Attitudes Toward Programming and Their Impact on C++ Academic Performance: An Exploratory Study

Attitudes Toward Programming and Their Impact on C++ Academic Performance: An Exploratory Study

Yahya Ibrahim, Nuridawati Mustafa, Kurk Wei Yi, Massila Kamalrudin

Fakulti Teknologi Maklumat dan Komunikasi

Universiti Teknikal Malaysia Melaka (UTeM), Durian Tunggal, Melaka, 76100, Malaysia

DOI: https://dx.doi.org/10.47772/IJRISS.2025.909000782

Received: 28 September 2025; Accepted: 04 October 2025; Published: 30 October 2025

ABSTRACT

This study investigated the relationship between students’ programming attitudes, gender, and performance in C++ courses. A total of 100 undergraduates completed a 25-item Programming Attitude Survey, and their scores were matched with final C++ examination results. Reliability analysis confirmed good internal consistency of the instrument, supporting its use for measuring programming-related perceptions. Descriptive findings indicated generally positive attitudes, though broad measures of attitude did not significantly predict programming performance. Item-level analyses revealed that specific positive attitudes, such as problem-solving persistence and enjoyment of programming, were positively correlated with C++ achievement, whereas negative orientations, such as reliance on memorization and feelings of helplessness, were linked to lower scores. Gender-based comparisons showed that male students reported more favourable attitudes and outperformed female peers in C++ assessments, suggesting that structural and contextual factors may contribute to these disparities. Regression analysis further revealed that gender was a significant predictor of programming performance, while programming attitudes alone did not explain substantial variance. These results highlight the importance of examining specific attitude dimensions rather than overall scores, while also addressing gender-related disparities in programming education. The findings contribute to ongoing discussions on factors influencing programming success, emphasizing the dual importance of nurturing adaptive problem-solving attitudes and implementing gender-sensitive instructional practices. Future research should explore pedagogical interventions, such as cognitive apprenticeships and active learning approaches, that may foster more equitable and effective programming learning outcomes.

Keywords: Programming Attitudes, C++ Education, Academic Performance, Exploratory Study, Computer Science Students

INTRODUCTION

Programming education remains a cornerstone of computer science and related disciplines, but success in this area depends on more than just mastering syntax or solving algorithmic problems. Increasingly, research shows that students’ attitudes, emotions, and motivational beliefs strongly influence their learning outcomes and academic performance [1], [2]. Positive attitudes such as self-efficacy, enjoyment, and interest have been consistently linked with better achievement, while negative emotions like anxiety and frustration can hinder engagement and persistence [3], [4].

Empirical evidence supports these perspectives across various contexts. For instance, Sun, Hu, and Zhou (2022) found that programming attitudes significantly predicted computational thinking skills among secondary school students, with notable gender differences—girls exhibited higher computational thinking performance but more negative attitudes [1]. Similarly, Wen, Wu, and Hsu (2023) reported that emotions such as enjoyment and anxiety mediated motivation and performance in introductory programming learning with Scratch, highlighting the role of affective factors [2]. At the tertiary level, game-based learning approaches have been shown to enhance students’ attitudes and interests in basic programming courses [3], while physical computing applications such as Arduino foster more positive learning experiences [4].

Other studies highlight the predictive power of self-efficacy and early programming achievement. A recent ACM conference paper (2024) demonstrated that students’ C programming grades and self-efficacy strongly predicted performance in subsequent object-oriented programming using Java [5]. Likewise, a study of 495 medical-field students learning C++ used structural equation modelling to show that attitudes (willingness, negativity, necessity) significantly influenced computational thinking perspectives and programming empowerment; students reporting negative (forced) attitudes exhibited lower computational-thinking perspectives and empowerment [6]. Moreover, large-scale analyses indicate that gender differences in attitudes and emotions may influence programming outcomes, though achievement gaps are inconsistent and shaped by instructional context [7].

