37 Graduate Student Journal of Psychology Spring 2024 - Vol. 22 Copyright 2024 by the Department of Counseling and Clinical Psychology Teachers College, Columbia University Changes in Undergraduate Students’ Self-Efficacy and Outcome Expectancy in an Introductory Statistics Course Milushka Elbulok-Charcape1, Evan Grandoit2, Anjali Krishnan1,3, Rose Bergdoll3, Joshua Fogel4, & Laura Rabin3 1Psychology Department, City University of New York Graduate Center, 2Psychology Department, Northwestern University, 3Psychology Department, City University of New York Brooklyn College, & 4Management, Marketing, and Entrepreneurship Department, City University of New York Brooklyn College The exploration of psychological variables that potentially impact college student performance in challenging ac- ademic courses can be useful for understanding success in introductory statistics. Although previous research has examined specific beliefs that students hold about their abilities and future outcomes, the current study is novel in its examination of changes in both self-efficacy (SE) and outcome expectancy (OE) in relation to performance over the course of an undergraduate introductory psychology statistics course. These psychological variables—relating to one’s belief about one’s ability to accomplish a task and the anticipated outcomes—may impact student motivation and performance. Students’ SE, OE, and other variables related to statistics performance were measured through a survey administered at the beginning and end of the course. Multivariate logistic regression and McNemar tests were conducted to examine factors that affected changes in SE and OE as the semester progressed. Students with lower scores on the final exam demonstrated a decrease in both high SE and positive OE. However, higher scores on exams earlier in the course were associated with increased odds for high SE but not for positive OE, suggesting that SE is less resilient to course performance. Based on these findings, the authors recommend that statistics instructors identify students at risk for decreasing SE. Instructors can help foster high SE in students struggling academically by connect- ing the course content to their everyday lives and suggesting strategies to enhance their confidence in their content knowledge and increase their comfort in navigating such a challenging course. Keywords: statistics education, research, self-efficacy, outcome expectancy, statistics Statistics is a mathematics-based course required by many undergraduate degree programs (Norcross et al., 2016). A statistics course is fundamental for stu- dents because it fosters critical thinking and reasoning skills (Cheng et al., 2018; Vanhoof et al., 2011; Wilson, 2013). Students of statistics courses become informed consumers of knowledge (Barber, 2002), and are equipped with the ability to evaluate and conduct anal- yses of research. As society’s reliance on complex data increases, the ability to analyze large sets of data and ap- ply statistical skills and techniques is required for those entering the workforce (Brown & Kass, 2009). As a result, enrollment in statistics courses from 2010-2015 increased by 19% (Blair et al., 2018) and employment opportunities for statisticians are projected to grow by 30% from 2018 to 2028, outpacing the rate of many other professions (The Bureau of Labor Statistics, 2020). Given that statistics courses carry such weight for students’ academic trajectory and their eventual en- try into the workforce, research devoted to variables af- fecting academic performance in undergraduate intro- ductory statistics courses has received much attention. Diverse student characteristics and behaviors have been researched in relation to statistics course performance. Studies focusing on the relationship between sex and statistics courses report that females hold more negative perceptions about statistics (Cen- dales et al., 2013; Zimprich, 2012) and that statistics anxiety may manifest differently in females com- pared to males. For example, procrastination—the tendency to delay initiation of academic tasks—and poorer learning strategies, such as memorization, were found more likely to lead to statistics anxiety in men (Rodarte-Luna & Sherry, 2008). Students with more experience and past achievements in sta- tistics or related fields (e.g., mathematics) have higher achievement in statistics courses (Johnson & Kuen- nen, 2006; Ramirez et al., 2012). Academic procras- tination (Schraw et al., 2007), is negatively correlated with statistics engagement (Onwuegbuzie, 2004) and academic performance (Goroshit, 2018; Paechter et al., 2017; Rodarte-Luna & Sherry, 2008; Wang & En- glander, 2010). Research on help-seeking behavior, the tendency to seek academic assistance (Newman, 2002), shows that the absence of help-seeking behav- ior may be related to predictors leading to lower sta- tistics performance (Rodarte-Luna & Sherry, 2008). Psychological factors such as motivation, student beliefs, attitudes, and emotions are particularly im- portant as they can impact academic performance in 38 ELBULOK-CHARCAPE ET AL. statistics courses. Undergraduate students who were motivated had better exam grades and valued statistics more than those not motivated (Budé et al., 2007). Students’ beliefs and attitudes about their knowledge of statistics do not solely impact their course perfor- mance (Chiesi & Primi, 2009; Dempster & McCorry, 2009; Ramirez et al., 2012), but can also disrupt learn- ing and application of statistical concepts outside of the classroom (Gal & Ginsberg, 1994). Even in class- es utilizing diverse teaching styles (e.g., lecture-based versus active learning), ambivalent attitudes toward statistics remain fairly stable throughout the semester (Bateiha et al., 2020). Finally, some emotional states such as fear (Slootmaeckers et al., 2014) and anxiety (Cui et al., 2019; Davis & Mirick, 2015; Esnard et al., 2021; Trassi et al., 2022) towards statistics nega- tively affect student academic performance in statis- tics. These emotional states impact cognitive factors, which ultimately alter students’ expectations of statis- tics courses, effort exerted, and academic