Effects of Locus of Control,Academic Self-Efficacy,and Tutoring on Academic Performance

Abstract

This study investigated the connection between locus of control (LOC), academic self-efficacy (ASE), and academic performance, and whether these variables are affected by tutoring. Additional variables of interest, including gender, students’ Pell Grant status, ethnicity, and class size, were also considered for the research models. The population for this study consisted of students enrolled at a mid-sized public university in northeastern United States who were pre- and posttested as part of a causal-comparative, quasi-experimental research design. Results of this study showed that LOC, tutoring, gender, and an ASE measure identified as self-assurance had positive and significant effects on academic performance as measured by students’ total grade point averages. However, tutoring had no effect on LOC but had only a small moderating effect on one component of ASE. It was postulated that assessing incoming students on LOC and ASE measures could aid in identifying students with external LOC and low ASE for possible intervention.

Keywords

locus of control academic self-efficacy tutoring repeated-measures design

Regardless of university efforts to retain students, nearly half of all students are still failing to graduate from 4-year institutions (Dennis, 1998; Fiske, 2004; Lederman, 2009). Data show that the proportion of first-year students who returned to their colleges as sophomores in 2007 to 2008, 65.7%, dropped to the lowest level in 25 years (Lederman, 2009). The intractability of this low retention rate has spawned numerous research studies involving strategies for addressing this issue. However, retention and academic performance remain hot topics in higher education and research is still needed to identify approaches to improving these outcomes. An important strategy for improving retention is to assess incoming students on factors that predict or are associated with academic performance to determine the readiness of students for college. Currently, most incoming college freshmen are evaluated on cognitive factors associated with academic performance, such as scholastic assessment test (SAT) scores and high school grade point averages (GPAs). However, there undoubtedly exist noncognitive factors that are also associated with postsecondary academic performance and on which undergraduate freshmen could, and should, be assessed. Two such factors that merit investigation are locus of control (LOC) and academic self-efficacy (ASE).

Numerous studies have examined the possible links between LOC, ASE, and academic performance (Choi, 2005; Desle, 2011; Ferla, Valcke, & Cai, 2009; Gifford, Briceno-Perriott, & Mianzo, 2006; Gore, 2006; Landine & Stewart, 1998; Putwain, Sander, & Larkin, 2013; Uguak, Elias, Uli, & Suandi, 2007; Zajacova, Lynch, & Espenshade, 2005). LOC refers to a person’s beliefs about control over life’s events, in which a person with internal LOC orientation attributes personal success or failure to his or her own efforts and abilities. Alternatively, a person with external LOC believes that outcomes are related more to extenuating circumstances beyond personal control, such as luck, fate, or God. Internal LOC is more conducive to high achievement and independent functioning than external LOC (Gifford et al., 2006; Rotter, 1966; Wood, Saylor, & Cohen, 2009).

LOC and Academic Performance

Several studies have demonstrated a positive relationship between LOC and academic performance (Carden, Bryant, & Moss, 2004; Desle, 2011; Gifford et al., 2006; Landine & Stewart, 1998; Uguak et al., 2007). Gifford et al. (2006) investigated the relationship between LOC and GPA and found that LOC was a significant predictor for cumulative GPA in a regression analysis, and that freshmen with internal LOC scores demonstrated a significantly higher GPA than freshmen who obtained external LOC scores. Landine and Stewart (1998) examined the relationship between a number of variables, including LOC and academic average. They found a significant positive relationship between LOC and academic average and concluded that LOC was related to academic achievement. In a correlational analysis, Uguak et al. (2007) considered LOC and a variable called academic achievement satisfaction and found a significant positive association between the two variables. Students scoring highest with internal LOC showed superior performance in comparison to those with external LOC.

Carden et al. (2004) conducted a study in which they divided students into two groups, an internally LOC-oriented grouping and an externally LOC-oriented grouping. The internal LOC students showed significantly lower academic procrastination and higher academic achievement than the external group. Desle (2011) compared high- and low-achieving junior college students with their LOC scores and found that high achievers had more internal LOC than low achievers, and that low achievers had more external LOC. Their study revealed that high achievers were more intelligent, self-sufficient, and controlled and relaxed and had better social and emotional adjustment than low achievers. Low achievers were more affected by feelings, more obedient, and tenser than high achievers. Since high achievers were associated with internal LOC, the implication is that internal LOC students were likely to be more intelligent and self-sufficient and have better social and emotional adjustment than students with an external LOC.

ASE and Academic Performance

ASE is a way of thinking about the beliefs students have about their competencies and can be defined as a belief, or confidence, that one can achieve a specific academic goal or attain a particular outcome on a specific academic task (as cited in Putwain et al., 2013). Studies have typically conceptualized ASE as domain-specific competence beliefs in a particular academic subject, as well as perceived competence in context-specific study-related skills and behaviors, typically those thought to contribute to self-regulated learning (Putwain et al., 2013). ASE influences cognition and self-regulation through the use of metacognitive strategies and correlates with success on homework, exams, and quizzes (as cited in Rheinheimer & Francois, 2001).

