Exploring Gender and Self-Efficacy Ratings of Athletic Training Students over Time

INTRODUCTION

Self-efficacy can be defined as a person's belief about his/her ability to successfully perform a given behavior. Albert Bandura¹ put forth a ground-breaking self-efficacy model as part of his social cognitive theory. In this model, self-efficacy is derived from 4 sources: (1) past accomplishments, (2) vicarious experiences, (3) verbal persuasion, and (4) emotional arousal. Past accomplishments are experiences in which the person mastered the designated task, while vicarious experiences are experiences in which someone observes a peer or mentor completing the task. An example of a previous accomplishment involves an individual producing a customized foot orthotic versus a vicarious experience of watching someone else produce the orthotic. Verbal persuasion requires social support from a credible source, and emotional arousal is an intrinsic reaction to the environment and task. According to this model, a student with high levels of efficacy will perform better than a student with low levels of efficacy, which leads to increasing levels of self-efficacy, giving way to higher performance.² Studies^3,4 have documented that students' increased efficacy may be a result of their education experiences.

Measures of self-efficacy need to be specific to a given task to increase the accuracy of predicted outcomes based upon a belief.² All too often, self-efficacy measures are global and lack specificity. Generalized self-efficacy instruments assess people's general confidence and abilities without specifying the tasks or situations.⁵ Self-efficacy research has focused on 2 main areas: (1) the relationship between efficacy and college major/career choices^6,7 and (2) the relationship between efficacy and motivation/achievement.^5,8–11 The first area of research provides insight into career development and gender differences. The second has supported that self-efficacy is correlated to various psychological constructs, such as self-regulation,¹ rewarding contingencies,¹⁰ and anxiety and self-concept.⁵ However, the manner in which self-efficacy has been measured varies widely, and direct comparisons between studies are complicated. The vast majority of self-efficacy–related research has focused on mathematics and reading/writing education, although the concept of self-efficacy has a place in health education. The instrument developed for this project used task-specific measures (ie, self-efficacy with a specific task/skill within the context of care for a real patient) as opposed to more global measures (ie, self-efficacy without a specific task/skill within the context of care real patient), and our purpose lends itself to the first research focus outlined above, in which self-efficacy is related to career development and gender differences.

The relationship between self-efficacy and gender has been studied extensively in the general area of science education.^12–20 A dubious distinction is held by the former president of Harvard University, Lawrence Summers, who in 2005 suggested in a conference presentation that “there are issues of intrinsic aptitude, and particularly of the variability of aptitude” for women in science.²¹ In general, the studies have determined that males display higher levels of self-efficacy than do females, and in accordance with Bandura's findings, they display higher levels of achievement in science education. While these findings are often statistically significant, it is important to note that the differences are typically small in practical terms. More study is needed in the field of athletic training.

A literature search within the field of “Athletic Training,” using self-efficacy and self-confidence as key words, has revealed that most studies fall into 1 of 3 distinct areas: (1) education of athletic training students,^22–36 (2) rehabilitation and return-to-play considerations,^37–45 and (3) confidence of athletic trainers in various practice settings with various populations.^46–53 Of the studies examining the education of athletic training students and discussing self-efficacy and/or self-confidence, several recorded gender but did not report a relationship between gender and self-efficacy and/or self-confidence.^23,27,29 Leaver-Dunn and colleagues²³ found no significant relationships between gender and self-confidence using the California Critical Thinking Disposition Inventory, which included a subscale of critical-thinking self-confidence. Weidner and Popp²⁷ studied peer-assisted learning but did not report any results related to gender and the effectiveness of peer-assisted learning or perceived benefits. Henning and colleagues²⁹ studied the prevalence of peer-assisted learning and the perceived benefits, such as increasing self-confidence, and reported no significant relationships between gender and self-confidence. All of these articles used interviews or 1-question survey methods to assess self-confidence or self-efficacy. To date, the authors are unaware of any research that has developed a reliable efficacy-measuring instrument that analyzed gender as a contributing factor in self-efficacy ratings. Athletic trainers play a critical role in the health care of physically active patients. If educators can develop interventions to increase students' perceptions of their efficacy, they may be able to increase the students' performance on a given task. In order to gauge the effectiveness of any intervention, it is necessary to obtain a baseline measure of the students' perception of efficacy with regard to a given task. The purpose of this project was to develop and assess the reliability of the Self-Efficacy in Athletic Training Student (SEATS) instrument and to explore changes by gender over the course of 1 year. Based upon the literature and our personal anecdotal observations, we hypothesized that males would display higher levels of self-efficacy. Further, we hypothesized that males would have greater increases in self-efficacy over a 1-year period.