performance. Specific cognitive variables and their positive relation- ship to academic performance in statistics includ- ing self-regulation (Acee & Weinstein, 2010; Dunn, 2014), the use of cognitive and meta-cognitive strate- gies (Rodarte-Luna & Sherry, 2008; Trogden & Royal, 2019), and analytical skills (Miller, 2019) have been in- vestigated. Among these cognitive variables, one of par- ticular importance is students’ self-efficacy in statistics. Self-efficacy is defined as an individual’s perceived ability to accomplish a designated task (Bandura, 1977) and is believed to be domain-specific (Bandu- ra, 1997; Pajares, 2009). Self-efficacy as a predictor of academic success in statistics has yielded conflicting findings. While most studies (Finney & Schraw, 2003; Mantooth et al., 2020; McGrath et al., 2015) report high self-efficacy change to be a strong predictor of ac- ademic success in statistics, other studies failed to find a significant association between increased self-effica- cy and success in statistics (Mihai-Bogdan et al., 2015; Olani et al., 2011; Walker & Brakke, 2017). In one study, student self-efficacy for statistics increased as the semester progressed (Walker & Brakke, 2017). Howev- er, self-efficacy alone does not contribute to achieve- ment in a statistics course (Franceschini et al., 2014). Outcome expectancy refers to beliefs about an- ticipated outcomes of action (Bandura, 1986; 1997); these expectancies are mechanisms thought to in- fluence performance and behavior (Bandura, 1986). Self-efficacy is also positively associated with outcome expectancy (Williams, 2010). Many studies exploring outcome expectancy focus on risky behavior such as smoking (Glock et al., 2012; Pokhrel et al., 2014; Urbán, 2010), drinking (Lopez-Vergara et al., 2012; Read, 2012), safe sex (Newby et al., 2013), and gam- bling (St-Pierre et al., 2013). One study focusing on academic aspects of outcome expectancy found that increasing high self-efficacy and outcome expectancy in low-performing students improved undergraduate students’ English course performance (Elborolosy & Al Thenyan, 2020). Another study on high school students enrolled in science and engineering-technol- ogy courses reported that students’ self-efficacy and outcome expectancy directly impacted their course performance (Han et al., 2021). Notably, student course performance was also impacted by the teachers’ self-efficacy and outcome expectancy regarding their own teaching ability—suggesting that student per- formance in other Science, Technology, Engineering, and Mathematics (STEM) courses, such as statistics, could also be enhanced by improving their confidence in their own ability and outcome expectation related to those courses. Another study found that Chinese el- ementary school students’ increased STEM stereotype beliefs negatively predicted their STEM self-efficacy and positive outcome expectancy. Additionally, their STEM self-efficacy and positive outcome expectancy predicted their STEM career interest (Luo et al., 2021). Three studies have focused on outcome expectan- cy concerning statistics courses, although they did not study change in outcome expectancy. In one study, outcome expectancy was influenced by perceived con- trollability and perhaps unsurprisingly, negative out- come expectancy had a negative association with ef- fect on first-year students enrolled in a health sciences statistics course (Budé et al., 2007). In another study, significant predictors of achievement in a psychology statistics course were past academic performance in statistics and positive expectancy (Hood et al., 2012). Finally, self-efficacy has been found to directly affect expectations for performance in statistics (Esnard et al., 2021). Further understanding the association of change in outcome expectancy with other vari- ables and statistics course performance is warranted. While numerous studies focus on self-efficacy in relation to undergraduate introductory statistics courses, few focus on outcome expectancy about this 39 CHANGES IN SE AND OE IN A STATISTICS COURSE course, and the authors are not aware of any study that focuses on both psychological constructs. Therefore, the current study tracks changes in undergraduate students’ academic self-efficacy and outcome expec- tancy while enrolled in an undergraduate introduc- tory psychology statistics course. We study how these constructs change according to students’ academic performance over the course of an academic semester. We include other relevant variables shown to impact academic performance in a statistics course of sex, college year, help-seeking behavior, and academic pro- crastination as covariates. We predict that students’ outcome expectancy will be more resilient to change than their self-efficacy based on their performance in a statistics course and associated covariates. We pre- dict that self-efficacy will change more than outcome expectancy, due to self-efficacy previously being asso- ciated with course performance (Hii et al., 2013; Mc- Grath et al., 2015; Waples, 2016). If our prediction is accurate, these results will support the conclusion that self-efficacy and outcome expectancy are not uniformly affected throughout a statistics course, in which students can begin the course with many pre- conceived fears and anxiety about the topic (Cui et al., 2019; Davis & Mirick, 2015; Esnard et al., 2021; Sloot- maeckers et al., 2014). As such, the use of different interventions to address these constructs in students enrolled in statistics courses would be warranted. Methods Participants Participants were undergraduate students en- rolled in introductory psychology statistics courses at an urban public college that is part of a large public university system in the Northeast United States. All students who attended class during the first week of the semester were invited to participate. The study was conducted in three introductory psychology statistics course sections, each from a different academic semes- ter. Participation entailed completion of study pre- and post-questionnaires at week 1 (beginning) and week 15 (end) of the semesters. Participants were not compensated nor were they