The relationship between ASE and academic performance has been the subject of a number of studies (Choi, 2005; Ferla et al., 2009; Gore, 2006; Landine & Stewart, 1998; Putwain et al., 2013; Zajacova et al., 2005). Ferla et al. (2009) studied ASE and academic self-concept to investigate the nature of the relationship between these two self-constructs and their impact on academic achievement and other outcome variables. They found that academic self-concept is a better predictor for affective-motivational variables, while ASE is the better predictor for academic achievement. Putwain et al. (2013) examined the relationship between three dimensions, or scales, of ASE and first and second semester academic performance for a sample of undergraduate students. The components of ASE they modeled were one’s ability to (a) study independently, (b) discuss material with one’s tutors and peers, and (c) attain expected grades. Their research revealed that self-efficacy (SE) in study-related skills positively predicted Semester 1 performance but not Semester 2 performance, and neither of the other two SE scales positively predicted Semester 1 or Semester 2 academic performance. Choi (2005) studied self-efficacy and self-concept constructs and their relationships with course grades in a university setting. Choi considered three levels of specificity for self-efficacy—general self-efficacy, ASE, and specific self-efficacy—and two levels of self-concept—academic self-concept and specific self-concept—in his analyses. Choi’s results showed that both self-concept variables, along with specific self-efficacy, were significant predictors for course grades. Although ASE was significantly correlated with course grades, it was not a significant predictor in the linear regression analysis.

Gore (2006) conducted incremental regression studies over three semesters to determine the extent to which ASE beliefs could account for variability in academic performance beyond that accounted for by standardized test scores. Gore found that ASE predicted college GPA, but the relationship depended on when the efficacy beliefs were measured and was much stronger when ASE was measured at the end, rather than the beginning, of the semester. Zajacova et al. (2005) investigated the joint effects of ASE and stress on the academic performance of college freshmen at a diverse urban commuter university. They utilized structural equation models to assess the relative importance of stress and ASE to predict first-year GPA, the number of credits earned, and retention. Their results suggested that ASE was more important than perceived stress in predicting accumulated credits and GPA, but stress was slightly more important in predicting retention. In their study about LOC and academic achievement, Landine and Stewart (1998) included self-efficacy in their variables of interest. Their results indicated that self-efficacy was also positively related to academic average.

Tutoring and Academic Performance

The impact of tutoring on academic performance in higher education has been less extensively studied than LOC and ASE, but a number of recent studies have explored this relationship (Arco-Tirado, Fernandez-Martin, & Fernandez-Balboa, 2011; Higgins, 2004; Rheinheimer, Grace-Odeleye, Francois, & Kusorgbor, 2010; Rheinheimer & McKenzie, 2011; Rheinheimer & Mann, 2000). In their research on peer tutoring, Arco-Tirado et al. (2011) had as one of their objectives determining the impact of a peer-tutoring program on preventing academic failure and dropouts among first-year students at a university in Spain. Their research model consisted of an experimental group in which students were tutored and a control group in which students were not. The results showed differences in favor of the tutored group GPA, performance rate, success rate, and learning strategies. Higgins (2004) investigated the use of a peer-tutoring program to increase retention of students at risk of failing a medical-surgical nursing class. Students considered as at risk were divided into two groups for analysis: one group of students received tutoring while the other group did not. The results from this study showed that peer tutoring had a significant effect on the academic performance and retention of at-risk students in a medical-surgical nursing class.

Rheinheimer et al. (2010) conducted a longitudinal research study on a sample of at-risk students in a state-funded equal opportunity program at a public university. They utilized a nonexperimental, ex post facto methodology to assess the impact of tutoring on academic performance variables over a 3-year period. Their results indicated that tutoring had significant positive relationships with retention, GPA, and total credits toward graduation. Rheinheimer and McKenzie (2011) tracked a cohort of undeclared freshmen over 4 years and used logistic regression and survival analysis to examine the relationship between tutoring and students’ retention rates and decision paths. Findings from their nonexperimental, causal-comparative study showed that tutoring had a significant impact on retention but not on GPA or time to selection of a major. Rheinheimer and Mann (2000) explored, as part of their study, the relationship between tutoring variables and the grades students received in courses in which they were tutored. Their study found a strong relationship between grades in courses tutored and hours of tutoring, and that the impact of tutoring was affected by the amount, or level, of tutoring received by students.

LOC, ASE, and Tutoring

Little research was found directly linking LOC, ASE, and tutoring. Several studies have been conducted on LOC and tutoring (Lazerson, Foster, Brown, & Hummel, 1988; Miller & Connolly, 2013; Yasutake, Bryan, & Dohrn, 1996), ASE and tutoring (Griffin & Griffin, 1997), and LOC and ASE (Cassidy, 2007). Lazerson et al. (1988) used truant and tardy junior high school students with learning disabilities as tutors for younger learning-disabled students to study the effect of tutoring on tutors’ LOC and other affective behaviors. The authors found that after 6 weeks of tutoring, the tutors demonstrated significant gains in LOC and showed a decrease in truant and tardy behaviors. In a model somewhat similar to that of Lazerson et al. and Yasutake et al. (1996), students with learning disabilities and students at risk for referral for special education assessment served as tutors for younger students. The elementary age students and tutors were divided into groups in which tutors were trained to use either strategy feedback or attribution plus strategy feedback. The results of this study indicated that peer tutoring with attribution feedback can positively influence third through eighth graders’ self-perceptions of confidence, which is one aspect of LOC. Miller and Connolly (2013) conducted a large-scale randomized controlled trial evaluation on the effect of a volunteer tutoring program for young children on a number of variables, including LOC and self-esteem. The authors found no effect for the tutoring program on LOC or self-esteem; however, misspecification and low dosage of the tutoring program may have contributed to the lack of evidence for significance.

