academic performance of college students

ABSTRACT. Today’s college students

are less prepared for college-level work

than their predecessors. Once they get to

college, they tend to spend fewer hours

studying while spending more hours work-

ing, some even full time (D. T. Smart, C. A.

Kelley, & J. S. Conant, 1999). In this study,

the authors examined the effect of both

time spent studying and time spent working

on academic performance. The authors fur-

ther evaluated the interaction of motivation

and ability with study time and its effect on

academic performance. The results suggest-

ed that nonability variables like motivation

and study time significantly interact with

ability to influence academic performance.

Contrary to popular belief, the amount of

time spent studying or at work had no

direct influence on academic performance.

The authors also addressed implications

and direction for future research.

Copyright © 2006 Heldref Publications

oday’s college students are spend-ing less time studying. The fall

2003 survey conducted by the HigherEducation Research Institute atUCLA’s Graduate School of Educationand Information Studies found thatonly 34% of today’s entering freshmenhave spent six or more hours per weekoutside of class on academic-relatedwork (e.g., doing homework, studying)during their senior year in high school.The sample consisted of 276,449 stu-dents at 413 of the nation’s 4-year col-leges and universities (over one fourthof entering freshmen in the UnitedStates), and the data were statisticallyadjusted to reflect responses of all first-time, full-time students entering allfour-year colleges and universities asfreshmen in 2003. In fact, in 1987 whenthis question was asked of enteringfreshmen, 47.0% claimed they spent 6or more hours per week studying out-side of class. Since then, the time spentstudying outside of class has declinedsteadily each year (Higher EducationResearch Institute, 2003).

Another trend that is emerging is theincrease in the number of college stu-dents who are employed either part timeor full time. According to Gose (1998),39% of college freshmen work 16 ormore hours per week, an increase of 4%since 1993. Among all business majors,marketing students typically work evenmore hours per week than do other stu-

dents (Smart, Tomkovick, Jones, &Menon, 1999). The 2002 survey con-ducted by the Higher EducationResearch Institute also found that65.3% of entering freshmen have either“some concern” or “major concerns”about not having enough money to com-plete their college degrees (Higher Edu-cation Research Institute, 2002). Thiswas an increase of almost 1% from2001 and is likely to increase in theyears ahead because of reduced fundingfor higher education by state legisla-tures. Although more women (70.9%)were concerned about whether theywould have enough funds to completecollege than were men (58.3%), all stu-dents seemed to be working out of theneed to make up for rising tuition andfewer available grants. In summary, theproportion of college students who areemployed either part or full time is like-ly to increase in the years to come, leav-ing greater numbers of students withless time for academic work.

Students spending less time studyingand more time working are two trendsthat all colleges and universities will haveto confront. Lowering academic stan-dards by rewarding minimum effort andachievement (expecting less) is certainlya short-term strategy, but one that willhave negative long-term consequences. Amore productive way to handle these con-cerns is to conduct empirical research todetermine to what extent these trends will

Academic Performance of CollegeStudents: Influence of Time SpentStudying and WorkingSARATH A. NONISGAIL I. HUDSONARKANSAS STATE UNIVERSITYJONESBORO, ARKANSAS

T

January/February 2006 151

negatively impact the academic perfor-mance of college students and use thefindings from these studies to improveour academic programs.

The influence that personal variables,such as motivation and ability, have onacademic success is well documented,but there is a paucity of research inves-tigating the influence that time collegestudents spend on various activitiessuch as studying outside of class andworking has on their academic success.One reason for a lack of research in thisarea may be the common belief amongmost students and academicians thatmore time spent studying outside ofclass positively influences academicperformance and that more time spentworking negatively influences academicperformance. Another, more plausiblereason for this lack of research may bethe complex nature of these relation-ships when evaluated in the presence ofother variables, such as student abilityand motivation. For example, it is likelythat time spent studying outside of classwill have a differential impact on theacademic performance of college stu-dents who vary in ability. That is, therelationship that ability has with studentperformance will be stronger for thosestudents who spend more time outsideof class studying than for students whospend less time studying.

With this study, we attempted to fillthis void in the literature. First, weattempted to determine the direct rela-tionship that time spent on academicsoutside of class and working had on aca-demic performance among business stu-dents. Second, we attempted to deter-mine whether the time spent onacademics outside of class interacts withvariables, such as student ability andmotivation, in influencing the academicperformance of business students.

