGRAM LEADER SUPPORTS AND ADOLESCENT AGENCY