Griffin and Griffin (1997) investigated the effects of reciprocal peer tutoring on academic achievement, ASE, and test anxiety. Reciprocal peer tutoring (RPT) is a procedure that allows both members of a tutoring pair to participate in the tutor role, so that students function reciprocally as both tutor and tutee. The authors conducted two experiments with a nonequivalent control group design in which one group of graduate students received RPT while the other group did not. The results of their experiments did not reveal any group differences on the achievement measures, ASE, or test anxiety, although students’ perceptions about RPT were uniformly positive. In a study to establish the level of students’ self-assessment skills, Cassidy (2007) compared students’ estimated marks of their work with tutors’ actual marks. Cassidy also examined the relationship between self-assessment skill, learning style, LOC, and ASE for the sample of comparatively inexperienced undergraduate students. While the findings indicated a good level of self-assessment skill for a majority of the students, no association was found between self-assessment skill and LOC or ASE.

Although relationships between LOC, ASE, tutoring, and achievement outcomes, such as academic performance, have been separately studied, the interplay between all four of these factors simultaneously has not. What is needed is research in which all four of these variables are studied together to determine how LOC, ASE, and tutoring interact with academic success. The purpose of this article is to discuss the results of a study conducted to investigate the connection between LOC, ASE, and academic performance, and whether these variables are affected by tutoring. Additional variables of interest, including gender, students’ Pell Grant status, and class size, were also considered for the research models.

Methodology

The research surveyed earlier establishes a need for a study to more deeply examine the effect that tutoring has on LOC and ASE, and the effects that all three factors have on academic performance. This study examined data collected on two sets of students, those who received tutoring and those who did not. The primary research questions of interest for this study were as follows:
RQ1: What is the impact of locus of control on student academic performance controlled for gender, class size, and Pell Grant status?

RQ2: What is the impact of academic self-efficacy on student academic performance controlled for gender, class size, and Pell Grant status?

RQ3: What is the impact of tutoring on locus of control controlled for gender, class size, and Pell Grant status?

RQ4: What is the impact of tutoring on academic self-efficacy controlled for gender, class size, and Pell Grant status?

RQ5: What is the impact of tutoring on academic performance controlled for gender, class size, and Pell Grant status?

Sample

The population for this study consisted of students enrolled at a mid-sized public university in northeastern United States in the Fall Semester of 2012. The sample consisted of students from eight psychology sections and students who were tutored in the university’s Learning Center. Some of the students who were tutored in the Learning Center were also in the psychology classes, and these students were considered as part of the treatment, or tutoring, group. The rest of the students from the psychology classes comprised the control group. All students in the control and treatment groups who agreed to participate in the study were administered a survey instrument near the beginning of the Fall 2012 Semester. Near the end of the Fall 2012 Semester, the survey instrument was again administered to the students in both the treatment and control groups. A total of 499 students completed the surveys in the initial (pre) administration of the instrument, 387 students participated in the second (post) administration of the survey, and 349 students completed both administrations of the survey. Information was requested and retrieved from the academic computing office for student academic information and to identify students who received Pell Grants. Tutoring databases were examined to determine which students had received tutoring.

Students considered for this study who were tutored received individual or small group (maximum four in a group) tutoring that was delivered by student (peer) or professional tutors. Student tutors were undergraduate students who were at least sophomores, juniors, or seniors, while professional tutors were individuals with at least a bachelor’s degree. All tutors attended 10 hours of tutor training soon after being hired. Students were allowed a maximum of 2 hours of tutoring a week which was conducted in a Learning Center central to the campus. All tutoring was coordinated through the University-Wide Tutoring Program which typically received between 1,500 and 2,000 requests for tutoring in the Fall Semester. While tutoring was primarily offered for 100/200 (freshman and sophomore) level classes, any undergraduate student could apply for and receive tutoring in one or more subjects at the time of this study. The only limitation to receiving tutoring was the availability of tutoring in the subject area in which the tutoring was requested. For this study, students who were assigned a tutor were considered to have been tutored and were part of a tutored group. Students in the control group were not assigned a tutor and received no tutoring.

Instrument

A survey instrument was developed from Rotter’s 29-item LOC instrument (Rotter, 1966) and a locally developed instrument to measure ASE (Rheinheimer & Francois, 2001). Items comprising two factors from the three-factor self-efficacy instrument were combined with the 29-item Rotter instrument, along with two demographic items, to create the 48-item survey instrument used for this study. The two self-efficacy factors in the survey were identified as academic maturity (AM) and self-assurance (SA), and the two demographic items requested students’ gender and parents’ college graduation status.

The Rotter component of the instrument comprised items with two selections and were scored on a 0, 1 scale. Only 23 of the Rotter items were included for the LOC scoring, resulting in a maximum possible individual LOC score of 23. The Rotter items were formatted so that a high score indicated an internal LOC and a low score was aligned with external LOC, which is a reversal of the original Rotter scale.

The two ASE factors were scored on a 5-point Likert scale which ranged from strongly disagree to strongly agree. These options were converted to a 1 to 5 numerical scoring rating for data analysis. Scores were summed over the 10-item AM and 7-item SA scales to create sum and mean scores for each of these factors for each student. The mean scores for each student ranged from 1 to 5 and these means were then subsequently used in the data analyses. The Cronbach’s alpha reliability for the ASE measures for this sample was .94.