Despite this growing body of evidence, limited research focuses specifically on C++, a language widely used in universities but known for its steep learning curve. Its syntax complexity and abstraction present unique challenges that may magnify the effects of attitudes on student success. This study aims to address this gap by examining the relationship between 25 dimensions of programming attitudes, gender, and C++ academic performance (measured by examination score T100). The research objectives are threefold:

1. To examine students’ attitudes toward programming and their overall academic performance in C++.
2. To investigate the relationship between programming attitudes and C++ academic performance.
3. To evaluate the influence of gender and programming attitudes on C++ performance.

RELATED WORK

Recent studies have highlighted that students’ programming attitudes, emotions, and learning environments play a crucial role in shaping their programming performance. For instance, Kovari and Katona [8] examined the effect of a software development course on students’ programming self-efficacy. They reported that practice-oriented learning improved problem-solving attitudes and algorithmic thinking, while low self-confidence negatively influenced outcomes.

Şenol Saygıner and Tüzün [9] compared block-based and text-based programming environments among novice learners. Their findings indicated that block-based training enhanced motivation and logical thinking more effectively than text-based instruction, suggesting that attitudes toward programming may differ depending on the learning medium.

Jiang, He, and Yan [10] investigated the factors influencing programming anxiety and found that higher anxiety was linked to lower motivation and performance, whereas confidence and positive emotions improved outcomes. This highlights the emotional dimension of programming attitudes and their connection to academic success.

Malkoc et al. [11] explored gender disparities in an online flipped programming course and found that male students generally achieved higher outcomes than females. Engagement and attitude factors significantly influenced performance, reinforcing the importance of considering gender differences in programming education research.

Finally, Yusuf, Md Noor, and Román-González [12] analysed block-based programming interaction patterns in relation to computational thinking. They showed that gender, spatial ability, and proficiency shaped how learners engaged with programming tasks, and these differences were predictive of performance. Table I shows summary of recent studies on programming attitudes, emotions, and performance.

Table I. Summary of Recent Studies on Programming Attitudes, Emotions and Performance

METHODOLOGY

Research Design

This study employed a quantitative exploratory research design to investigate the relationship between students’ programming attitudes and their academic performance in C++. Data were collected using a structured survey instrument that measured 25 attitude dimensions, supplemented with demographic information (gender) and students’ final examination scores in C++ (T100) [13].

Participants

The dataset consisted of undergraduate students enrolled in C++ programming course at Universiti Teknikal Malaysia Melaka (UTeM). A total of N = 95 valid responses were retained after data cleaning and removal of incomplete entries. Gender distribution was recorded as 52 male and 43 female, ensuring adequate representation for comparative analysis.

Instruments

The primary instrument was the Computing Attitude Survey, which comprised 25 items designed to measure students’ perceptions, beliefs, and emotions toward programming. Each item was rated on a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree). The scale covered multiple dimensions, including self-efficacy, motivation, anxiety, willingness, and perceived usefulness. In addition, students’ gender and C++ academic performance score (T100) were collected.

Data Preparation

Raw data were imported from Microsoft Excel into Python and R for statistical analysis. Preliminary checks included handling missing values, identifying invalid responses, and coding categorical variables (e.g., gender: 0 = male, 1 = female). Negatively worded survey items were reverse coded to maintain consistency in attitude direction.

Data Analysis

The internal consistency of the 25-item programming attitude scale was first assessed using Cronbach’s alpha to ensure reliability of the instrument. Descriptive statistics, including means and standard deviations, were then computed to provide an overview of students’ attitudes toward programming and their academic performance in C++.

To explore relationships between attitudes and performance, Pearson product–moment correlations were calculated. Both the composite attitude score and item-level correlations were examined to identify whether specific beliefs about programming were more strongly linked to performance outcomes. In addition, independent-samples t-tests were conducted to compare male and female students’ programming attitudes and C++ performance, thereby assessing potential gender differences.

Finally, a multiple regression analysis was carried out to determine whether programming attitudes and gender jointly predicted C++ performance. This model allowed for the examination of the unique contribution of each predictor while controlling for the other. Scatterplots with fitted regression lines were also generated to visually illustrate the relationships between attitudes and performance, as well as the interaction between gender and performance outcomes.