penalized for non-partici- pation. There were 341 students initially enrolled, and 286 students participated for an 83.9% response rate. The study was ethically conducted and received IRB approval. All participants provided informed consent. Procedure The course instructor and two research as- sistants created the self-report questionnaire used in the current study, which inquired about stu- dent demographics, academic self-efficacy, aca- demic outcome expectancy, academic help-seek- ing behavior, and academic procrastination. Psychological Measurements Academic self-efficacy was assessed by self-report response to the statement, “I am [was] quite capable of mastering the material in this class.” Academic outcome expectancy was assessed by self-report re- sponse to the statement, “I will [would] never [be able to] do well in this class” where disagreeing indi- cates positive outcome expectancy. The past tense of the verb [words included in brackets] was used for the follow-up questionnaire. Both statements utilized a 4-point Likert-type scale with response options of 1 = strongly agree, 2 = somewhat agree, 3= somewhat disagree, and 4 = strongly disagree. The self-efficacy item used in this study was influ- enced by the statistics self-efficacy assessment developed by Hall and Vance (2010), for which they reported a reliability coefficient of 0.92. The statistics self-efficacy item used in this study was inspired by the question in their assessment “How confident are you with solving statistical problems?”. To reduce the time and effort for students to respond to the entire survey used in this study, the full statistics self-efficacy assessment by Hall and Vance (2010) was not adapted. Additionally, the number of scale responses for the adapted self-efficacy item was changed from five to four so the response va- lence would not be obscured by a neutral answer option. The outcome expectancy item used in this study was adapted from the statistics outcome expectan- cy assessment created by Hood et al. (2012). Their assessment question, “I expect to do well in this re- search methods and statistics course”, inspired the outcome expectancy item included in this study. To reduce the time and effort for students to respond to the entire survey used in this study, the full statistics outcome expectancy assessment by Hood et al. (2012) was not adapted. The response scale for the adapted outcome expectancy item was reduced from a 7-point to a 4-point response scale to match the self-efficacy item. Hood et al. (2012) reported a Cronbach’s al- pha of 0.69 for their outcome expectancy assessment. Variability in participants’ responses to the ac- 40 ELBULOK-CHARCAPE ET AL. ademic self-efficacy and outcome expectancy scales was low. For the self-efficacy scale given at the beginning of the course, 28.3% of participants selected “strongly agree,” 64.7% of participants selected “somewhat agree,” 6.6% of participants selected “somewhat disagree,” and 0.3% of participants selected “strongly disagree.” For the self-ef- ficacy scale given at the end of the course, 23.1% of par- ticipants selected “strongly agree,” 53.8% of participants selected “somewhat agree,” 19.6% of participants selected “somewhat disagree,” and 3.5% of participants “strongly disagree.” For the outcome expectancy scale given at the be- ginning of the course, 1% of participants selected “strong- ly agree,” 3.5% of participants selected “somewhat agree,” 33.9% of participants selected “somewhat disagree,” and 61.5% of participants “strongly disagree.” For the out- come expectancy scale given at the end of the course, 2.1% of participants selected “strongly agree,” 15.4% of partic- ipants selected “somewhat agree,” 35.7% of participants selected “somewhat disagree,” and 46.9% of participants “strongly disagree.” Due to this limited variability, respons- es for both scales were categorized into agree or disagree. Help-seeking behavior was assessed by self-report responses to the following two questions on the pre-and post-questionnaire: “If you did not understand something in class or got stuck when working on problems outside of class, how likely were you to: (1) Attend a peer tutor session; and (2) Go to the learning center.” Both state- ments utilized a 4-point Likert-type scale with response options of 1 = very likely, 2 = somewhat likely, 3 = some- what unlikely, and 4 = never would. Due to the limited variability, responses for those who responded positive- ly to either question (i.e., very likely or somewhat likely) were categorized as self-reported help-seekers while those who endorsed negative responses (i.e., somewhat unlike- ly or never would) were categorized as non-help-seek- ers. Kuder-Richardson 20 for the baseline questionnaire was 0.78 and for the follow-up questionnaire was 0.73. Academic procrastination was assessed by self-report responses to the following four questions on the pre-and post-questionnaire: With regard to academic tasks (e.g., reading for class, completing homework assignments, pre- paring for exams): (1) To what degree did you tend to de- lay or procrastinate?; (2) To what degree did you typically have to rush to complete a class-related task on time?; (3) How often did you begin assignments shortly after they are assigned? (reverse coded); and (4) To what degree was procrastination on academic tasks a problem for you? All statements utilized a 5-point Likert-type scale with re- sponse options of: Item 1: 1 = never procrastinate, 2 = almost never, 3 = sometimes, 4 = nearly always, and 5= always procrastinate; Item 2: 1 = never rush, 2 = almost never, 3 = sometimes, 4 = nearly always, 5 = always rush; Item 3: 1 = never begin shortly after they are assigned, 2 = almost never, 3 = sometimes, 4 = nearly always, 5 = always begin shortly after they are assigned; and Item 4: 1 = not at all a prob- lem, 2 = a small problem, 3 = a moderate problem, 4 = a large problem, and 5 = a very large problem. Scores on the four items were summed to create a total academic procrastination score, with higher scores indicating a greater tendency to procrastinate. Cronbach alpha for the baseline questionnaire was 0.79 and for the follow-up questionnaire