Hypotheses Tested

It is commonly believed that studentswho spend more time on academic-related activities outside of class (e.g.,reading the text, completing assign-ments, studying, and preparing reports)are better performers than students whospend less time on these activities. Thereis some empirical support for this belief.For example, Pascarella and Terenzini

(1991) found that the study habits offreshmen relate significantly to their firstyear cumulative grade point average(GPA). In their investigation of 143 col-lege students, McFadden and Dart(1992) reported that total study timeinfluenced expected course grades. Incontrast, Mouw and Khanna (1993) didnot find study habits to significantlyimprove the explanatory power of thefirst year cumulative GPA of college stu-dents. Ackerman and Gross (2003) havefound more recently that students withless free time have a significantly higherGPA than those with more free time.Because of this conflicting evidence,there is a need to reinvestigate this rela-tionship. Thus, our first hypothesis was

H1: There is a relationship between timespent studying outside of class and aca-demic performance.

Along with the present trend of stu-dents spending less time on academic-related activities, a growing number ofcollege and university administrators areconcerned that today’s postsecondarystudents are working more hours thantheir counterparts were years ago (Gose,1998). It can be reasonably assumed thatworking more hours per week will leavestudents less time for studying outside ofclass and that this will negatively influ-ence their academic performance.Although working more hours per weekcan be one key reason for a student to bein academic trouble, available researchdoes not seem to support this hypothesis.Strauss and Volkwein (2002) reportedthat working more hours per week posi-tively related to a student’s GPA. Light(2001), who interviewed undergraduatestudents of all majors, found no signifi-cant relationship between paid work andgrades. According to Light, “studentswho work a lot, a little, or not at all sharea similar pattern of grades” (p. 29).Because empirical evidence to date hasbeen counterintuitive, testing thishypothesis using different samples anddifferent methodologies is importantbefore generalizations can be made. Thisled to our next hypothesis that

H2: There is a relationship between timespent working and academic performance.

According to Pinder (1984) and oth-ers (Chan, Schmitt, Sacco, & DeShon,1998; Chatman, 1989; Dreher & Bretz,

1991; Nonis & Wright, 2003; Wright &Mischel, 1987), performance is a multi-plicative function of both ability andmotivation.

Performance = Ability × Motivation

For example, a student with very highability but low motivation is unlikely toperform well, whereas a student withlow ability but high motivation is likelyto perform well. That is, the variabilityin motivation across students maydampen associations between abilityand performance.

In the same vein, one can argue thatit is simply the study behavior that ulti-mately brings about the desired perfor-mance and not students’ inner desires ormotivations. This is supported by thewidely held belief that it is hard work(i.e., time spent on academic activitiesoutside of class by a student) thatresults in academic success and thatlaziness and procrastination ultimatelyresult in academic failure (Paden &Stell, 1997). Therefore, similar to howmotivation interacts with ability toinfluence academic performance, onecan infer that behavior such as hardwork interacts with ability to influenceperformance among college students.This led us to our third hypothesis to betested in this study.

H3: Behavior (time spent studying out-side of class) will significantly interactwith ability in that the influence that abil-ity has on academic performance will behigher for students who spend more timestudying outside of class than for studentswho spend less time studying.

All indications are that today’s collegefreshmen are less prepared for collegethan their predecessors. American Col-lege Testing (ACT) Assessment reportsthat fewer than half of the students whotake the ACT are prepared for college.According to the Legislative Analyst’sOffice (2001), almost half of those stu-dents regularly admitted to the CaliforniaState University system arrive unpre-pared in reading, writing, and mathemat-ics. Although these statistics are commonat most colleges and universities in thenation, how institutions handle theseconcerns varies. Strategies includeattempting to develop methods to diag-nose readiness for college-level workwhile students are still in high school or

152 Journal of Education for Business

requiring remedial courses of enteringfreshmen, thereby lowering academicstandards.

There are others who believe that it isnot only reading, writing, and mathemat-ics abilities that influence academic per-formance, but also nonability variables,such as motivation (Barling & Charbon-neau, 1992; Spence, Helmreich, & Pred,1987), self-efficacy (Bandura & Schunk,1981; Multon, Brown, & Lent, 1991;Zimmerman, 1989), and optimism(Nonis & Wright, 2003). Although aminimum level of ability is required, it isplausible that nonability variables willcompensate for ability inadequacies tobring about the required level of perfor-mance. One question that interests allparties is whether hard work (i.e., moretime spent studying) will influence therelationship between motivation and per-formance. That is, will the relationshipbetween motivation and academic per-formance be stronger if a student putsmore effort or time into studying outsideof class compared with those who put inless time? This led to our final hypothe-sis, one that was speculative in nature,but nevertheless has implications forboth students and academicians.

H4: Behavior (more time spent studyingoutside of class) will significantly interactwith motivation in that the influence thatmotivation has on academic performancewill be higher for students who spendmore time studying outside of class com-pared with students who spend less timestudying outside of class.

METHOD

Sample

We secured the data for this studyfrom a sample of undergraduate stu-dents attending a medium-sized(10,000+), Association to AdvanceCollegiate Schools of Business(AACSB)-accredited, public universityin the mid-south United States. Toobtain a representative sample of allstudents, we selected classes from avariety of business courses (e.g., man-agement, accounting, MIS, finance),offered at various levels (freshmen,sophomore, junior, and senior) and atdifferent times (day or night), for thestudy. Data collection occurred duringthe 9th week of a 15-week semester.