Data Collection

The 48-item survey instrument was administered near the beginning and at the end of the Fall 2012 Semester to students in eight sections of first-year psychology classes and a sample of students who were tutored in the Learning Center at the university. Four of the psychology classes were large, or mega, classes, with enrollments of 100 or more students. These classes were chosen for this study because together they comprised a large sample of mostly first-year students. A sample of tutors from the Learning Center were asked to administer the survey instrument to their tutees, the students they were tutoring, at approximately the same time as the surveys were administered in the psychology classes. The tutors chosen to administer the surveys represented a cross section of disciplines offered at the university, and most of the students surveyed were taking first-year classes. All students were given informed consent forms to sign if they were willing to participate, and they were told that their participation was entirely voluntary and that their responses would be kept confidential.

Procedure

The methodological design of this study was causal-comparative, quasi-experimental research, with both descriptive and inferential procedures used to analyze the data. Quasi-experimental designs are appropriate for research studies when the data cannot be completely randomized (Polit & Hungler, 1997). The primary independent variable in this study was a grouping variable involving students who were and were not tutored. Since it is not ethical to randomly assign students to such a group, a true experimental study was not possible for this research.

For the descriptive analyses of this study, frequency counts, means, standard deviations, and correlations were calculated, while t tests, chi-square procedures, multiple regression, and repeated measures analyses were utilized to conduct the inferential analyses. The level of significance, α, for all statistical tests was set at .05, and all statistical analyses were conducted with the SAS statistical package.

Repeated measures analysis was used to examine the effect of the treatment variables on LOC and ASE over time. This within-subject procedure determines the effect of independent and treatment variables when the subjects are subjected to a pre- or posttreatment or treatments.

The dependent variables for this study included LOC, an ASE measure called AM, an ASE measure identified as SA, and students’ GPAs. The primary independent variable was a grouping variable (Group) which consisted of the control and treatment groups. The control group consisted of students who were not tutored, and the treatment group consisted of students who were tutored. Additional independent variables for this study included gender, Pell Grant status, ethnicity, a variable that represented whether or not a student was in a large, or mega, class (Megacourse), and students’ total earned credits toward graduation (TCTG). LOC, AM, and SA also served as independent variables for the multiple regression analysis. For the repeated measures and regression analyses, group, gender, Pell Grant status, ethnicity, and megacourse were coded as 0/1 dummy variables.

Results

The breakdown of the students by gender, treatment group, Pell Grant status, ethnicity, and whether or not they were in a megacourse (Megacourse) is given in Table 1 for the pre- and postadministrations of the survey and for the 349 students who completed both administrations of the survey. Of the 349 students who completed both of the survey administrations, 226 (64%) were females and 127 (36%) were males. The breakdown for the treatment group was 226 (64%) for the control group and 127 (36%) for the tutoring group. Also, there were 107 (30.3%) students who received Pell Grants and 246 (69.7%) who did not, 232 (65.7%) Caucasian and 121 (34.3%) other ethnicities, and 179 (50.7%) students in a megacourse and 174 (49.3%) who were not.
Table 1.
Frequency Counts and Percentages for Selected Study Variables.

Variable Pre (n = 499) Post (n = 387) Both (n = 353)

Gender

Male 210 (42.1) 145 (37.5) 127 (36.0)

Female 289 (57.9) 289 (57.9) 226 (64.0)

Treatment group

Control 330 (66.0) 254 (65.6) 226 (64.0)

Treatment 170 (34.0) 133 (34.4) 127 (36.0)

Pell Grant status

Yes na 117 (30.2) 107 (30.3)

No na 270 (69.8) 246 (69.7)

Ethnicity

Caucasian na 253 (65.4) 232 (65.7)

Other na 134 (34.6) 121 (34.3)

Megacourse

Yes na 121 (34.3) 179 (50.7)

No na 184 (47.6) 174 (49.3)

Note. Pre = students who took the first administration of survey; Post = students who took the second administration of survey; Both = students who took both administrations of the survey; Numbers in parentheses in table = relative frequencies; na = no data available for that variable.

The results of selected data analyses are given in Tables 2 through 7. Significant results were found for correlations between LOC and GPA, LOC and AM, LOC and SA, and between AM and SA (Table 2).
Table 2.
Intercorrelations Between Selected Study Variables.

LOC GPA AM SA

LOC – .12* (353) .18 (385) .15* (385)

GPA – −.04 (385) .06 (385)

AM – .77** (385)

SA –

Note. LOC = locus of control posttest score; GPA = student’s cumulative grade point average; AM = Academic Maturity posttest score; SA = Self-Assurance posttest score. Numbers in parentheses represent correlation sample sizes.

p < .05. p < .001. p < .01.

From the t-test results in Table 3, it can be seen that students who were tutored had significantly higher GPAs, more credits toward graduation (TCTG), and higher AM self-efficacy scores than students who were not tutored. The students who were tutored also had significantly lower verbal SAT scores than students who were not tutored. No significant differences were found between students who were and were not tutored for LOC scores and SA self-efficacy scores. Effect sizes for significant differences ranged from 0.25 to 0.39, which Cohen suggested would be between a small (0.2) and medium (0.5) effect (as cited in Field, 2009).
Table 3.
Descriptive Statistics and t-Test Comparisons for Selected Study Variables.