was 0.82. Course Performance Measurements Students attended bi-weekly lectures taught by an instructor for 75 minutes per lecture and weekly laboratory sessions led by graduate student instruc- tors for 110 minutes per session. During lecture, the instructor taught students how to use and compute various types of analyses. Each of the three exams, administered during the lecture portion of the class, was semi-cumulative, covered approximately one- third of the course material, and was graded out of 100 possible points. The first examination consisted of multiple-choice questions that tested descriptive statistics, z-scores, correlation, regression, funda- mentals of the normal curve, and basic probability theory. The second examination consisted of mul- tiple-choice questions and one hypothesis testing procedure (i.e., complete a t-test by hand) on the principles and steps of hypothesis testing using sin- gle sample t-tests, dependent mean t-tests, z-tests, decision errors, effect size, power, and computation of confidence intervals. The third examination con- sisted of multiple-choice questions and one hypoth- esis testing procedure covering independent mean t-tests, analysis of variance, chi-square tests, rank-or- der tests, and specific advanced statistical procedures (e.g., hierarchical multiple regression, factor anal- ysis, structural equation modeling). Examinations were graded objectively by the lecturing professor and a graduate student instructor. Partial credit for hypothesis testing responses was possible and awarded according to an objective scoring rubric. In the laboratory sessions, graduate student instructors reviewed lecture material, demonstrat- 41 CHANGES IN SE AND OE IN A STATISTICS COURSE ed the use of statistical analysis using IBM SPSS Sta- tistics (IBM Corporation, 2021), and administered weekly quizzes. Additionally, two undergraduate peer tutors, who had previously taken the course and earned high marks, were available for voluntary tu- toring for approximately 4-6 hours per week to rein- force course concepts and prepare students for exams. The laboratory grade was computed based on average scores of weekly homework assignments (60%), mul- tiple-choice quizzes (15%), and attendance/participa- tion (25%). Homework assignments and laboratory quizzes were graded objectively by a graduate student instructor according to a detailed scoring rubric. Students’ final course grades were calculated based on scores on the three in-class exams (each worth 23%), overall lab grade (26%), attendance during lec- tures (2%), and a brief in-class presentation (3%). Statistical Analyses Descriptive statistics were calculated for the vari- ables with means and standard deviations for the con- tinuous variables and frequency and percentage for the categorical variables. The McNemar test was performed to compare the self-efficacy and outcome expectancy variables from baseline to follow-up on the third exam because it was the performance measure closest in time to follow-up. Multivariate logistic regression was per- formed to examine the factors that affected the self-effi- cacy and outcome expectancy variables. IBM SPSS Sta- tistics version 28 (IBM Corporation, 2021) was used for all the analyses. All p-values were two-tailed. Of the 286 participants, nine total students were not included in regression analyses. One student was not included because of missing laboratory quiz grades. Six other students were excluded because they did not respond to all the procrastination questions in the post-ques- tionnaire and two other students were excluded be- cause they did not indicate their current year in college. Results Table 1 shows sample descriptive statistics. More than three-quarters (76.6%) of participants identified as female, which could indicate overall increased neg- ative outcome expectancy for course outcomes (van Es & Weaver, 2018). College year was almost equally distributed with approximately one-third for each cat- egory, with third-year students representing the larg- est category. The fourth and other category included fourth-year students (n=64), students in a year greater than their fourth year (n=11), those obtaining their second bachelor’s degree/non-graduate post-baccalau- reate status (n=2), graduate students (n=2), non-de- gree-seeking students (n=3), and unknown (n=1). Toward the end of the semester (i.e., week 15), ap- proximately two-thirds of students reported that they were likely to engage in help-seeking behavior when they got stuck or did not understand course content. Mean exam scores were slightly above 80 for Exam 1, approximately 75 for Exam 2, and slightly above 70 for Exam 3. Mean lab quiz scores were slightly above 80. Mean procrastination scores were approximately 10.4 (lowest score was 4 and highest score was 20), which was slightly above the midpoint of possible scores, slightly tilting in the direction of higher pro- crastination. Self-efficacy lowered from slightly above 90% of participants at baseline endorsing a statement of feeling capable of mastering course material to ap- proximately three-quarters feeling the same way to- wards the end of the course. Additionally, outcome expectancy shifted negatively with about 95% of par- ticipants expecting to do well at the beginning of the course to slightly above 80% by the end of the course. Table 2 shows McNemar test comparisons from baseline to follow-up for agreeing with the statistics self-efficacy item, “I am [was] quite capable of mas- tering the material in this class.” For the entire sam- ple, for those who scored below 70, and those who scored below 80 on Exam 3, there was a statistically significant percentage decrease in statistics self-effi- cacy from baseline to follow-up. There were no sta- tistically significant percentage decreases in report- ed statistics self-efficacy from baseline to follow-up for those who scored 80-89.9 or 90-100 on Exam 3. Table 3 shows McNemar test comparisons from baseline to follow-up for disagreeing with the out- come expectancy item “I will [would] never [be able to] do well in this course.” In both the entire sample and those who scored below 70 on Exam 3, there was a statistically