Phase 1 Cross-Sectional Study Results

• In Event: Symposium | Testing a Dynamic and Iterative Process of Leadership Identity and Competency Development in Youth

Short Description

The purpose of this symposium presentation is to share the results from the initial cross-sectional analytic phase (Phase 1) within a comprehensive longitudinal examination of leadership identity and leadership competency development in youth. The Phase 1 study yielded results that suggest that progressing to longitudinal Phase 2 of the study is warranted.

Detailed Abstract

The emergent theory outlined by [Author] et al. (2020) led to a series of propositions that helped facilitate the creation of candidate models, addressing Hypothesis 1: Due to the hypothesized and tested relationships between leader identity and leadership skills (e.g., Day & Sin, 2011; Lord & Hall, 2005; Miscenko et al., 2017), it is expected that holding a stronger positive youth leadership identity will be associated with higher self-ratings of leadership competency.

Using the confirmatory factor analysis model for the Positive Youth Leadership Inventory (PYLI, [Author] & [Author], 2018, 2019) as the measurement model for leadership identity, structural equation models were formed for each of the four domains (Intrapersonal, Societal, Interpersonal, and Strategic) of the Student Leadership Competencies Youth Model ([Author], 2018). In these models, latent factors formed from individual competencies (each having items) were predicted by the general factor of the PYLI. Model selection was informed by theory posited by Hypothesis 1.

The models were estimated on a random sample of 4,000 youth stratified to be representative of the U.S. population with respect to gender, race and ethnicity, and rural/non-rural residence. Models were fit using the lavaan package in R (Rosseel, 2012), treating the data as ordinal with the WLSMV estimator. Due to the intentional over-sampling of population cross-sections with small proportions, those cases were weighted accordingly in the final analysis. Model fit was evaluated by comparing selected fit indices to recommended cuts. RMSEA was good to adequate for all models (min: 0.048, max: 0.072). The TLIs from the four models ranged from 0.926 to 0.966, indicated good fit for all models.

Given that all scales used in this study have been developed using CFA/EFA approaches, the parameter estimates for the measurement model for both the PYLI and Student Leadership Competencies are of secondary interest as they were consistent with results from prior studies. Conversely, the results from the structural portion of each model are of primary concern, as these results provide support for or against Hypothesis 1. For the model including the intrapersonal domain, there were four competencies that each were predicted by the general factor of the PYLI. The Wald test (asymptotic Z-test) for each parameter was positive and significant with p < .001, compared to an a priori alpha = 0.05. The model implied R^2 values for each competency ranged from 0.393 to 0.490, indicating large proportion of the variance in self-rated competencies was accounted for by leadership identity. The results for the other three domains closely mirror those from the analysis of the intrapersonal domain. That is, all regression paths in the structural portion of the model were significant. For the strategic domain, the R^2 values for the seven competencies ranged from 0.459 to 0.511. For the societal domain, the R^2 values for the two competencies were 0.400 and 0.442. For the interpersonal domain, the R^2 values for the seven competencies ranged from 0.379 to 0.551.

The results from the four structural equation models strongly support the hypothesis of an association between the PYLI and student leadership competencies. However, due to study limitations care should be taken in order to not over-interpret the results. First, these data were collected in a cross-sectional fashion and without any experimental manipulations. This means that there can be no inference made on causality and (equivalently) the direction of the effects. Indeed, the model in which competencies predict leadership identity would produce the exact same model fit. Second, there may be temptation to infer individual level characteristics from this population level model. That is, one may be attempt to infer that increasing student leadership identity leads to an increase in acquisition of leadership competencies in the same way for all students. Making such inferences from these data would require strict assumptions that development in leadership identity and leadership competency acquisition is an ergodic process (Molenaar, 2004). However, these assumptions have been shown to rarely hold in practice. That said, these limitations do not render the results of this Phase 1 study useless. It would be equally unlikely for a zero cross-sectional relationship to occur in conjunction with a dynamic system in which the two factors are associated (i.e. stochastically dependent).

Therefore, in light of both the results and limitations of this study, we make the following conclusions. First, if we desire to understand the dynamic process of change in leadership identity and leadership competencies, we need to collect intra-individual data (i.e. within a longitudinal study). Second, given the large cost and technical difficulty of conducting a longitudinal study, we first attempted to infer the presence of an inter-individual relationship. Finally, this Phase 1 study yielded results that suggest progressing into Phase 2 of the study is a worthwhile endeavor.

Participants

• Alex Brodersen, University of Notre Dame
  Co-Author

• Lindsay Hastings, University of Nebraska-Lincoln
  Presenter

• L.J. McElravy, University of Nebraska-Lincoln
  Presenter

• Corey Seemiller, Wright State University
  Presenter