RESULT

Reliability and descriptive statistics

The 25-item attitude scale showed acceptable internal consistency (Cronbach’s α = 0.819), so items were combined into a composite attitude score by taking the item mean for each participant.

Descriptive statistics for the main variables are shown in Table II. On average, participants reported moderately positive attitudes toward programming (M = 3.51, SD = 0.41; possible range 1–5). The mean C++ score (T100) was 56.58 (SD = 19.92; possible range 0–100).

Table II. Descriptive Statistics for Attitude and C++ Performance

Correlations between attitudes and C++ performance

To further examine the link between specific attitudes and academic performance, item-level correlations between each survey statement and C++ performance were computed.

Table III. Pearson correlation between programming attitude and C++ performance

To explore which specific programming attitudes were most strongly associated with academic performance in C++, item-level correlations were examined (see Table III). Results revealed that the strongest positive associations with exam performance emerged for items reflecting enjoyment and confidence in problem solving. Students who reported that they enjoy solving programming problems (r = 0.36, p < 0.001), can usually figure out a way to solve programming problems (r = 0.35, p < 0.001), and find the challenge of solving programming problems enjoyable (r = 0.32, p = 0.002) tended to achieve higher C++ scores. Other positively correlated items included relating programming study to one’s life (r = 0.28, p = 0.007) and expressing general interest in learning more about programming (r = 0.27, p = 0.009). These findings suggest that students who view programming as enjoyable, personally meaningful, and solvable through persistence are more likely to succeed academically in C++.

In contrast, several items showed significant negative correlations with C++ performance. Students who agreed with statements emphasizing memorization or limited problem-solving strategies tended to perform worse. For instance, endorsing “After I study a topic in Programming C++ and feel that I understand it, I have difficulty solving problems on the same topic” was associated with the lowest performance (r = –0.40, p < 0.001). Similarly, higher endorsement of items such as “If I get stuck on a programming problem, there is no chance I’ll figure it out on my own” (r = –0.35, p < 0.001), “To learn programming, I only need to memorize solutions to sample problems” (r = –0.29, p = 0.005), and “There is usually only one correct approach to solving a programming problem” (r = –0.27, p = 0.008) were linked with lower scores. These patterns indicate that rigid or defeatist beliefs about programming may undermine students’ ability to perform well in C++.

Taken together, the item-level analyses reveal that while the composite attitude scale showed only a weak and nonsignificant correlation with C++ performance, certain individual attitudes were more diagnostic. Specifically, positive problem-solving orientations aligned with higher achievement, whereas reliance on memorization and lack of persistence predicted poorer outcomes.

Gender comparisons

Independent-samples t-tests were conducted to examine gender differences in programming attitudes and C++ performance. Results revealed a statistically significant difference in C++ scores, with male students (M = 60.62) scoring higher than female students (M = 51.70), t (df ≈ 91) = 2.22, p = 0.029 (see Table IV). Similarly, male students reported more positive attitudes toward programming (M = 3.61) compared to female students (M = 3.39), t (df ≈ 91) = 2.78, p = 0.007. These findings indicate that, within this sample, both programming attitudes and academic performance favoured male students.

Table IV. Independent-samples t-tests comparing males and females on C++ performance and programming attitudes

Predicting C++ Performance: Multiple Regression

A multiple linear regression was conducted to examine whether programming attitudes and gender predicted students’ C++ performance (T100). The overall model was statistically significant, F(2, 92) = 3.40, p = 0.038, and explained approximately 7% of the variance in performance (R² = 0.07).

As shown in Table V, programming attitudes did not significantly predict C++ scores, B = –4.83, t(92) = –0.96, p = 0.339. However, gender emerged as a significant predictor, B = –9.78, t(92) = –2.18, p = 0.032, indicating that female students scored lower on average than male students after controlling for programming attitudes. These findings suggest that while attitudes toward programming were not directly associated with performance, gender differences contributed to the prediction of C++ outcomes.