significant percentage decrease in posi- tive outcome expectancy from baseline to follow-up. There were no statistically significant percentage decreases from baseline to follow-up for those who scored from 70-79.9, 80-89.9, or 90-100 on Exam 3. Table 4 shows multivariate logistic regression anal- yses for follow-up of agreeing with the self-efficacy item “I was quite capable of mastering the course material in this class.” Higher scores on the first two exams were each 42 ELBULOK-CHARCAPE ET AL. statistically significantly associated with increased odds for high self-efficacy. The other variables were not sta- tistically significantly associated with high self-efficacy. Table 5 shows multivariate logistic regres- sion analyses for follow-up of positive out- come expectancy. No variables were significant- ly associated with positive outcome expectancy. Discussion In the current study, both high self-efficacy and positive outcome expectancy significantly decreased from baseline to follow-up in univariate analyses for lower-performing students on the third (and final) exam of the semester. In our multivariate analyses, we found that test performance on Exams 1 and 2 was significantly associated with increased odds of high self-efficacy. Positive outcome expectancy, however, was not associated with any of the aforementioned variables; thus, as predicted, outcome expectancy was less susceptible to change throughout the course. As reported, higher scores on the first two exams were associated with high self-efficacy. This is consis- tent with previous findings of a moderate correlation between self-efficacy and course performance in sta- tistics (Hii et al., 2013; McGrath et al., 2015; Waples, 2016). Importantly, we found that as performance decreased so did self-efficacy for the course. This was expected, as student performance is positively con- nected to student confidence (Sucuoğlu, 2018). We did not find a significant association between Exam 3 performance and increased odds for high self-efficacy. We speculate that students adjusted their self-efficacy based on performance on the earlier exams and by the final exam (Exam 3), their self-efficacy was less suscep- tible to change. In light of this finding, instructors could track critical points where students’ self-efficacy decreases, discuss self-efficacy directly at the outset of the course, and provide explicit strategies to enhance self-efficacy in those who are ambivalent or struggling in an introductory statistics course. After all, stu- dents with higher self-efficacy are more likely to con- tinue engaging with statistics potentially leading to improved statistics performance (Gopal et al., 2018). In our univariate analyses, we found a similar pattern for high self-efficacy and positive outcome ex- pectancy. Students who obtained lower Exam 3 scores from 0-69.9 showed decreases in both high self-effica- cy and positive outcome expectancy from baseline to follow-up. Conversely, for those who obtained high- er scores of 80 or higher on Exam 3, no statistically significant changes were found for high self-efficacy and positive outcome expectancy from baseline to fol- low-up. Essentially, self-efficacy and outcome expec- tancy changed in the same direction as performance on the final exam. These directional changes are expected as students’ perceived competence is based on mastery experiences (Bandura, 1986; 1997), including past suc- cesses as well as failures (Fong & Krause, 2014). In our multivariate analysis, no variables were significantly as- sociated with positive outcome expectancy. This adds to the mixed findings of previous literature reporting achievement (exam grades) as associated with (Hood et al., 2012) and not associated with outcome expectancy (Budé et al., 2007). However, our finding is similar to the finding of effort (self-reported help-seeking) and outcome expectancy having no significant relationship (Budé et al., 2007; Esnard et al., 2021; Hood et al., 2012). Limitations This study has certain limitations. We combined first year with second year students and fourth year with more senior students because our samples of those students were much smaller compared to the number of third year students. Responses to the self-efficacy and outcome expectancy questions were also dichot- omized because of limited variability in the responses. Second, self-efficacy and outcome expectancy were each self-assessed by only one item to reduce partici- pant burden (as the current study was part of a larger study of performance in statistics courses). Lengthier and validated measures of statistics self-efficacy, such as Current Statistics Self-Efficacy (CSSE) or Self-Ef- ficacy to Learn Statistics (SELS; Finney & Schraw, 2003) should be used in future research to determine if there are similar patterns as in our study. Third, contextual factors such as sequence of courses, office hour meetings, and prerequisites were not included. Recommendations Considering our findings, we recommend that statistics instructors identify students who are at risk for decreasing self-efficacy. Furthermore, instructors should help foster high self-efficacy in students by di- rectly connecting course content to students’ everyday lives, incorporating formative low-stakes assessments that may help increase students’ mastery experiences (Zientek et al., 2019), training students to employ study strategies (Liao & Wang, 2018), and teaching about 43 CHANGES IN SE AND OE IN A STATISTICS COURSE growth mindset (Samuel & Warner, 2021). Given the anxiety felt by students enrolling in statistics courses, a brief training based on this targeted instruction could be offered to statistics instructors to help them compre- hend common pitfalls and psychological variables that could help students navigate such a challenging course. Researchers found an improvement in statistics performance that correlated with post self-effica- cy scores only for undergraduate students who first solved statistics problems in a group setting, in which each member provided explanations for their own answers, gave feedback on the answers and explana- tions of the other group members, and then solved the same