Predictors Of Java Programming Self Efficacy Among Engineering Students In A Nigerian University

The study examined the relationship between Java programming self-efficacy and programming background of engineering students in a Nigerian University. One hundred and ninety two final year engineering students randomly selected from six engineering departments of the university participated in the study. Two research instruments: Programming Background Questionnaire and Java Programming Self-Efficacy Scale were used in collecting relevant information from the subjects. The resulting data were analyzed using Pearson product correlation and Multiple regression analysis. Findings revealed that Java Programming self-efficacy has no significant relationship with each of the computing and programming background factors. It was additionally obtained that the number of programming courses offered and programming courses weighed scores were the only predictors of Java self-efficacy.

💡 Research Summary

The study investigated how engineering students’ background in computing and programming relates to their self‑efficacy in Java programming at a Nigerian university. A total of 192 final‑year students were randomly selected from six engineering departments (Electrical & Electronic, Mechanical, Chemical, Civil, Computer & Software, and Architectural Engineering). Data were collected using two instruments: a Programming Background Questionnaire (PBQ) that captured variables such as computer usage frequency, prior exposure to programming languages, participation in projects, the number of programming courses taken, and a weighted score reflecting grades in those courses; and a Java Programming Self‑Efficacy Scale (JPSES) consisting of ten items measuring confidence in tasks like understanding syntax, debugging, and designing Java applications. Both instruments demonstrated high internal reliability (Cronbach’s α > 0.85).

Descriptive statistics confirmed a balanced sample in terms of gender and department representation. Pearson product‑moment correlations were first computed to explore bivariate relationships between each background variable and JPSES scores. Contrary to many prior studies, none of the computing experience variables (e.g., weekly computer use, operating‑system knowledge) nor the programming experience variables (e.g., number of previously learned languages, number of projects completed) showed a statistically significant correlation with Java self‑efficacy (correlation coefficients ranged from –0.03 to 0.09, all p > 0.05). This suggests that mere exposure to computers or prior programming activities does not automatically translate into higher confidence when tackling a specific language such as Java.

To identify which factors actually predict Java self‑efficacy, a multiple regression analysis was performed using a stepwise entry method. The independent variables entered were: (1) computing experience, (2) programming experience, (3) the total number of programming courses taken, and (4) the weighted programming‑course score (a composite of grades and credit weight). The final regression model retained only two predictors: (a) the number of programming courses completed (β = 0.31, p < 0.01) and (b) the weighted programming‑course score (β = 0.27, p < 0.01). The model’s adjusted R² was 0.22, indicating that roughly 22 % of the variance in Java self‑efficacy is explained by these two variables, while the remaining 78 % is attributable to other unmeasured influences.

The authors interpret these findings to mean that quantitative exposure to programming (i.e., taking more courses) and qualitative success within those courses (higher grades) are the primary levers for boosting confidence in Java. They argue that curriculum designers should consider expanding the breadth of programming offerings and implementing support mechanisms—such as timely feedback, scaffolded assignments, and grade‑enhancement strategies—to improve students’ performance, thereby indirectly raising self‑efficacy.

Several limitations are acknowledged. First, the sample is confined to a single university in Nigeria, which restricts the external validity of the results. Second, reliance on self‑report questionnaires may introduce social desirability bias and measurement error, especially since the psychometric validation of the instruments is not described in depth. Third, the modest explanatory power of the regression model signals that important psychological constructs (e.g., intrinsic motivation, goal orientation, self‑regulation, instructor feedback) were omitted. The authors recommend future research to adopt a mixed‑methods approach, incorporate a broader set of psychosocial variables, and draw on multi‑institutional, cross‑cultural samples to develop a more comprehensive model of programming self‑efficacy.

In sum, the paper contributes a nuanced perspective to the literature on computing education in developing‑country contexts: while general computing and programming experience do not predict Java self‑efficacy, the structural aspects of the academic program—specifically the number of programming courses taken and the grades earned in those courses—play a decisive role. This insight offers practical guidance for educators seeking to cultivate confidence and competence in specific programming languages among engineering undergraduates.