Variable N M SD t Score F ratio Cohen’s d^a

GPA 255 133 2.66 2.94 0.88 0.64 3.60* 1.89** 0.35

TCTG 254 133 25.54 34.58 22.04 31.85 2.93* 2.09 0.35

Locus of control 254 133 11.61 11.97 3.26 3.57 0.99 1.20 0.11

Academic maturity 254 131 3.86 4.16 0.78 0.73 3.64* 1.15 0.39

Self-assurance 254 131 4.12 4.25 0.89 0.76 1.50 1.39 0.15

Verbal SAT 237 131 474.64 459.67 61.85 56.91 2.22 1.18 −0.25

Note. GPA = student’s cumulative grade point average; TCTG = total credits toward graduation; SAT = scholastic assessment test. The first column for N, M and SD for each variable represents the statistics for students who did not receive tutoring (control group), and the second column is for students who did receive tutoring. The F ratios are values for the tests for variance homogeneity between the group levels (control or tutoring).

^aEffect sizes calculated with pooled standard deviations.

p < .01. p < .0001. p < .001. ****p < .05.

The results of the repeated measures analyses (Table 4 –6) showed significant effects for time by group and time by gender by Pell Grant status for AM (Table 5), and for time by SA (Table 6). There were no significant effects for LOC (Table 4).
Table 4.
Repeated Measures Analysis of Variance for LOC.

Effect Numerator df Denominator df λ Value F ratio

Time 1 345 1.00 0.17

Time × Group 1 345 1.00 0.98

Time × Gender 1 345 1.00 0.30

Time × Group × Gender 1 345 1.00 1.68

Time × Pell Grant 1 345 1.00 1.16

Time × Group × Pell Grant 1 345 1.00 1.01

Time × Gender × Pell Grant 1 345 1.00 0.01

Note. Wilks’ Lambda (λ) was the multivariate test statistic value reported for each effect.

Table 5.
Repeated Measures Analysis of Variance for Academic Maturity.

Effect Numerator df Denominator df λ Value F ratio

Time 1 342 1.00 1.54

Time × Group 1 342 0.99 5.07*

Time × Gender 1 342 1.00 1.69

Time × Group × Gender 1 342 1.00 0.02

Time × Pell Grant 1 342 1.00 0.00

Time × Group × Pell Grant 1 342 0.99 2.66

Time × Gender× Pell Grant 1 342 0.99 4.15*

Category Least Square Means for Time by Group

Level of group
N
Pretest
Posttest

Not tutored 225 4.03 3.83

Tutored 125 4.02 4.08

Category Least Square Means for Time by Gender by Pell Grant

Level of gender
Level of Pell Grant
N
Pretest
Posttest

Male No 90 3.99 3.87

Male Yes 36 3.74 3.86

Female No 155 4.14 4.11

Female Yes 69 4.23 3.97

Note. Wilks’ Lambda (λ) was the multivariate test statistic value reported for each effect.

p < .01.

Table 6.
Repeated Measures Analysis of Variance for SA.

Effect Numerator df Denominator df λ Value F ratio

Time 1 342 0.99 4.03

Time × Group 1 342 1.00 1.02

Time × Gender 1 342 1.00 0.13

Time × Group × Gender 1 342 1.00 0.81

Time × Pell Grant 1 342 1.00 0.43

Time × Group × Pell Grant 1 342 1.00 0.11

Time × Gender × Pell Grant 1 342 1.00 0.36

Category Least Square Means for Time

Variable
N
Pretest
Posttest

SA 350 4.29 4.18

Note. SA = self-assurance. Wilks’ Lambda (λ) was the multivariate test statistic value reported for each effect.

p < .01.

The multiple regression analysis (Table 7) found LOC, AM, SA, gender (Female), and whether or not students were tutored (Tutoring) to be significant predictors for GPA. Nonsignificant predictors for GPA included Pell Grant status (Pellgrant), megacourse status (Megacourse), and ethnicity (White).
Table 7.
Multiple Regression Model With GPA as Dependent Variable.

Predictor B SE (B)* β

Intercept 2.39* 0.26 0

Locus of control 0.03 0.01 .13

Academic maturity −0.30* 0.08 −.29

Self-assurance 0.21**** 0.07 .22

Female 0.42* 0.09 .25

Tutoring 0.22** 0.09 .13

Pell Grant −0.01 0.09 −.01

Megacourse −0.01 0.08 −.01

White −0.02 0.09 −.01

Note. n = 385. GPA = grade point averages; B = unstandardized beta coefficients; SE = standard error of beta; β = standardized beta coefficients. R²= .13.

p < .0001. p < .05. p < .001. ****p < .01.

Discussion

Research Question 1: What is the impact of LOC on student academic performance controlled for gender, class size, and Pell Grant status?

The results of this study demonstrate that LOC had a significant and positive effect on academic performance as measured by GPA, when controlling for gender, class size, Pell grant status, and ethnicity. From Table 2, we see the significant positive correlation between LOC and GPA, and the results of the multiple regression analysis regression model in Table 7 show the positive impact of LOC on GPA.

The value of a predictor for a multiple regression model provides insight into the impact of this parameter on the dependent measure. In Table 7, the values of the predictor variables are indicated by the column headed by B, and these values represent the coefficients of the variables in the regression model. Each significant coefficient represents the individual contribution of that variable in the model, indicating the degree to which each predictor affects the outcome if the effects of all other predictors are held constant (Field, 2009). The value of a significant coefficient represents, therefore, the change in the dependent variable caused by a unit change in the predictor when the other predictors are held constant. For the model in Table 7, an increase of one point in a student’s LOC score translates into an increase of 0.03 in the student’s GPA, independent of the other predictors in the model.