problems again individually with no feedback, compared to students who only solved the problems once, individually, with no feedback (Hall & Vance, 2010). Additionally, they suggested students be pro- vided an explanation of the course material prior to measuring their statistics self-efficacy so that students can give an accurate assessment of their ability. Based on their findings, we recommend that students be given opportunities to compare their statistics prob- lem-solving and answer-choice rationale to their peers’ statistics problem-solving and answer-choice rationale in low-stakes assessments. It is also vital that these op- portunities be provided numerous times throughout a statistics course so that changes in their self-effica- cy and performance remain congruent and improve. Conclusion In conclusion, through our multivariate analy- ses, we found that greater performance on class exams early in the semester was associated with high self-ef- ficacy. Change in positive outcome expectancy was not associated with exam performance. However, lower-performing students demonstrated decreases in both higher self-efficacy and positive outcome ex- pectancy. Self-efficacy was more vulnerable to course performance and strategies to address this should be considered as self-efficacy is a predictor of academ- ic success and achievement in statistics (Johnson & Kuennen, 2006; Mihai-Bogdan et al., 2015; Olani et al., 2011; Ramirez et al., 2012; Walker & Brakke, 2017). References Acee, T. W., & Weinstein, C. E. (2010). Effects of a val- ue-reappraisal intervention on statistics students’ motivation and performance. The Journal of Ex- perimental Education, 78(4), 487-512. https:// doi.org/10.1080/00220970903352753 Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Re- view, 84, 191-215. http://doi.org/10.1037/0033- 295X.84.2.191 Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall. Bandura, A. (1997). Self-efficacy: The exercise of con- trol. NY: W.H. Freeman. Barber, P. (2002). Critical analysis of psychological research: Rationale and design for a proposed course for the undergraduate psychology curric- ulum. Psychology Learning & Teaching, 2(2), 95- 101. https://doi.org/10.2304/plat.2002.2.2.95 Bateiha, S., Marchionda, H., & Autin, M. (2020). Teaching style and attitudes: a comparison of two collegiate introductory statistics classes. Journal of Statistics Education, 28(2), 154-164. http://doi.or g/10.1080/10691898.2020.1765710 Blair, R. M., Kirkman, E. E., & Maxwell, J. W. (2018). Statistical abstract undergraduate programs in the mathematical sciences in the United States: 2018 CBMS Survey. American Mathematical Society. Brown, E. N., & Kass, R. E. (2009). What is statis- tics? The American Statistician, 63(2), 105-110. https://doi.org/10.1198/tast.2009.0019 Budé, L., Van De Wiel, M. W., Imbos, T., Candel, M. J. J. M., Broers, N. J., & Berger, M. P. (2007). Stu- dents’ achievements in a statistics course in rela- tion to motivational aspects and study behaviour. Statistics Education Research Journal, 6(1), 5-21. Bureau of Labor Statistics, U.S. Department of Labor, Occupational Outlook Handbook, Mathematicians and Statisticians, at https://www.bls.gov/ooh/ math/mathematicians-and-statisticians.htm (vis- ited October 25, 2020) Cendales, B., Vargas-Trujillo, E., & Barbosa, C. (2013). Psychological factors associated with academic performance in university statistics courses: Dif- ferences by sex and degree area. Advances in Latin American Psychology, 31(2), 363-375. Cheng, S., Ferris, M., & Perolio, J. (2018). An inno- vative classroom approach for developing critical thinkers in the introductory statistics course. The American Statistician, 72(4), 354-358. https:// doi.org/10.1080/00031305.2017.1305293 Chiesi, F., & Primi, C. (2009). Assessing statistics at- 44 ELBULOK-CHARCAPE ET AL. titudes among college students: Psychometric properties of the Italian version of the Survey of Attitudes toward Statistics (SATS). Learning and Individual Differences, 19(2), 309-313. http://dx. doi.org/10.1016/j.lindif.2008.10.008 Crocker, J., & Major, B. (1989). Social stigma and self-esteem: The self-protective properties of stig- ma. Psychological Review, 96(4), 608–630. https:// doi.org/10.1037/0033-295X.96.4.608 Cui, S., Zhang, J., Guan, D., Zhao, X., & Si, J. (2019). Antecedents of statistics anxiety: An integrated ac- count. Personality and Individual Differences, 144, 79-87. https://doi.org/10.1016/j.paid.2019.02.036 Davis, A., & Mirick, R. G. (2015). MSW students’ perceptions of relevance and application of sta- tistics: Implications for field education. Journal of Teaching in Social Work, 35(3), 317-336. https:// doi.org/10.1080/08841233.2015.1040531 DeFreitas, S. C. (2012). Differences between Afri- can American and European American first- year college students in the relationship between self-efficacy, outcome expectations, and academ- ic achievement. Social Psychology of Education, 15(1), 109-123. http://doi.org/10.1007/s11218- 011-9172-0 Dempster, M., & McCorry, N. K. (2009). The role of previous experience and attitudes toward statistics in statistics assessment outcomes among under- graduate psychology students. Journal of Statistics Education, 17(2). https://doi.org/10.1080/10691 898.2009.11889515 Dunn, K. (2014). Why wait? The influence of aca- demic self-regulation, intrinsic motivation, and statistics anxiety on procrastination in online sta- tistics. Innovative Higher Education, 39(1), 33-44. https://doi.org/10.1007/s10755-013-9256-1 Elborolosy, S. A. M., & Al Thenyan, A. A. A. (2020). Improving flunked English majors performance through enhancing student’s perception of self-ef- ficacy and outcome expectancy at Shaqra Univer- sity. English Teaching Language, 13(9), 83-93. https://doi.org/10.5539/elt.v13n9p83 Esnard, T. R., Alladin, F. M., & Samlal, K. C. (2021). Prior mathematics performance, statistics anxi- ety, self-efficacy and expectations for performance in statistics: A survey of social sciences students in a Caribbean institution of higher education. Statistics Education Research Journal, 20(1), 4-4. https://doi.org/10.52041/serj.v20i1.98 Finney, S. J., & Schraw, G. (2003). Self-efficacy be- liefs