Short-Term and Long-Term Consequences of Achievement …siegler/418-Guthrie.pdf · 318 HARACKIEWICZ, BARRON, TAUER, CARTER, AND ELLIOT effects will be mediated through their short-term

Journal of Educational Psychology2000, Vol. 92. No. 2, 316-330

Copyright 2000 by the American Psychological Association, Inc.O022-0663/O0/J5.0O DOt: 10.1O3W0022-O663.92.2.316

Short-Term and Long-Term Consequences of Achievement Goals:Predicting Interest and Performance Over Time

Judith M. Harackiewicz, Kenneth E. Barron,John M. Tauer, and Suzanne M. Carter

University of Wisconsin—Madison

Andrew J. ElliotUniversity of Rochester

Why do some students excel in their college classes and develop interest in an academic discipline? Theauthors examined both the short-term and long-term consequences of students' achievement goals in anintroductory psychology course. Mastery goals positively predicted subsequent interest in the course, butnot course grades. Performance goals positively predicted grades, but not interest. Three semesters later,the authors obtained measures of continued interest in die discipline and long-term performance. Masterygoals predicted subsequent enrollment in psychology courses, whereas performance goals predictedlong-term academic performance. These positive and complementary effects of mastery and performancegoals on different measures of academic success are consistent with a multiple-goals perspective in whichboth goals can have beneficial consequences in college education.

College students often pursue multiple goals in their classes.Some goals may be fairly general and concern students' reasonsfor taking a class, whereas other goals are more specific to whatstudents hope to accomplish in the course—to learn as much asthey can about the subject or to obtain a high grade. These lattertypes of goals have been labeled achievement goals (Ames, 1992;Dweck, 1986; Nicholls, 1989), and they can influence how stu-dents approach, experience, and perform in their classes. A greatdeal of research has investigated the relationship between achieve-ment goals and a wide range of educationally relevant measuressuch as learning strategies, attributions for success, and academicperformance (see Ames, 1992, for a review). However, there hasbeen relatively little research attention given to motivational vari-ables such as interest and intrinsic motivation (see Harackiewicz,Barron, & Elliot, 1998, for review), and even less attention to thelong-term consequences of achievement goals.

We believe that educational research needs to consider not onlyhow well students perform in their courses, but also whetherstudents develop interest in the topics taught in those courses(Harackiewicz et al, 1998; Jackson, 1968; Maehr, 1976; Nicholls,1979). Students who perform well in an introductory course shouldbe well prepared for future courses in that discipline. Whether they

Judith M. Harackiewicz, Kenneth E. Barron, John M. Tauer, and Su-zanne M. Carter, Department of Psychology, University of Wisconsin—Madison; Andrew J. Elliot, Department of Psychology, University ofRochester.

We thank Krissie Augustine, Rachel Dean, Roberta Deppe, BrookeGallagher, Leonard Gicas, Chasity Henry, and Sarah Terry for their help inconducting this research, and we thank Suzanne Hidi, Lisa Linnenbrink,Ann Renninger, and Carol Sansone for helpful comments on drafts of thisarticle.

Correspondence concerning this article should be addressed to Judidi M.Harackiewicz, Department of Psychology, University of Wisconsin, 1202West Johnson Street, Madison, Wisconsin 53706. Electronic mail may besent to [email protected].

actually enroll in more courses in that discipline, however, maydepend on whether they become interested in the course materialand want to learn more about the subject. Our focus is on theinterest that develops during a course (Deci, 1992; Lepper &Hodell, 1989), and we believe that it can influence continuinginterest in a discipline. Thus, interest and academic performanceare both important indicators of success in the short term, and theymay also predict future success in related courses.

In fact, motivational variables may be especially important atthe college level, where students play a major role in decidingwhether they take any more courses in a particular field. Althoughcourses are often required across a general education plan andwithin fields of specialization, students typically choose the fieldin which to concentrate their studies, and they often have consid-erable latitude in selecting the particular courses they take. Theymay choose their subsequent courses on die basis of their interestin the material, their prior performance in related courses, or somecombination of these factors, and achievement goals may play acritical role in this motivational process.

In sum, the present research has two major objectives. The firstis to examine the role of students' achievement goals in promotingtwo important short-term educational outcomes: interest and per-formance in an introductory college course. The second is toexamine the longer term consequences of achievement goals bytracking their effects on subsequent academic performance andcontinued interest beyond an introductory course.