This timing was deliberate becausedata were being collected for the moti-vation variable, one that is likely tochange among students during theearly and late parts of a semester. Inaddition, information on such variablesas time spent on academics or work-related activities is also likely to varyduring the beginning, middle, and endof a semester.

We distributed surveys and explainedthem to those students who participatedin the study. The survey consisted oftwo parts. The first part required stu-dents to maintain a journal during a 1-week period, documenting how muchtime they spent on various activitieseach day of the week (there were over25 activities listed under three broadcategories: academics, personal, andwork related). For accuracy purposes,we asked students to complete theirjournal each morning, recording theprevious day’s activities. The secondpart of the survey contained demo-graphic information, such as gender,age, and race, as well as measures ofseveral other constructs including moti-vation (only motivation was used in thisstudy). Participants had to provide theirsocial security numbers for documenta-tion purposes. We assured them thattheir responses would be pooled withothers and no effort would be made toevaluate how any one individual mayhave responded to the survey. We urgedstudents to take the task seriously andto be accurate in their responses to eachquestion. A cover letter signed by thedean of the college of business wasincluded in each student’s journal. Weadministered 440 surveys, and 288were returned. Two hundred and sixty-four of the returned surveys wereusable, yielding an effective responserate of 60.0%.

Measures

We used the social security numbersprovided by the respondents to collectuniversity data for the variable gradepoint average for the semester (SGPA),semester courseload, number of hourscompleted to date, and ACT compositescore. As such, these variables were notself-reported and should provide morevalidity to the study’s findings.

Achievement Striving

We used six items from a Spence etal. (1987) Likert-type 1–5-point scale,to measure students’ achievement striv-ing, which we used as a surrogate formotivation. In several prior studies,researchers have used this variable as ameasure of motivation (Barling & Char-bonneau, 1992; Barling, Kelloway, &Cheung, 1996). The reported coefficientalpha for this scale is high (0.87), andthis scale has been used in several othersimilar studies (Carlson, Bozeman,Kacmar, Wright, & McMahan, 2000;Nonis & Wright, 2003).

Demographic Variables

Students reported demographic infor-mation, such as gender, age, and racialor ethnic group membership, in theirjournals.

Behavior Variables

We also used student journal data todetermine the time spent outside ofclass on academic activities like readingthe text and lecture notes for classpreparation, going over the text and lec-ture notes to prepare for exams, andcompleting assignments and homework.The researcher added these items for theweek to derive the total amount of timestudents spent outside of class on aca-demic activities during the week (TSA).

Students also reported the time theyspent working, as well as the time ittook for them to travel to and from workeach day, during the given week. Thesetwo items were also added to derive thetotal amount of time students spentworking during a given week (TSW).

Analysis

As Table 1 shows, sample character-istics were comparable to availabledemographic characteristics of collegestudents in the United States (StatisticalAbstract of the United States, 2002).Other pertinent demographic character-istics for the sample were as follows:average age = 23.8 years; majors = 16%accounting, 13.1% business administra-tion, 12.3% finance, 13.5% manage-ment, 14.5% marketing, 14.5% MIS,and the remainder “other” businessmajors.


We coded gender and racial or ethnicgroup membership and used them asdummy variables consisting of two cat-egories (coded 0 or 1), such as male orfemale and African American or other,because 97.5% of the sample was eitherCaucasian or African American. Todetermine the bivariate relationshipsthat the plausible predictor (indepen-dent) variables had with the academicsuccess (dependent) variables, we cal-culated Pearson’s product moment cor-relation coefficients. Table 2 shows boththe descriptive statistics and the Pear-son’s correlation coefficients. Theachievement-striving measure demon-

strated an acceptable reliability coeffi-cient as per Nunnally (1978).

Prior to testing the hypotheses, it wasimportant for us to control for variablesthat were likely to have an impact onacademic performance other than thevariables that we were testing. Studieshave found that demographic variables,such as gender, age, and race (Cubeta,Travers, & Sheckley, 2001; Strauss &Volkwein, 2002), influence the academ-ic performance of college students.Therefore, we tested H1 and H2 usingpartial correlation coefficients, control-ling for the extraneous variables gender,age, and racial or ethnic group. Aca-

demic load was also included as a con-trol variable because students who takemore courses are likely to spend moretime studying outside of class comparedwith students who take fewer courses.In addition, we treated TSA and TSW asindependent variables, and we usedSGPA as the dependent variable.