in college statistics courses. Contemporary Educational Psychology, 28, 161-186. http://doi. org/10.1016/S0361-476X(02)00015-2 Fong, C. J., & Krause, J. M. (2014). Lost confidence and potential: A mixed methods study of under- achieving college students’ sources of self-efficacy. Social Psychology of Education, 17(2), 249-268. https://doi.org/10.1007/s11218-013-9239-1 Franceschini, G., Galli, S., Chiesi, F., & Primi, C. (2014). Implicit gender–math stereotype and women's susceptibility to stereotype threat and stereotype lift. Learning and Individual Differ- ences, 32, 273-277. https://doi.org/10.1016/j.lin- dif.2014.03.020 Glock, S., Unz, D., & Kovacs, C. (2012). Beyond fear appeals: Contradicting positive smoking outcome expectancies to influence smokers' implicit atti- tudes, perception, and behavior. Addictive Behav- iors, 37(4), 548-551. https://doi.org/10.1016/j. addbeh.2011.11.032 Gopal, K., Salim, N. R., & Ayub, A. F. M. (2018). In- fluence on self-efficacy and attitudes towards sta- tistics on undergraduates’ statistics engagement in a Malaysian public university. Journal of Physics: Conference Series, 1132(1), 12-42. https://doi. org/10.1088/1742-6596/1132/1/012042 Goroshit, M. (2018). Academic procrastination and academic performance: An initial basis for inter- vention. Journal of Prevention & Intervention in the Community, 46, 131-142. https://doi.org/10. 1080/10852352.2016.1198157 Graham, S. (1994). Motivation in African Americans. Review of Educational Research, 64, 55–118. Hall, S., & Vance, E. A. (2010). Improving self-efficacy in statistics: Role of self-explanation & feedback. Journal of Statistics Education, 18(3). https://doi. org/10.1080/10691898.2010.11889583 Han, J., Kelley, T., & Knowles, J. G. (2021). Factors influencing student STEM learning: Self-efficacy and outcome expectancy, 21st century skills, and career awareness. Journal for STEM Education Research, 4(2), 117-137. https://doi.org/10.1007/ s41979-021-00053-3 Hii, L., Myint, K. T., & Chieng, F. (2013). Students' self-efficacy in statistics and academic achieve- ment. Second International Higher Education 45 CHANGES IN SE AND OE IN A STATISTICS COURSE Teaching and Learning Conference, 113-118. Sar- awak, Malaysia. Hood, M., Creed, P. A., & Neumann, D. L. (2012). Using the expectancy value model of motivation to understand the relationship between student attitudes and achievement in statistics. Statistics Education Research Journal, 11(2), 72-85. http:// hdl.handle.net/10072/49231 IBM Corporation. (2021). IBM SPSS Statistics for Windows, Version 28.0. Armonk, NY: IBM Cor- poration. Johnson, M., & Kuennen, E. (2006). Basic math skills and performance in an introductory statis- tics course. Journal of Statistics Education, 14(2), 1–17. https://doi.org/10.1080/10691898.2006.1 1910581 Lent, R. W., Brown, S. D., & Hackett, G. (1994). To- ward a unifying social cognitive theory of career and academic interest, choice, and performance. Journal of Vocational Behavior, 45, 79-122. https://doi.org/10.1006/jvbe.1994.1027 Liao, H. C., & Wang, Y. H. (2018). Using comprehen- sion strategies for students' self-efficacy, anxiety, and proficiency in reading English as a foreign language. Social Behavior and Personality: An In- ternational Journal, 46(3), 447-458. https://doi. org/10.2224/sbp.6648 Lopez-Vergara, H. I., Colder, C. R., Hawk Jr, L. W., Wieczorek, W. F., Eiden, R. D., Lengua, L. J., & Read, J. P. (2012). Reinforcement sensitivity the- ory and alcohol outcome expectancies in early adolescence. The American Journal of Drug and Alcohol Abuse, 38(2), 130–134. https://doi.org/1 0.3109/00952990.2011.643973 Luo, T., So, W. W. M., Wan, Z. H., & Li, W. C. (2021). STEM stereotypes predict students’ STEM career interest via self-efficacy and outcome expectations. International Journal of STEM Education, 8(36). https://doi.org/10.1186/s40594-021-00295-y Mantooth, R., Usher, E. L., & Love, A. M. A. (2020). Changing classrooms bring new questions: Envi- ronmental influences, self-efficacy, and academic achievement. Learning Environments Research. https://doi.org/10.1007/s10984-020-09341-y McGrath, A. L., Ferns, A., Greiner, L., Wanamaker, K., & Brown, S. (2015). Reducing anxiety and in- creasing self-efficacy within an advanced graduate psychology statistics course. Canadian Journal for the Scholarship of Teaching and Learning, 6, 1–19. https://doi.org/10.5206/cjsotl-rcacea.2015.1.5 Mihai-Bogdan, I., Runcan, P., & Runcan, R. (2015). A study of the attitudes, self-efficacy, effort and academic achievement of social work students towards research methods and statistics: A dia- chronic approach. Revista de Asistenta Sociala, 1, 103–114. Miller, E. M. (2019). Promoting student success in statistics courses by tapping diverse cognitive abilities. Teaching of Psychology, 46(2), 140-145. https://doi.org/10.1177/0098628319834198 Newby, K. V., Brown, K. E., French, D. P., & Wallace, L. M. (2013). Which outcome expectancies are im- portant in determining young adults’ intentions to use condoms with casual sexual partners?: A cross-sectional study. BMC Public Health, 13(1), 133. https://doi.org/10.1186/1471-2458-13-133 Newman, R. S. (2002). How self-regulated learners cope with academic difficulty: The role of adaptive help seeking. Theory Into Practice, 41(2), 132-138. https://doi.org/10.1207/s15430421tip4102_10 Norcross, J. C., Hailstorks, R., Aiken, L. S., Pfund, R. A., Stamm, K. E., & Christidis, P. (2016). Under- graduate study in psychology: Curriculum and assessment. American Psychologist, 71(2), 89-101. https://doi.org/10.1037/a0040095 Olani, A., Hoekstra, R., Harskamp, E., & Van Der Werf, G. (2011). Statistical reasoning ability, self-efficacy, and value beliefs in a university sta- tistics course. Electronic Journal of Research in Educational Psychology, 9(1), 49-72. https://doi. org/10.25115/ejrep.v9i23.1427 Onwuegbuzie, A. J. (2004). Academic procrastination and statistics anxiety. Assessment & Evaluation in Higher Education, 29(1), 3-19. https://doi. org/10.1080/0260293042000160384 Paechter, M., Macher, D., Martskvishvili, K., Wim- mer, S., & Papousek, I. (2017). Mathematics anxi- ety and statistics anxiety. Shared but also unshared components and antagonistic contributions to performance in statistics. Frontiers in Psychology, 8, 1196. https://doi.org/10.3389/fpsyg.2017.01196 Pajares, F. (2009). Toward a positive psychology of aca- demic motivation: The role of self-efficacy beliefs. In R. Gilman, E. S. Huebner, & M. J. Furlong (Eds.), Handbook of positive psychology in schools (p. 149–160). Routledge/Taylor & Francis Group. 