Research Question 2: What is the impact of ASE on student academic performance controlled for gender, class size, and Pell Grant status?

From Table 7, we see that both measures of ASE, AM and SA, are significant predictors of GPA. The coefficient for SA indicates that a unit change in SA results in an increase of 0.21 in GPA, when all other predictors are held constant. However, the coefficient for AM is negative, indicating that GPA decreases as AM increases. Since AM and SA were highly correlated, it is possible that AM was suppressed in the regression model.

Research Question 3: What is the impact of tutoring on LOC controlled for gender, class size, and Pell Grant status?

The repeated measures analysis in Table 4 revealed no significant effect for group on LOC. The variable group represented the grouping variable for the tutored and nontutored groups, that is, those students who were and were not tutored. The repeated measures analysis investigated the impact of the set of variables on changing students’ LOC scores over the course of the treatment period, a little more than half of the semester. An absence of significance implies no change in the LOC scores, indicating that tutoring had no impact on LOC in this study.

Research Question 4: What is the impact of tutoring on ASE controlled for gender, class size, and Pell Grant status?

The repeated measures analysis results in Table 5 show a significant effect for time by group. A significant time by group effect indicates that the variable group significantly changed the dependent variable, in this case AM. To interpret this effect, it is necessary to examine the category least square means for time by group in Table 5. These means reveal that the AM scores for the students who were not tutored decreased from the pre- to the posttest, while the AM scores for the students who were tutored remained relatively stable. Therefore, it would appear that tutoring had a neutral effect on AM but not being tutored had a negative impact on this component of ASE.

Alternatively, the repeated measures results in Table 6 show no significant effect for time by group for SA, indicating that the group variable showed no change in SA scores over the course of the study. This result indicates that tutoring had no impact on the ASE component SA.

Research Question 5: What is the impact of tutoring on academic performance controlled for gender, class size, and Pell Grant status?

Table 3 shows that there was a significant difference in the final GPAs between students who were and were not tutored. An examination of Table 3 also shows that students who were tutored had significantly lower verbal SAT scores than students who were not tutored. This difference in SAT scores indicates that the students who were tutored had a precollege profile more closely aligned with that of at-risk students, yet when they were tutored they earned GPAs that were nearly 0.30 higher than students who were not tutored.

This finding is further strengthened by the results from Table 7, which show that tutoring was a significant predictor for GPA. The coefficient for tutoring in the regression model from Table 7 suggests that being tutored corresponds to an increase of 0.22 in a student’s GPA, when the effects of the other predictors are held constant.

Limitations and Recommendations

One limitation to the study is that the results of this research are restricted in application to the population of undergraduate students at the university at which this study was conducted. These findings are certainly very useful and most important for this institution, and since many colleges and universities have similar populations, the outcomes from this research would likely apply to numerous university settings. However, to make generalizations to a broader population, this study needs to be replicated with student samples from a variety of colleges and universities with different demographics.

A second limitation to this research is that the study was restricted to just a few variables. As evidenced from the R² value in the multiple regression model, there are more variables that need to be considered for studies such as this one. Variables such as major field of study, self-regulation, and emotions (i.e., boredom, anxiety, and enjoyment) may also account for variability in academic performance and impact LOC. Ruthig et al. (2008), for example, found that emotions moderated the effects of LOC on academic achievement. Future studies should investigate the effects of these and other variables, along with the problem of self-selection bias (Gattis, 2002), to more accurately evaluate the effect of LOC, tutoring, and ASE on academic performance.

A third limitation is that there may not have been enough time between the pre- and postadministrations of the survey instrument to accurately measure changes in LOC and ASE. Because of circumstances beyond the control of the researchers, the preadministration of the survey instrument was delayed. This delay reduced the time available for the full effects of teaching and tutoring to be realized. Realistically, it might take two or more semesters to detect any significant changes in LOC, AM, and SA.

Another possible limitation is that tutoring was measured as a dichotomous variable. A tutoring variable that was measured as interval or ratio might have enhanced the interpretation of the influence of tutoring in this model, allowing the strength of tutoring to be manifested on a gradient. It should be noted, however, that the dichotomous structure of the variable Tutoring can also signify the importance of tutoring in the regression model. As a significant variable in a yes or no format, Tutoring is indicating that the mere presence of tutoring is positively associated with the outcome variable, which can be viewed as a compelling endorsement of the value of tutoring.

Conclusions

LOC and an ASE measure identified as SA had positive and significant effects on academic performance as measured by students’ total GPAs. The impact these variables had on GPA was demonstrated by a regression model in which both were significant predictors. An implication of this finding is that if incoming students could be assessed on LOC and self-efficacy measures, students with high external (low internal) LOC and low SA could be targeted for interventions. These interventions would be designed to alter students’ LOC and SA, consequently increasing their chances of academic success. While this outcome is applicable for students in the university in which this study was undertaken, it most probably is relevant to many other universities as well.

The within-subjects research design failed, however, to produce any significant pre or post difference in students’ LOC scores; therefore, tutoring was not a factor in enhancing LOC. Although it was initially hypothesized that tutoring might have a positive impact on LOC, the delay in completing the first administration of the survey instrument may have contributed to the lack of significance in the repeated measures findings. It is also very likely that any measurable impact on LOC by tutoring would need more time to be detected, underlining the need for a follow-up study.