We tested moderator relationshipsproposed in H3 and H4 through moder-ated multiple regression analysis(Cohen & Cohen, 1983; Wise, Peters, &O’Conner, 1984). We performed threeregressions: (a) We regressed the depen-dent variable (SGPA) on the controlvariables (gender, age, racial or ethnicgroup membership, and academic load);(b) we regressed the dependent variableon the control variables, plus the inde-pendent variable (i.e., ACT compositescore as a surrogate for ability), plus themoderator variable (i.e., TSA as a surro-gate for hard work or behavior); and (c)we regressed the dependent variable onthe control variables, plus the indepen-dent variable, plus the moderator vari-able, plus the interaction (i.e., ACTcomposite score and TSA).

The process involved conductingthree regression models for each mod-erator hypothesis. This process facili-tated the investigation of a potentialdirect influence of the moderator vari-ables (when they serve as predictors)and the extent to which the positedmoderator influence actually exists.When both the independent and themoderator variable are continuous


TABLE 1. Demographic Characteristics of the Sample Compared With thePopulation, in Percentages

Demographic characteristic Populationa Sample

GenderMale 43.6 44.2Female 56.3 55.8

Racial/Ethnic Group White 77 85African American 12.1 12Other 11 2.5

Employment StatusDo not work 35.6 34Work part time 30.3 28Work full time 34.1 37

Note. The sample consisted of undergraduate students enrolled in business courses at a medium-sized, Association to Advance Collegiate Schools of Business-accredited public university in themid-south.aBased on Statistical Abstract of the United States (2002).

TABLE 2. Descriptive Statistics and Pearson Product–Moment Correlations for Study Variables

Variable M SD 1 2 3 4 5 6 7 8

1. Gender — — —2. Age 23.76 6.29 –0.04 —3. Race — — 0.04 –0.07 —4. ACT composite (ACT) 22.00 3.91 –0.01 –0.24* 0.31* —5. Achievement striving (AST)a 3.53 0.71 –0.17* 0.21* –0.09 –0.02 —6. Time spent outside of class on

academic activities (TSA) vs. academic load 12.94 8.57 –0.07 0.34* –0.16 –0.18* 0.29* —

7. Time spent working (TSW) vs. academic load 16.84 14.55 0.03 –0.03 0.06 –0.03 –0.16* –0.06 —

8. Semester grade point average (SGPA) 2.97 0.76 –0.11 –0.09 0.27* 0.45* 0.35* 0.05 –0.10 —

areliability coefficient = 0.77.*p < .05 (one-tailed).

(ACT composite and TSA), as in thisstudy, the appropriate statistical proce-dure to detect interaction is the moder-ated multiple regression analysis (Bar-ron & Kenny, 1986).

Because the measurement units asso-ciated with the various scales used inthis study were different, we standard-ized variables investigated in the analy-

ses and used z scores when testinghypotheses.

RESULTS

The partial correlation coefficientbetween TSA and SGPA, controlling forthe variables gender, age, race, and aca-demic load (r = .10, p = .19), was in the

expected direction, but not significant.Therefore, H1 was not supported. Thepartial correlation coefficient betweenTSW and SGPA (r = −.08, p = .28) wasalso statistically insignificant, failing tosupport H2.

Moderated Multiple Regression(MMR), controlling for gender, age,race, and academic load, provided thestatistics required to test the remainingtwo hypotheses. For H3, the R2 for thecontrol variables was statistically signif-icant (R2 = .06, p < .05). In the secondstep, the increment to R2 was statistical-ly significant for the addition of themain effects of ACT composite andTSA (∆R2 = .19, p < .05). In fact, themain effects of both ACT composite andTSA were also significant (p < .05).From Step 2 to Step 3, the increment ofR2 was also significant for the additionof the interaction term (∆R2 = .03, p <.05); this supported H3, which statedthat TSA would interact with ability(see Table 3). Predicted values generat-ed from the regression equation thatwere one standard deviation above andbelow the mean for ACT compositescore and TSA indicated that studentswho were high in ACT composite andTSA most likely had a very high semes-ter GPA (y-hat or predicted value =3.95), relative to students high in ACTcomposite score with low TSA (y-hat =3.1) and relative to students low in ACTcomposite with either high (y-hat = 2.5)or low (y-hat = 2.7) TSA. This is theappropriate technique to interpret inter-action terms when moderated multipleregression is implemented (Cleary &Kessler, 1982; Cohen & Cohen, 1983).Results are shown in Figure 1.

For H4, the R2 for the control variableswas once again statistically significant(R2 = .10, p < .05). In the second step, theincrement to R2 was statistically signifi-cant (∆R2 = .14, p < .05) for the additionof the main effects of achievement striv-ing and TSA. However, the main effectof TSA was not statistically significant.From Step 2 to Step 3, the increment ofR2 was also not statistically significant(∆R2 = .01, p > .05) for the addition ofthe interaction term. These results did notprovide support for H4, which stated thattime spent studying outside of classwould interact with motivation (see Table4). Therefore, H4 was not supported.