46 ELBULOK-CHARCAPE ET AL. Pokhrel, P., Little, M. A., Fagan, P., Muranaka, N., & Herzog, T. A. (2014). Electronic cigarette use out- come expectancies among college students. Ad- dictive Behaviors, 39(6), 1062-1065. https://doi. org/10.1016/j.addbeh.2014.02.014 Ramirez, C., Schau, C., & Emmīoglu, E. (2012). The importance of attitudes in statistics education. Statistics Education Research Journal, 11(2), 57- 71. https://doi.org/10.52041/serj.v11i2.329 Read, J. P. (2012). Reinforcement sensitivity theory and alcohol outcome expectancies in early adoles- cence. The American Journal of Drug and Alcohol Abuse, 38(2), 130-134. https://doi.org/10.3109/0 0952990.2011.643973 Rodarte-Luna, B., & Sherry, A. (2008). Sex differenc- es in the relation between statistics anxiety and cognitive/learning strategies. Contemporary Ed- ucational Psychology, 33(2), 327-344. https://doi. org/10.1016/j.cedpsych.2007.03.002 Samuel, T. S., & Warner, J. (2021). “I can math!”: reduc- ing math anxiety and increasing math self-efficacy using a mindfulness and growth mindset-based intervention in first-year students. Community College Journal of Research and Practice, 45(3), 205- 222. https://doi.org/10.1080/10668926.2019.166 6063 Schraw G., Watkins, T., & Olafson, L. (2007). Doing the things we do: A grounded theory of academic procrastination. Journal of Educational Psycholo- gy, 99(1), 12-25. https://doi.org/10.1037/0022- 0663.99.1.12 Slootmaeckers, K., Kerremans, B., & Adriaensen, J. (2014). Too afraid to learn: Attitudes towards statistics as a barrier to learning statistics and to acquiring quantitative skills. Politics, 34(2), 191- 200. https://doi.org/10.1111/1467-9256.12042 St-Pierre, R. A., Temcheff, C. E., Gupta, R., Dereven- sky, J., & Paskus, T. S. (2013). Predicting gambling problems from gambling outcome expectancies in college student-athletes. Journal of Gambling Studies, 30(1), 47-60. https://doi.org/10.1007/ s10899-012-9355-4 Sucuoğlu, E. (2018). Economic status, self-efficacy and academic achievement: The case study of under- graduate students. Quality & Quantity, 52(1), 851- 861. https://doi.org/10.1007/s11135-018-0692-y Trassi, A. P., Leonard, S. J., Rodrigues, L. D., Rodas, J. A., & Santos, F. H. (2022). Mediating factors of statistics anxiety in university students: a system- atic review and meta‐analysis. Annals of the New York Academy of Sciences, 1512(1), 76-97. https:// doi.org/10.1111/nyas.14746 Trogden, B. G., & Royal, J. E. (2019). Using exam wrappers in chemistry, mathematics, and statistics courses to encourage student metacognition. Jour- nal on Excellence in College Teaching, 30(3), 71-96. Urbán, R. (2010). Smoking outcome expectancies me- diate the association between sensation seeking, peer smoking, and smoking among young adoles- cents. Nicotine & Tobacco Research, 12(1), 59-68. https://doi.org/10.1093/ntr/ntp174 van Es, C., & Weaver, M. M. (2018). Race, sex, and their influences on introductory statistics educa- tion. Journal of Statistics Education, 26(1), 48-54. https://doi.org/10.1080/10691898.2018.1434426 Vanhoof, S., Kuppens, S., Castro Sotos, A. E., Verschaf- fel, L., & Onghena, P. (2011). Measuring statistics attitudes: Structure of the survey of attitudes to- wards statistics. Statistics Education Research Jour- nal, 10(1), 35-51. https://doi.org/10.52041/serj. v10i1.354 Walker, E. R., & Brakke, K. E. (2017). Undergraduate psychology students’ efficacy and attitudes across introductory and advanced statistics courses. Scholarship of Teaching and Learning in Psychol- ogy, 3(2), 132-140. http://dx.doi.org/10.1037/ stl0000088 Wang, Z., & Englander, F. (2010). A cross-disciplinary perspective on explaining student performance in introductory statistics--what is the relative impact of procrastination? College Student Journal, 44(2), 458-472. Waples, J. A. (2016). Building emotional rapport with students in statistics courses. Scholarship of Teach- ing and Learning in Psychology, 2(4), 285-293. https://doi.org/10.1037/stl0000071 Williams, D. M. (2010). Outcome expectancy and self-efficacy: Theoretical implications of an un- resolved contradiction. Personality and Social Psychology Review, 14(4), 417-425. https://doi. org/10.1177/1088868310368802 Wilson, S. G. (2013). The flipped class: A method to address the challenges of an undergraduate statis- tics course. Teaching of Psychology, 40(3), 193-199. https://doi.org/10.1177/0098628313487461 Yeager, D. S., Purdie-Vaughns, V., Garcia, J., Apfel, 47 N., Brzustoski, P., Master, A., . . . & Cohen, G. L. (2014). Breaking the cycle of mistrust: Wise in- terventions to provide critical feedback across the racial divide. Journal of Experimental Psychology: General, 143, 804–824. https://doi.org/10.1037/ a0033906 Zientek, L. R., Fong, C. J., & Phelps, J. M. (2019). Sources of self-efficacy of community college students enrolled in developmental mathemat- ics. Journal of Further and Higher Education, 43(2), 183-200. http://dx.doi.org/10.1080/030 9877X.2017.1357071 Zimprich, D. (2012). Attitudes toward statistics among Swiss psychology students. Swiss Jour- nal of Psychology, 71(3), 149-155. https://doi. org/10.1024/1421-0185/a000082 48 CHANGES IN SE AND OE IN A STATISTICS COURSE Table 1 Descriptive Statistics of Sample 49 ELBULOK-CHARCAPE ET AL. Table 2 Comparisons of High Self-Efficacy from Baseline to Follow-Up Note. N/A= not applicable since baseline had 100% agreement. 50 CHANGES IN SE AND OE IN A STATISTICS COURSE Table 3 Comparisons of Positive Outcome Expectancy from Baseline to Follow-up Note. N/A= not applicable since baseline had 100% agreement. 51 ELBULOK-CHARCAPE ET AL. Table 4 Logistic Regression Analyses for Follow-up of High Self-Efficacy 52 CHANGES IN SE AND OE IN A STATISTICS COURSE Table 5 Logistic Regression Analyses for Follow-up of Positive Outcome Expectancy