Achievement Goals and Short-Term Consequences

Recent years have seen a convergence of theory and researcharound the constructs of mastery and performance goals, whichrepresent two different ways of pursuing competence in achieve-ment situations (Ames, 1992; Dweck & Leggett, 1988; Nicholls,1989). Mastery goals are self-referential and focus on learning anddeveloping skills. Performance goals are normative in nature andfocus on demonstrating competence. Of course, not all students arepositively motivated in the classroom, and theorists have also

316

ACHIEVEMENT GOALS 317

identified work avoidance goals, which focus on effort minimiza-tion (Brophy, 1983; Nicholls, 1989). Each of these goal orienta-tions fosters distinct patterns of cognition, affect, and behavior. Forexample, mastery goals have been linked to a number of adaptivelearning behaviors such as task involvement, challenge seeking,and deep processing of course material (Ames & Archer, 1988;Nicholls, Cheung, Lauer, & Patashnick, 1989; Nolen & Haladyna,1990). Performance goals have been linked to maladaptive learn-ing strategies such as challenge avoidance and surface level pro-cessing (E. S. Elliott & Dweck, 1988; Nolen, 1988). Work avoid-ance goals have been associated with the most negative effects onlearning (Archer, 1994; Duda & Nicholls, 1992; Meece, 1991;Meece, Blumenfeld, & Hoyle, 1988).

For these reasons, achievement goal theorists have long advo-cated mastery goals as opposed to performance or work avoidancegoals as the optimal approach to course work. Although there islittle controversy concerning the effects of work avoidance goals,there has been more recent debate about the effects of performancegoals. In fact, results have recently begun to accumulate suggest-ing that strong conclusions about the negative consequences ofperformance goals may be premature (Harackiewicz et al., 1998;Hidi & Harackiewicz, in press; Urdan, 1997). A number of theo-rists have suggested that the performance goal construct may betoo general, confounding theoretically distinct and separable com-ponents. For example, Elliot and colleagues (A. J. Elliot, 1997;A. J. Elliot & Harackiewicz, 1996) distinguish between approachand avoidance performance goals, and Wolters, Yu, and Pintrich(1996) distinguish between extrinsic goals and relative abilitygoals, which are similar to performance-approach goals. Whenboth components are taken into consideration, performance-avoidance and extrinsic goals have been linked to the most nega-tive outcomes (A. J. Elliot & Church, 1997; Middleton & Midgley,1997; Skaalvik, 1997; Wolters et al., 1996). When the definition ofperformance goals is constrained to positive approach strivingstoward normative competence (which is the definition we will usethroughout this paper), a different pattern of findings emerges.

In a recent review, Harackiewicz et al. (1998) noted that per-formance goals are both conceptually and empirically independentof mastery goals and that positive mastery goal effects do notnecessarily imply negative performance goal effects. They exam-ined studies in which researchers tested the independent effects ofmastery and performance (approach) goals and found several pos-itive effects of performance goals on measures of adaptive learningstrategies and academic performance (Archer, 1994; Bouffard,Boisvert, Vezeau, & Larouche, 1995; A. J. Elliot & Church, 1997;Harackiewicz, Barron, Carter, Lehto, & Elliot, 1997; Kaplan &Midgley, 1997; Midgley & Urdan, 1995). For example, Harac-kiewicz et al. (1997) found that performance goals positivelypredicted academic performance in introductory college classes,whereas mastery goals did not. They argued that performancegoals can promote academic performance because they are con-sistent with the general context in which performance is assessedand grades are assigned. Specifically, large lecture classes incollege typically use normative grading structures that explicitlydefine success in terms of outperforming others. Moreover, intro-ductory college courses tend to rely on multiple-choice exams thatmay assess superficial rather than deep understanding. Either ofthese factors may create a performance-oriented classroom envi-ronment. In such contexts, students who adopt performance goals

may actually be striving to attain good grades in an optimalmanner.

In contrast, Harackiewicz et al. (1997) found that mastery goalspredicted interest in an introductory class, whereas performancegoals did not. Thus, each goal was positively associated with oneindicator of success (interest or performance) but not the other,suggesting that students who endorsed both goals were the mostlikely to attain both outcomes. Furthermore, A. J. Elliot andChurch (1997) found a similar pattern of results in an intermediate-level college course at a different university. Considered together,the results support a multiple-goals model, in which adopting bothmastery and performance goals can facilitate motivation and per-formance in college classes (Barron & Harackiewicz, in press).

Achievement Goals and Long-Term Consequences

The goals that students adopt in an introductory course may alsoinfluence their motivation and performance in subsequent coursesin that discipline. Although Harackiewicz et al. (1997) found thatperformance goals were the sole predictor of grades in an intro-ductory psychology course, a different pattern of results mightemerge over time. Mastery goals could prove advantageous inmore advanced courses where course work may require deepprocessing, thoughtful integration of the course content, and sus-tained effort and involvement. Mastery goals might also haveindirect effects on later grades by fostering interest in a particulardiscipline, which may facilitate learning and subsequent perfor-mance in related courses (Alexander, Kulikowich, & Jetton, 1994;Hidi, 1990; Maehr, 1976; Sansone & Harackiewicz, 1996). Thus,we might see direct or indirect effects of mastery goals on perfor-mance in later courses in that discipline.

However, even as advanced courses require more integrativework from students, the general climate of later college coursesmay continue to be performance oriented as professors evaluatestudents' work relative to others and assign grades on a normativecurve. In this case, we should continue to find positive perfor-mance goal effects on subsequent academic performance, andthese effects may be mediated by students' performance in theintroductory course. In other words, performance goals may facil-itate subsequent performance by promoting learning and perfor-mance in an introductory course that provides a strong foundationfor later courses. In sum, mastery and performance goals mightboth contribute to academic performance over the long term.