TABLE 3. Results of Moderated Multiple Regression Analysis of TimeSpent Outside of Class on Academic Activities (TSA), ACT CompositeScore (ACT), and Semester Grade Point Average (SGPA)

Independent variable Slope SE t p R 2

Control (Step 1) .06*Gender –0.12 0.13 –0.90 .36Age 0.03 0.10 0.26 .78Race 0.70 0.22 3.22 .00*Academic load 0.05 0.07 0.74 .46

Predictor (Step 2) .25*ACT composite (ACT) 0.43 0.07 6.60 .00*Time studying (TSA) 0.17 0.07 2.54 .01*

Moderator (Interaction) (Step 3) .28*ACT × TSA 0.18 0.07 2.69 .01*

*p < .05.

FIGURE 1. Time spent studying (TSA) and ACT composite score (ACT)interaction on semester grade point average (SGPA). Graph is based onpredicted values (y-hat) generated from the regression equation for indi-viduals 1 standard deviation above and below the mean for TSA and ACT.

SGPA TSA (Low)

TSA (High)

LOW HIGH

ACT

DISCUSSION

We drew the following conclusionsfrom the analyses.

1. Contrary to popular belief, thefindings suggest that TSW has no directinfluence on SGPA.

2. Based on the partial correlation,findings suggest that TSA has no directinfluence on academic performance(measured as SGPA).

3. The main effects of both ACT com-posite score and achievement strivingare statistically significant.

4. In the presence of ACT compositescore, the main effect of TSA also has astatistically significant relationship withSGPA. However, in the presence ofachievement striving, the main effect ofTSA does not have a significant interac-tion with SGPA.

5. The interaction between ACT com-posite score and TSA significantlyinfluences SGPA.

6. The interaction between TSA andachievement striving did not significant-ly influence SGPA.

Based on partial correlation coeffi-cients, neither of the hypotheses thattested direct relationships (H1 and H2)was supported. However, one of thehypotheses that investigated the mod-erator relationship was supported(H3). These results indicate that therelationships that college students’abilities (ACT composite score), moti-vation (achievement striving), andbehavior (TSA and TSW) have with

academic performance are more com-plex than what individuals believethem to be.

One important finding of this studyis the lack of evidence for a direct rela-tionship between TSW and academicperformance (H2). TSW did notdirectly affect academic performance.At a time when the percentage of col-lege students who work is at an all-time high and administrators are con-cerned about its influence on academicperformance, these results are encour-aging. Although more empirical evi-dence may be required prior to makingany definitive conclusions, theseresults did not contradict the findingsof Strauss and Volkwein (2002) orLight (2001). Contrary to popularbelief, both Strauss and Volkwein andLight found that working more hourswas positively related to GPA and sug-gested that students apply the samework ethic to both their academic andpaid work (i.e., those who earn highergrades are students who are moremotivated, and work harder and longerthan others). Perhaps academicallystrong students are better at balancingacademic and job-related work, there-by reducing the negative effects thatTSW may have on academic perfor-mance.

Based on the partial correlation (r =.10, p > .05), the expected influencethat TSA has on academic performance(H1) was not supported. When we test-ed H4, the insignificant main effectbetween time spent outside of class on

academic activities (TSA) and academ-ic performance (see Table 4) also sup-ports the above conclusion. However,when we tested H3, the significantmain effect between TSA and academ-ic performance (Table 3) was not con-sistent with the previous findings in H1and H4. That is, when ACT compositescore was used as a predictor (in theabsence of achievement striving), TSAhad an impact on academic perfor-mance (see Table 3). Also, whenachievement striving was used as a pre-dictor (in the absence of ACT compos-ite), TSA did not impact academic per-formance (see Table 4). In summary,when ACT and TSA were used as pre-dictors, TSA was able to explain varia-tion in academic success that was notexplained by ACT (Table 3). However,when achievement striving and TSAwere used as predictors, TSA wasunable to explain any variation in aca-demic performance that was notexplained by achievement striving.

Results from H3 show that TSA wasa predictor and a moderator in the pres-ence of ACT composite (a quasimoder-ator). Results suggest the importance ofboth ability (i.e., ACT composite score)and behavior (TSA) measures in deter-mining academic performance (H3). Asindicated by the significant and positiveslope coefficient for the interaction termbetween ability and behavior (slope =0.18), it is simply not ability alone thatbrings about positive performance out-comes. Variables such as TSA strength-en the influence that ability has on stu-dent performance. At a time when mostefforts by administrators and instructorsare focused on curriculum and pedagog-ical issues, this study’s results show theneed to also give attention to the com-position of today’s college student pop-ulations in terms of what they bring toclass (i.e., study habits).