Tutoring had a significant positive effect on GPA as evidenced by the significantly higher GPA for students who were tutored over those who were not and by the significant entry of tutoring into the multiple regression model for GPA. The results of these analyses demonstrate that, for a local population of students, tutoring is an effective strategy for improving academic performance and can level the playing field even for students who exhibit characteristics of less academically prepared learners.

Variable	Pre (n = 499)	Post (n = 387)	Both (n = 353)
Gender
Male	210 (42.1)	145 (37.5)	127 (36.0)
Female	289 (57.9)	289 (57.9)	226 (64.0)
Treatment group
Control	330 (66.0)	254 (65.6)	226 (64.0)
Treatment	170 (34.0)	133 (34.4)	127 (36.0)
Pell Grant status
Yes	na	117 (30.2)	107 (30.3)
No	na	270 (69.8)	246 (69.7)
Ethnicity
Caucasian	na	253 (65.4)	232 (65.7)
Other	na	134 (34.6)	121 (34.3)
Megacourse
Yes	na	121 (34.3)	179 (50.7)
No	na	184 (47.6)	174 (49.3)

	LOC	GPA	AM	SA
LOC	–	.12* (353)	.18** (385)	.15*** (385)
GPA		–	−.04 (385)	.06 (385)
AM			–	.77** (385)
SA				–

Variable	N	M	SD	t Score	F ratio	Cohen’s d^a
GPA	255 133	2.66 2.94	0.88 0.64	3.60*	1.89**	0.35
TCTG	254 133	25.54 34.58	22.04 31.85	2.93*	2.09**	0.35
Locus of control	254 133	11.61 11.97	3.26 3.57	0.99	1.20	0.11
Academic maturity	254 131	3.86 4.16	0.78 0.73	3.64***	1.15	0.39
Self-assurance	254 131	4.12 4.25	0.89 0.76	1.50	1.39****	0.15
Verbal SAT	237 131	474.64 459.67	61.85 56.91	2.22****	1.18	−0.25

Effect	Numerator df	Denominator df	λ Value	F ratio
Time	1	345	1.00	0.17
Time × Group	1	345	1.00	0.98
Time × Gender	1	345	1.00	0.30
Time × Group × Gender	1	345	1.00	1.68
Time × Pell Grant	1	345	1.00	1.16
Time × Group × Pell Grant	1	345	1.00	1.01
Time × Gender × Pell Grant	1	345	1.00	0.01

Effect	Numerator df	Denominator df	λ Value	F ratio
Time	1	342	1.00	1.54
Time × Group	1	342	0.99	5.07*
Time × Gender	1	342	1.00	1.69
Time × Group × Gender	1	342	1.00	0.02
Time × Pell Grant	1	342	1.00	0.00
Time × Group × Pell Grant	1	342	0.99	2.66
Time × Gender× Pell Grant	1	342	0.99	4.15*
Category Least Square Means for Time by Group
Level of group	N	Pretest	Posttest
Not tutored	225	4.03	3.83
Tutored	125	4.02	4.08
Category Least Square Means for Time by Gender by Pell Grant
Level of gender	Level of Pell Grant	N	Pretest	Posttest
Male	No	90	3.99	3.87
Male	Yes	36	3.74	3.86
Female	No	155	4.14	4.11
Female	Yes	69	4.23	3.97

Effect	Numerator df	Denominator df	λ Value	F ratio
Time	1	342	0.99	4.03*
Time × Group	1	342	1.00	1.02
Time × Gender	1	342	1.00	0.13
Time × Group × Gender	1	342	1.00	0.81
Time × Pell Grant	1	342	1.00	0.43
Time × Group × Pell Grant	1	342	1.00	0.11
Time × Gender × Pell Grant	1	342	1.00	0.36
Category Least Square Means for Time
Variable	N	Pretest	Posttest
SA	350	4.29	4.18

Predictor	B	SE (B)	β
Intercept	2.39*	0.26	0
Locus of control	0.03**	0.01	.13
Academic maturity	−0.30***	0.08	−.29
Self-assurance	0.21****	0.07	.22
Female	0.42*	0.09	.25
Tutoring	0.22**	0.09	.13
Pell Grant	−0.01	0.09	−.01
Megacourse	−0.01	0.08	−.01
White	−0.02	0.09	−.01

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Anthony Drago is chair and professor of Psychology at East Stroudsburg University. Dr. Drago is also a psychologist in private practice and consultant to numerous Behavioral Health agencies in the Pocono Mountain region of Pennsylvania. Dr. Drago's areas of expertise include DSM-5, suicide, and trauma. He has lectured on numerous occasions in the US, China, and Costa Rica, and his current research interests include self-efficacy, locus of control, suicide subtypes, and caregiver response to autism.

David C. Rheinheimer is distinguished professor emeritus from East Stroudsburg University, and currently serves as a statistical and research methods consultant and advisor. Dr. Rheinheimer is conversant with a wide range of statistical methods and applications, and has published articles on research conducted in tutoring and various statistical applications in education. His current research interests include student self-efficacy, locus of control, and other issues involving student success and retention.

Thomas N. Detweiler served as a professional tutor for 20+ years in the University-Wide Tutoring Program at East Stroudsburg University, focusing on the Biological Sciences. Mr. Detweiler also served as district supervisor for the PA Department of Environmental Protection. In his tutoring practice, Mr. Detweiler worked with students in developing an internal locus of control to improve self-efficacy in addition to assisting them in understanding course materials.

References

Arco-Tirado

J. L.

Fernandez-Martin

F. D.

Fernandez-Balboa

(2011) The impact of a peer-tutoring program on quality standards in higher education. Higher Education 62: 773–778.