METHODS

Participants

A convenience sample of undergraduate athletic training students (N = 30 students, 13 males and 17 females) ranging from 18 to 22 years in age and enrolled in a large state-sponsored Midwest university were solicited to participate. All students were enrolled in the professional phase of an athletic training program. There were 16 seniors (7 females and 9 males) and 14 juniors (10 females and 4 males). Table 1 depicts the course work they had completed or were enrolled in at the time of the study. No students were excluded from the project as all students provided consent. A post hoc power analysis using G*Power 3⁵⁴ with gender and time as the 2 independent variables and self-efficacy as the dependent variable produced a statistical power level of .963. The project was approved by the Human Subject Review Board, and informed consent was obtained.

Table 1.  Athletic Training Courses Enrolled or Completed During Year of Study by Academic Cohort

View Fullscreen

SEATS Instrument Validity and Reliability Procedure

The SEATS instrument was derived from a previously developed set of Learning Over Time (LOT) instruments.⁵⁵ The LOT instruments gave students the context of specific skills/activities with which to rate their self-efficacy. The LOT instruments were designed for completion by preceptors (formerly referred to as approved clinical instructors). To measure self-efficacy, we took the content areas (Injury Assessment, Preventative Injury Care, and Therapeutic Modalities) and components directly from the LOT instruments (see Table 2 for the components within each content area). This was done to allow for a future comparative study of instructor ratings of clinical ability and student ratings of self-efficacy. The face and content validity of the 3 content areas and associated components was established by the review of 3 education experts and 6 clinical staff members. Through several rounds of review, comments and suggestions were used to make modifications to the content areas and associated components.^55,56 As the SEATS instrument was derived from the LOT instrument to be used by preceptors, the reliability of the SEATS instrument needed to be established from the perspective of the student. An internal reliability analysis using Cronbach α was conducted with the data from a previous project to measure and establish the reliability of the SEATS instrument.

Table 2.  Descriptive Statistics at Each Measure and Friedman's 2-Way Analysis of Variance by Ranks; Significance Across the 3 Measures of Self-Efficacy Rating by Component for Males (n = 13)

View Fullscreen

RESULTS

Nonparametric Analysis

The dependent variables used in this study were single Likert scale items and are considered ordinal-level variables, not interval-level variables. Additionally, sample size was small, and variables were likely not fully normally distributed. Therefore, nonparametric analyses using Friedman's 2-way analysis of variance by ranks were conducted. Sixteen analyses were conducted for males, and 16 were conducted for females. The same Bonferroni-corrected α level was chosen (.05/16 = 0.003).

Table 2 illustrates the descriptive statistics at each measure and the analysis of variance significance level across the 3 measures for males. Males displayed significant increases in self-efficacy for only 3 components: Neurovascular (P = .015) of the Injury Assessment Content Area, Discuss Return to Play (P = .003), and Demonstrate Return to Play (P = .014) of the Preventative Injury Care Content Area.

Table 3 illustrates the descriptive statistics at each measure and the analysis of variance significance level across the 3 measures for females. Females displayed significant increases (P ≤ .003) in self-efficacy for all but 4 (Palpation and Special Tests of the Injury Assessment Content Area and Equipment Set-Up and Demonstrate Treatment of the Therapeutic Modalities Content Area) of the 16 components.

Table 3.  Descriptive Statistics at Each Measure and Friedman's 2-Way Analysis of Variance by Ranks; Significance Across the 3 Measures of Self-Efficacy Rating by Component for Females (n = 17)

View Fullscreen

Table 4 illustrates the pairwise post hoc comparisons of self-efficacy by component. Males had only 5 of the 16 pairwise comparisons that were statistically significant (P ≤ .003). None were statistically significant from measure 1 to 2. Three (Neurovascular, Discuss Return to Play, and End Treatment) of the 16 were statistically significant from measure 1 to 3. Two (Discuss Return to Play and Demonstrate Return to Play) of the 16 were statistically significant from measure 2 to 3. For females there were no statistically significant (P ≤ .003) increases in self-efficacy from measure 1 to 2 for the Injury Assessment Content Area and the Therapeutic Modalities Content Area. There were statistically significant increases for all but 1 component (Assessment of Injury) of the Preventative Injury Care Content Area. For females, 14 of the 16 components were statistically significant from measure 1 to 3. Six of the 16 components were statistically significant from measure 2 to 3 (Observation/Inspection, Neurovascular, and Assessment/Diagnosis of the Injury Assessment Content Area and Discuss Indications, Equipment Set-Up, and End Treatment of the Therapeutic Modalities Content Area).