H4, which stated that the influencethat behavior (i.e., TSA) has on aca-demic performance would be higher forstudents with high levels of motivationthan for students with low levels ofmotivation, was not supported. In thisinstance, it is clear that, in the absenceof ability as a predictor, high levels ofmotivation or behavior will not bringabout the desired academic perfor-mance or outcome.


TABLE 4. Results of Moderated Multiple Regression Analysis of TimeSpent Outside of Class on Academic Activities (TSA), Achievement Striving (AST), and Semester Grade Point Average (SGPA)

Independent variable Slope SE t p R 2

Control (Step 1) .10*Gender –0.23 0.13 –1.81 .07Age –0.07 0.07 –1.08 .28Race 0.87 0.28 4.42 .00*Academic load 0.06 0.06 0.98 .33

Predictor (Step 2) .24*Achievement striving (AST) 0.40 0.07 6.36 .00*Time studying (TSA) 0.01 0.06 0.18 .85

Moderator (Interaction) (Step 3) .24AST × TSA 0.04 0.06 0.58 .57

*p = .05.

Implications

At a time when students spend lesstime studying and more time working,our results provide food for thought,although it may be premature to deriveimplications from the findings of thisstudy. Should subsequent researchersusing different samples validate find-ings of this study, there are implicationsfor both students and administrators.

Students

Results from studies such as this canbe passed on to students. This can beeasily done at a student orientation, instudent newsletters, on the Web, or inthe classroom. It should be clearly com-municated to them that their abilities,motivation, and behavior work in tan-dem to influence their academic perfor-mance. If students are lacking in evenone of these areas, their performanceswill be significantly lower. Once stu-dents have a better understanding ofhow ability, motivation, study time, andwork patterns influence academic per-formance, they may be more likely tounderstand their own situations and takecorrective action. More important, theymay be less likely to have unreasonableexpectations about their academic per-formance and take more individualresponsibility for its outcome ratherthan conveniently putting the blame onthe instructor. For example, it is notuncommon for intelligent students tobelieve that ability will result in highlevels of academic performance regard-less of their level of motivation or effort.The results of this study show theimpact of ability on academic perfor-mance to be much higher for studentswho spend more time studying than forthose who spend less.

Also, the results did not show a directlink between TSW and academic per-formance. Although this can be anencouraging finding at a time when alarge percentage of college students areworking longer hours while attendingcollege (Curtis & Lucus, 2001), moreresearch is needed prior to making gen-eralizations. For example, it is plausiblethat the direct relationship betweenTSW and academic performance can bemoderated by several personal (i.e.,ability, motivation, study habits) and sit-

uational (i.e., level of stress, courseload)variables, and, as such, the impact thatTSW has on academic performancemay be different for different studentpopulations under different situations orcircumstances. We did not investigatethose relationships in this study.

Administration

Study results also have implicationsfor both the recruitment and retention ofstudents. According to ACT, only 22%of the 1.2 million high school graduateswho took the ACT assessment in 2004achieved scores that would make themready for college in all three academicareas: English, math, and science (ACTNews Release, 2004a). First, universityadministrators as well as faculty shouldrealize the importance of recruiting stu-dents who are academically prepared forcollege as indicated by ACT compositeor SAT scores. Having the motivation ora strong work ethic may not bring aboutdesired performance outcomes in theabsence of ability, as evidenced by H4.This can be a potential concern for col-leges and universities that have lowadmission standards (i.e., low ACT orSAT score requirements and loweracceptable high school GPAs) or openadmission policies. Due to low admis-sion requirements, these institutions aremore likely to have a larger percentageof students who lack the minimum abil-ity needed to succeed in college com-pared with a smaller percentage of suchstudents in colleges and universities thathave high admission standards. There-fore, colleges and universities that haverelatively low admission standards needto have a process in place to identifythose students who lack the necessaryabilities (e.g., quantitative skills, verbalskills) to succeed in college and providethem with ample opportunities to devel-op those abilities while in college byoffering remedial courses. Failure todevelop those abilities prior to takingcollege-level courses can be a recipe forpoor academic performance and lowretention rates. Data compiled by ACTshow a strong inverse relationshipbetween admission selectivity anddropout rates: Highly selective = 8.7%,selective = 18.6%, traditional = 27.7%,liberal = 35.5%, and open = 45.4%(ACT Institutional Data File, 2003).

Also, on the basis of the results fromH3, students with high ability who alsospend more time studying are the oneswho are most likely to excel in college asindicated by their GPA (Figure 1). Theseare the type of students who are mostlikely to perform well academically andbring universities as well as individualprograms a high-quality academic repu-tation, and, as such, a process should bein place to recruit and retain them.