Carden

Bryant

Moss

(2004) Locus of control, test anxiety, academic procrastination, and achievement among college students. Psychological Reports 95(2): 581–582.

Cassidy

(2007) Assessing ‘inexperienced’ students’ ability to self-assess: Exploring links with learning style and academic personal control. Assessment & Evaluation in Higher Education 32(3): 313–330.

Choi

(2005) Self-efficacy and self-concept as predictors of college students’ academic performance. Psychology in the Schools 42(2): 197–205.

Dennis

M. J.

(1998) A practical guide to enrollment management. In: Hossler

Bean

J. P.

(eds) The strategies management of college enrollments, San Francisco, CA: Jossey-Bass, pp. 170–185.

Desle

V. F.

(2011) Personality, adjustment factors and locus of control of high and low achievers. Indian Journal of Community Psychology 7(1): 139–146.

Ferla

Valcke

Cai

(2009) Academic self-efficacy and academic self-concept: Reconsidering structural relationships. Learning and Individual Differences 19(4): 499–505.

Field

(2009) Discovering statistics using SPSS, 3rd ed. London, England: Sage Publications.

Fiske, E. (2004). Refuse to lose: Today’s colleges and universities must work to foster student success. A report of the Lumina Foundation, Focus [PowerPoint slides]. Retrieved from http://www.luminafoundation.org/publications/focus.pdf.

10.

Gattis

K. W.

(2002) Responding to self-selection bias in assessments of academic support programs: A motivational control study of supplemental instruction. The Learning Assistance Review 7(2): 26–36.

11.

Gifford

D. G.

Briceno-Perriott

Mianzo

(2006) Locus of control: Academic achievement and retention in a sample of university first-year students. Journal of College Admission 191: 18–25.

12.

Gore

P. A.

Jr. (2006) Academic self-efficacy as a predictor of college outcomes: Two incremental validity studies. Journal of Career Assessment 14(1): 92–115.

13.

Griffin

B. W.

Griffin

M. M.

(1997) The effects of reciprocal tutoring on graduate students’ achievement, test anxiety, and academic self-efficacy. Journal of Experimental Education 65(3): 197–210.

14.

Higgins

(2004) Relationship between retention and peer tutoring for at-risk students. Journal of Nursing Education 43(7): 319–321.

15.

Landine

Stewart

(1998) Relationship between metacognition, motivation, locus of control, self-efficacy, and academic achievement. Canadian Journal of Counseling 32(3): 200–212.

16.

Lazerson

D. B.

Foster

H. L.

Brown

S. I.

Hummel

J. W.

(1988) The effectiveness of cross-age tutoring with truant, junior high school students with learning disabilities. Journal of Learning Disabilities 21(4): 253–255.

17.

Lederman, D. (2009, January 23). As talk about retention rises, rates drop [electronic version]. Inside Higher Ed. Retrieved from http://www.insidehighered.com/news/2009/01/23/retain.

18.

Miller

Connolly

(2013) A randomized controlled trial evaluation of Time to Read, a volunteer tutoring program for 8- to 9-year-olds. Educational Policy and Analysis 35(1): 23–37.

19.

Polit

D. F.

Hungler

B. P.

(1997) Essentials of nursing research: Methods, appraisals, and utilization, Philadelphia, PA: Lippincott.

20.

Putwain

Sander

Larkin

(2013) Academic self-efficacy in study-related skills and behaviours: Relations with learning-related emotions and academic success. British Journal of Educational Psychology 83: 633–650.

21.

Rheinheimer

McKenzie

(2011) The impact of tutoring on the academic success of undeclared students. Journal of College Reading & Learning 41(2): 22–36.

22.

Rheinheimer

Mann

(2000) Gender matching, floor effects, and other tutoring outcomes. Journal of Developmental Education 24(2): 10–18.

23.

Rheinheimer, D. C., & Francois, G. (2001). Assessing student self-efficacy and its effect on academic performance for traditional and at-risk students. Third National Conference on Research in Developmental Education, Charlotte, NC.

24.

Rheinheimer

D. C.

Grace-Odeleye

Francois

G. E.

Kusorgbor

(2010) Tutoring: A support strategy for at-risk students. The Learning Assistance Review 15(1): 23–33.

25.

Rotter

(1966) Generalized expectancies for internal versus external control of reinforcements. Psychological Monographs 80(1): 1–28.

26.

Ruthig

J. C.

Perry

R. P.

Hladkyj

Hall

N. C.

Pekrun

Chipperfield

J. G.

(2008) Perceived control and emotions: Interactive effects on performance in achievement settings. Social Psychology of Education 11(2): 161–180.

27.

Uguak

U. A.

Elias

Uli

Suandi

(2007) The influence of causal elements of locus of control on academic achievement satisfaction. Journal of Instructional Psychology 34(2): 120–128.

28.

Wood

A. M.

Saylor

Cohen

(2009) Locus of control and academic success among ethnically diverse baccalaureate nursing students. Nursing Education Perspectives 30(5): 290–294.

29.

Yasutake

Bryan

Dohrn

(1996) The effects of combining peer tutoring and attribution training on students’ perceived self-competence. Remedial and Special Education 17(2): 83–91.

30.

Zajacova

Lynch

S. M.

Espenshade

T. J.

(2005) Self-efficacy, stress, and academic success in college. Research in Higher Education 46(6): 677–706.