A graphical depiction of the regression results is provided in Figure I. The scatterplot shows fitted regression lines for males and females, illustrating that performance scores tended to be higher among males across the range of attitude scores, whereas attitudes themselves did not strongly differentiate performance outcomes.

Table V. Multiple regression predicting C++ performance (T100) from programming attitude and gender.

Fig: 1 Regression of C++ Performance on Attitudes by Gender

DISCUSSION

The first aim of this study was to examine the reliability and descriptive patterns of students’ attitudes toward programming. The 25-item scale demonstrated acceptable internal consistency, confirming its usefulness for analyzing programming perceptions. On average, students reported moderately positive attitudes toward programming. However, the lack of a strong overall correlation between composite attitudes and C++ performance suggests that broad measures may mask the influence of specific beliefs. Recent research similarly shows that learning outcomes depend less on general attitudes and more on targeted dimensions, such as self-efficacy and engagement [14].

The item-level correlation analyses provided greater insight. Positive orientations, such as enjoying problem solving, persisting through challenges, and relating programming to personal relevance were linked with higher C++ performance. In contrast, beliefs emphasizing memorization, rigid problem-solving approaches, or helplessness were associated with poorer outcomes. These findings align with Laitinen et al. [15], who found that students with higher self-efficacy tended to adopt deeper learning approaches, while surface strategies like rote learning correlated with weaker achievement. Thus, attitudes rooted in persistence and confidence may act as protective factors in programming education.

Gender comparisons further highlighted disparities. Male students both reported more favorable attitudes and performed better in C++ assessments than female students. Importantly, regression analyses showed that gender remained a significant predictor even after accounting for attitudes, underscoring structural influences that go beyond individual beliefs. Childs [16], for example, found that interventions such as pair programming for primary-aged girls produced limited attitudinal changes, suggesting that gender gaps are shaped by broader contexts like prior exposure, confidence, and socialization. This reinforces the need for gender-responsive teaching strategies in programming courses.

Finally, the regression model indicated that while attitudes did not significantly predict C++ performance, gender contributed meaningfully to the variance explained. This modest explanatory power echoes findings from Chen et al. [14], who demonstrated that instructional design factors (e.g., collaborative or immersive approaches) play a substantial role in shaping programming outcomes. Together, these findings suggest that improving performance requires a multifaceted approach: cultivating adaptive problem-solving attitudes, addressing gender disparities, and enhancing learning environments through innovative pedagogical strategies.

CONCLUSION

This study examined the interplay between students’ programming attitudes, gender, and performance in C++ courses. The results demonstrated that while the overall attitude scale did not strongly predict programming achievement, specific attitudes such as enjoyment of problem solving, persistence, and confidence were positively associated with higher performance. In contrast, reliance on memorization and perceptions of helplessness negatively affected outcomes. These findings suggest that focusing on cultivating adaptive problem-solving orientations may be more beneficial than simply measuring general attitudes.

The study also revealed significant gender differences, with male students reporting more favorable programming attitudes and achieving higher scores than their female counterparts. Regression analysis confirmed gender as a significant predictor of performance, underscoring the need to address gender disparities in programming education through targeted support and inclusive teaching strategies.

Overall, this research highlights the importance of moving beyond global measures of attitude to examine specific belief patterns that directly shape learning outcomes. Educators and curriculum designers should consider integrating problem-solving–oriented practices and gender-sensitive approaches such as pair-programming in order to foster both equity and effectiveness in programming instruction. Future studies could expand on these findings by investigating larger and more diverse samples, as well as by testing interventions designed to enhance adaptive programming attitudes across different learning contexts.

ACKNOWLEDGEMENT

The authors would like to express gratitude to Fakulti Teknologi Maklumat dan Komunikasi (FTMK), Universiti Teknikal Malaysia Melaka (UTeM)  for their invaluable support and resources provided throughout this research.