In addition to recruiting, retaining thestudents and helping them to achievetheir goals is an important issue forinstitutes of higher education. Researchresults indicate that just over half of stu-dents (63%) who began at a 4-year insti-tution with the goal of a bachelor’sdegree have completed that degree with-in 6 years at either their initial institu-tion or at another institution (U.S.Department of Education, 2002).Unfortunately, an alarming number ofschools have no specific plan or goals inplace to improve student retention anddegree completion (ACT News Release,2004b). This shows the need for insti-tutes of higher education to have theirown models to precisely predict andtrack the academic performance of theirprospective students to ultimately mon-itor and control student retention anddropout rates. Although measures ofability such as ACT and SAT scores andhigh school GPA are widely used forcollege admission and GPA at college isused to evaluate the progress of the stu-dent, the results of this study show that,if included, nonability variables such asmotivation and TSA may significantlyimprove these prediction models. Thisinformation, if collected and monitored,would be useful in terms of decisionmaking for university administrators aswell as faculty.

Limitations and Direction forFuture Research

We made significant efforts to mini-mize measurement error in variablesthat are normally self-reported, such asACT composite scores and academicperformance (GPA), as well as thosevariables that rely on memory of pastevents, such as TSA and TSW (i.e., aquestion such as time spent studying ina given week or time spent studying


the previous week). By using universi-ty data for variables such as ACT com-posite scores and academic perfor-mance as well as collecting the timedata based on a diary maintained byparticipants during a 1-week period,we minimized measurement error.Nevertheless, although results can begeneralized to the university where wecollected the data, additional evidencewill be required prior to generalizingstatements to all university settings. Inthis respect, a national sample thatinvestigates these relationships caneither support or refute this study’sfindings.

The study did not include a variablethat measured the effectiveness level orquality of the time students spentstudying, which may be one reasonwhy H1 was not supported. It is verylikely that both the time that studentsspend studying as well as how thistime is spent should be measured. Thatwas certainly a limitation of this study.Results of future studies in whichresearchers include this variable (i.e.,time management perceptions andbehaviors measured by Macan, 1994;Macan, Shahani, Dipboye, & Phillips,1990) will provide more insight intothis issue.

If TSA moderates the relationshipbetween ACT composite and academicperformance, it is plausible for TSWalso to moderate the relationshipbetween ACT composite and academicperformance. Therefore, in a futurestudy, researchers might investigatewhether the relationship between ACTcomposite score and academic perfor-mance is stronger for students whospend less time working compared withthe students who spend more timeworking. We did not investigate theserelationships because they were outsidethe scope of this study.

We limited the personality variableunder investigation to achievementstriving. Other variables such as opti-mism and self-efficacy are likely toinfluence academic performance, andfuture studies will be able to addressthese issues in more depth. However, inthis study, we addressed an importantconcern of the academic community at atime when such empirical research isnot widely available, and, as a result, we

have contributed to the higher educationliterature.

NOTE

Correspondence concerning this article should beaddressed to Sarath A. Nonis, Professor of Market-ing, Department of Management and Marketing,Box 59, Arkansas State University, State University,AR 72467. E-mail: [email protected]

REFERENCES

Ackerman, D. S., & Gross, B. L. (2003, Summer).Is time pressure all bad? Measuring betweenfree time availability and student performanceperceptions. Marketing Education Review, 12,21–32.

ACT Institutional Data File. (2003). National col-legiate dropout and graduation rates. RetrievedJanuary 9, 2005, from http://www.act.org/path/postsec/droptables/index.html

ACT News Release. (2004a). Crisis at the core:Preparing all students for college and work.Retrieved January 9, 2005, from http://www.act.org/path/policy/index.html

ACT News Release. (2004b). U.S. colleges fallingshort on helping students stay in school.Retrieved January 9, 2005, from http://www.act.org/news/release/2004/12-13-04.html

Bandura, A., & Schunk, D. H. (1981). Cultivatingcompetence, self-efficacy, and intrinsic interestthrough proximal self-motivation. Journal ofPersonality and Social Psychology, 41(3),586–598.

Barling, J., & Charbonneau, D. (1992). Disentan-gling the relationship between the achievementstriving and impatience-irritability dimensionsof Type-A behavior, performance, and health.Journal of Organizational Behavior, 13,360–378.

Barling, J., Kelloway, K., & Cheung, D. (1996).Time management and achievement strivinginteract to predict car sales performance. Jour-nal of Applied Psychology, 81(6), 821–826.

Barron, R. M., & Kenny, D. A. (1986). The mod-erator-mediator variable distinction in socialpsychological research: Conceptual, strategic,and statistical considerations. Journal of Per-sonality and Social Psychology, 51(6),1173–1182.

Carlson, D. S., Bozeman, D. P., Kacmar, M. K.,Wright, P. M., & McMahan, G. C. (2000, Fall).Training motivation in organizations: An analy-sis of individual-level antecedents. Journal ofManagerial Issues, 12, 271–287.

Chan, D., Schmitt, N., Sacco, J. M., & DeShon, R.P. (1998). Understanding pretest and posttestreactions to cognitive ability and personalitytests. Journal of Applied Psychology, 83(3),471–485.