Table 4.  Pairwise Comparisons of Self-Efficacy Ratings Using Friedman's 2-Way Analysis of Variance by Ranks; Significance by Component (N = 30)

View Fullscreen

DISCUSSION

Our results indicated that both time and gender were contributing factors to changes in perceptions of self-efficacy. One would expect that with time, as students learn and practice the various skill sets associated with our assessment, their perceptions of self-efficacy would change. We found significant changes in self-efficacy over the 3 measurement occasions with regard to gender. Female participants exhibited significant increases in self-efficacy across the 3 measure occasions for 12 of the 16 components. Conversely, male participants exhibited significant changes for only 3 components.

Martin⁵⁹ explored the relationship between gender and teachers' perceptions of self-confidence in their students' engagement and motivation. Male teachers' ratings of self-confidence were markedly higher than those of females for perceptions of their students' persistence and planning. We found that males had higher initial ratings of self-efficacy for 11 of the 16 (68%) components at measure 1 but for only 9 of the 16 (56%) components at measure 3. Females displayed greater average overall increased changes in self-efficacy across the 3 measures (.90 ± .36) when compared with males (.61 ± .32).

Kruger and Dunning⁶⁰ found that people tend to hold overly favorable views of their abilities in many domains. They found that people scoring in the bottom quartile on various tests grossly overestimated their performance and abilities. As they stated, “Not only do these people reach erroneous conclusions and make unfortunate choices, but their incompetence robs them of the metacognitive ability to realize it.” In other words, they are unconsciously ignorant about their lack of ability. This is a similar concept to that of The Reflective Practitioner, as discussed by Donald Schön.⁶¹ Schön⁶¹ suggested that people move from being unconsciously ignorant about a given task or skill through 2 intermediary stages to finally become unconsciously competent. This may explain why we found that males had greater initial levels of self-efficacy for the majority of components at measure 1 (68%) but not at measure 3 (56%). Perhaps, as time progressed between measures, the males began to realize their true abilities and adjusted their ratings of efficacy accordingly. Stewart et al⁶² explained that confidence must be tempered with knowledge of limitations, weaknesses, and competence.

Gardner-Medwin and Gahan⁶³ studied confidence-based assessments in which students rated their confidence in giving an answer to a question. They found no differences in ratings of confidence based upon gender. While this finding contradicts our results for the initial measure, it is supported by our results for the final measure, in which the difference between genders is negligible. Their findings are related to those of Kruger and Dunning⁶⁰ in that appropriate feedback is important: “The aim is to encourage reflection, self-awareness, and the expression of appropriate levels of confidence.”

Students require appropriate corrective feedback in order to properly assess their competence level and thus properly calibrate ratings of self-efficacy. With proper assessments of competence and self-efficacy students can self-determine areas of strength and weakness. Dependence upon student self-assessments alone is perilous at best. Research has shown that students' self-assessment of their abilities often differs from the assessment afforded by peer assessments⁶⁴ or actual skill demonstration.^65,66

Development of interventions to increase students' self-efficacy may improve the students' performance. The development of a reliable method of measuring self-efficacy will allow educators to determine a baseline of their students' perceptions of self-efficacy and, thus, the effectiveness of the interventions. The development and reliability analysis of the SEATS instrument was sound and based upon established scientific principles.⁶⁷ The content validity argument is based upon a review of experts in both the didactic and clinical settings of athletic training education. The estimates of reliability for the instrument were high and are thus acceptable.⁶⁷

CONCLUSIONS

As Bandura's self-efficacy model stated, increased levels of self-efficacy can lead to increased performance. However, Morgan and Cleave-Hogg⁶⁶ found no correlation between level of efficacy and performance on a standardized simulation test or clinical grades. Educators need to be aware of their students' self-efficacy on given skills, develop interventions to increase the ratings of self-efficacy, and perhaps improve performance on those skills. Our findings can be summarized as follows:

• Males had higher initial levels of self-efficacy than did females.

• Females had higher average increases in self-efficacy across time.

This study found that males have higher levels of self-efficacy but that this difference was attenuated as females' overall levels of self-efficacy increased more over time than did those of males, ultimately to similar levels. A reliable method for measuring levels of self-efficacy has been developed and tested for athletic training education on a defined set of skills. It is the hope of the authors that readers will adapt our instrument to their particular needs and use it as part of an overall assessment system.