REFERENCES

1. Sun, L., Hu, L., & Zhou, D. (2022). Programming attitudes predict computational thinking: Analysis of differences in gender and programming experience. Computers & Education, 181, 104457. https://doi.org/10.1016/j.compedu.2022.104457
2. Wen, F.-H., Wu, T., & Hsu, W.-C. (2023). Toward improving student motivation and performance in introductory programming learning by Scratch: The role of achievement emotions. SAGE Open, 13(4), 21582440231205985. https://doi.org/10.1177/00368504231205985
3. Quah, W. B., Banua, N. M., & Zulkifle, Z. (2023). Game-based learning for basic programming courses: Discover students’ attitudes and interests. Journal of ICT in Education, 10(1), 2. https://doi.org/10.37134/jictie.vol10.1.2.2023
4. Zaini, Z. H., Arshad, N. I., Singh, B. S. M., et al. (2024). A study on student attitudes in learning programming using physical computing. Dinamika: Jurnal Ilmiah Pendidikan Dasar, 12(1), 36–44. https://doi.org/10.30595/dinamika.v12i1.6499
5. Alhassan, M., & Alhassan, A. (2024). Assessing the influence of self-efficacy and C programming grades as predictors of object-oriented programming performance in Java. In Proceedings of the 2024 10th International Conference on Education and Training Technologies (pp. 123–129). ACM. https://doi.org/10.1145/3661904.3661911
6. Lai, C.-H., Chen, Y.-K., Wang, Y.-h., & Liao, H.-C. (2022). The study of learning computer programming for students with medical fields of specification—An analysis via structural equation modeling. International Journal of Environmental Research and Public Health, 19(10), 6005. https://doi.org/10.3390/ijerph19106005
7. Atman Uslu, N., & Onan, A. (2025). Exploring students’ emotions towards programming: Analysing sentiments using concurrent conversion mixed methods. Education and Information Technologies, 30, 19879–19909. https://doi.org/10.1007/s10639-025-13482-z
8. Kovari, A., & Katona, J. (2023). Effect of software development course on programming self-efficacy. Education and Information Technologies, 28(9), 10937–10963. https://doi.org/10.1007/s10639-023-11617-8
9. Şenol Saygıner, & Tüzün, H. (2022). The effects of block-based visual and text-based programming training on students’ achievement, logical thinking skills, and motivation. Journal of Computer Assisted Learning, 38(2), 644–658. https://doi.org/10.1111/jcal.12771
10. Jiang, Y., He, W., & Yan, Z. (2024). A study of factors influencing programming anxiety among non-computer science students. In Proceedings of the 2024 International Conference on Advanced Learning Technologies (pp. 210–217). ACM. https://doi.org/10.1145/3664934.3664956
11. Malkoc, S., Steinmaurer, A., Gütl, C., Luttenberger, S., & Paechter, M. (2024). Coding decoded: Exploring course achievement and gender disparities in an online flipped classroom programming course. Education Sciences, 14(6), 634. https://doi.org/10.3390/educsci14060634
12. Yusuf, A., Md Noor, N., & Román-González, M. (2024). Interaction patterns during block-based programming activities predict computational thinking: Analysis of the differences in gender, cognitive load, spatial ability, and programming proficiency. IntechOpen Journal. https://doi.org/10.5772/acrt.36
13. Ibrahim, Y. (2025). GenderPlusComputingAttitudeSurvey (v1.0) [Microsoft Excel file]. Zenodo. https://doi.org/10.5281/zenodo.17221584
14. Chen, C. M., Chang, Y.-H., & Wang, P.-Y. (2024). Enhancing programming learning performance through metaverse-based Jigsaw collaboration. STEM Education Journal, 8, Article 495. https://doi.org/10.1186/s40594-024-00495-2
15. Laitinen, S., Christopoulos, A., & Laitinen, P. (2024). Relationships between self-efficacy and learning approaches among computer science students. Frontiers in Education, 9, 118161. https://doi.org/10.3389/feduc.2024.1181616
16. Childs, K. (2024). Investigating the impact of introducing pair programming to primary-aged girls: A randomized controlled trial. International Journal of Computer Science Education Studies, 12(2). https://ijcses.org/index.php/ijcses/article/view/218