Chatman, J. A. (1989). Improving interactionalorganizational research: A model of person-organization fit. Academy of ManagementReview, 14(3), 333–349.

Cleary, P. D., & Kessler, R. C. (1982). The esti-mation and interpretation of modified effects.Journal of Health and Social Behavior, 23(2),159–169.

Cohen, J., & Cohen, P. (1983). Applied multipleregression/correlation analysis for behavioralsciences. Hillsdale, NJ: Lawrence Erlbaum.

Cubeta, J. F., Travers, N. L., & Sheckley, B. G.(2001). Predicting the academic success ofadults from diverse populations. Journal ofCollege Student Retention, 2(4), 295–311.

Curtis, S., & Lucus, R. (2001). A coincidence ofneeds? Employers and full-time students.Employee Relations, 23(1), 38–54.

Dreher, G. F., & Bretz, R. D. (1991). Cognitiveability and career attainment: Moderatingeffects of early career success. Journal ofApplied Psychology, 76(3), 392–397.

Gose, B. (1998, January 16). More freshmen thanever appear disengaged from their studies, sur-vey finds. The Chronicle of Higher Education,A37–A39.

Higher Education Research Institute. (2002). Theofficial press release for the American freshmen2002. Los Angeles: University of CaliforniaPress.

Higher Education Research Institute. (2003). Theofficial press release for the American freshmen2002. Los Angeles: University of CaliforniaPress.

Legislative Analyst’s Office. (2001). Improvingacademic preparation for higher education.Sacramento, CA: Author.

Light, R. J. (2001). Making the most of college.Cambridge, MA: Harvard University Press.

Macan, T. H. (1994). Time management: Aprocess model. Journal of Applied Psychology,79(3), 381–391.

Macan, T. H., Shahani, C., Dipboye, R. L., &Phillips, A. P. (1990). College students’ timemanagement: Correlation with academic per-formance and stress. Journal of EducationalPsychology, 82(4), 760–768.

McFadden, K., & Dart, J. (1992). Time manage-ment skills of undergraduate business students.Journal of Education for Business, 68, 85–88.

Mouw, J., & Khanna, R. (1993). Prediction of aca-demic success: A review of the literature andsome recommendations. College Student Jour-nal, 27(3), 328–336.

Multon, K. D., Brown, S. D., & Lent, R. W. (1991).Relation of self-efficacy beliefs to academic out-comes: A meta-analytic investigation. Journal ofCounseling Psychology, 38(1), 30–38.

Nonis, S. A., & Wright, D. (2003). Moderatingeffects of achievement striving and situationaloptimism on the relationship between abilityand performance outcomes of college students.Research in Higher Education, 44(3), 327–346.

Nunnally, J. C. (1978). Psychometric theory. NewYork: McGraw-Hill.

Paden, N., & Stell, R. (1997, Summer). Reducingprocrastination through assignment and coursedesign. Marketing Education Review, 7, 17–25.

Pascarella, E. T., & Terenzini, P. T. (1991). Howcollege affects students. San Francisco: Jossey-Bass.

Pinder, C. (1984). Work motivation. Glenview, IL:Scott, Foresman.

Smart, D. T., Kelley, C. A., & Conant, J. S. (1999).Marketing education in the year 2000: Changesobserved and challenges anticipated. Journal ofMarketing Education, 21(3), 206–216.

Smart, D. T., Tomkovick, C., Jones, E., &Menon, A. (1999). Undergraduate marketingeducation in the 21st century: Views fromthree institutions. Marketing EducationReview, 9(1), 1–10.

Spence, J. T., Helmreich, R. L., & Pred, R. S.(1987). Impatience versus achievement striv-ings in the Type-A pattern: Differential effectson student’s health and academic achievement.Journal of Applied Psychology, 72(4), 522–528.

Statistical Abstract of the United States. (2002).The national databook (pp. 168–169). Wash-ington, DC: U.S. Department of Commerce andStatistics Administration.

Strauss, L. C., & Volkwein, F. J. (2002). Compar-


ing student performance and growth in 2- and4-year institutions. Research in Higher Educa-tion, 43(2), 133–161.

U.S. Department of Education. (2002). Descrip-tive summary of 1995–1996 beginning postsec-ondary students: Six years later. Washington,DC: Author.

Wise, S. L., Peters, L. H., & O’Conner, E. J.(1984). Identifying moderator variables usingmultiple regression: A reply to Darrow andKahl. Journal of Management, 10(2),227–236.

Wright, J., & Mischel, W. (1987). A conditionalapproach to dispositional constructs: The local

predictability of social behavior. Journal ofPersonality and Social Psychology, 53(6),1159–1177.

Zimmerman, B. J. (1989). A social-cognitiveview of self-regulated academic learning.Journal of Educational Psychology, 81(3),329–339.