As noted earlier, college students typically decide for them-selves whether to enroll in further courses, and their subsequentcourse choices may serve as a behavioral indicator of their con-tinued interest in a particular discipline (Maehr, 1976). We wouldexpect to find positive mastery goal effects on continued interest,and these effects should be mediated by students' interest in anintroductory course. In other words, the interest developed in anintroductory course should promote continued interest in a partic-ular field, and mastery goals may promote continued interestthrough their effects on interest in an introductory course. Perfor-mance goals might also have indirect effects on continued interest,if students choose their subsequent courses on the basis of theirprior performance in related courses.

In sum, although mastery and performance goals may havesome indirect effects over time, we predict that long-term goal

318 HARACKIEWICZ, BARRON, TAUER, CARTER, AND ELLIOT

effects will be mediated through their short-term consequences.Specifically, mastery goals should affect continued interest in aparticular discipline by promoting interest in an introductory class,and performance goals should influence subsequent performanceby promoting performance in an introductory course.

Interest: Catch Versus Hold Components

In addition to our primary objectives, a third objective of thisresearch was to explore the nature of interest in college classes.Theorists have identified two types of interest: individual interest,which reflects a relatively enduring predisposition to reengage inparticular activities over time, and situational interest, which isused to characterize engagement that arises in a particular context(Krapp, Hidi, & Renninger, 1992). As Alexander (1997) pointsout, both types of interest may contribute to the development ofcontinuing interest in a topic. For example, the decision to enrollin a college course reveals a certain level of interest, and somestudents may enter a course with knowledge about the topic andindividual interest (Alexander, Kulikowich, & Schulze, 1994; Ren-ninger, 1998a; Schiefele & Krapp, 1996). This preexisting interestmay be strengthened and deepened in the course. Other studentsmay only develop continuing interest if the course stimulates theirinterest (Hidi & Harackiewicz, in press; Hidi & McLaren, 1990,1991).

Hidi and Baird (1986) raised an important distinction betweenfactors that "catch" interest and those that "hold" it (Mitchell,1993). In a college classroom, for example, an entertaining lectureor high-tech audiovisuals might grab students' attention and catchtheir interest, but course material that is inherently interesting orthat becomes personally valued may better hold students' interestover time (Malone & Lepper, 1987; Rathunde, 1993; Schiefele,1991). Thus, interest based on hold factors should predict continu-ing interest in a discipline more strongly than interest based oncatch factors.

Mediators and Moderators

A fourth objective of the current research was to examinepotential mediators and moderators of achievement goal effects inthe short term. First, to explore the link between achievement goalsand academic performance, we measured students' study strategiesand tested whether they mediated goal effects. A number oftheorists have distinguished between two dimensions of learningstrategies, and although labels and measures differ, these distinc-tions converge on one between deep or elaborative processing ofmaterials versus surface or shallow learning (Biggs, 1993; Marton,Hounsell, & Entwistle, 1984; Nolen, 1988; Pintrich & Schrauben,1992; Schmeck, 1983). We therefore included measures of bothdimensions to assess the link between students' achievement goalsand their self-reported strategy use and, in turn, the link betweenstrategy use and performance in their introductory class. Second,Dweck and Leggett (1988) argued that negative performance goaleffects may only be revealed when students experience lowerlevels of perceived competence. Therefore, in this study, we alsomeasured students' performance expectations at the outset andmidpoint of the semester to test whether the effects of performancegoals were moderated by performance expectations.

Our Study

We used a prospective longitudinal design to examine the con-sequences of students' achievement goals over the course of anacademic semester, measuring achievement goals and performanceexpectations at the outset of an introductory psychology course,study strategies and performance expectations at the midpoint ofthe semester, interest (both catch and hold factors) near the end ofthe course, and performance at the end of the course. We thentracked students' course choices and subsequent academic perfor-mance over the following three semesters. To measure continuedinterest, we counted the number of psychology credits taken overthe subsequent three semesters and also counted the total numberof credits taken over that same time period.

The measurement of academic performance was more complex.To examine performance in the introductory class, we measuredfinal course grades. Then, for those students who did enroll inadditional psychology courses, we computed a grade-point average(GPA) for their subsequent psychology courses. Because we couldnot measure subsequent performance in psychology courses for allstudents, however, we also examined more general measures onwhich we could compare short- and long-term effects for allstudents. Specifically, we computed a GPA for the semester of theintroductory course, as well as an overall GPA for subsequentsemesters. Thus, we collected both psychology-specific and gen-eral measures of performance in both the short and long term.

We expected that achievement goals for the introductory psy-chology course would predict interest and specific measures ofperformance and that they might also be related to more generalmeasures of performance. Specifically, we predicted that masterygoals, but not performance goals, would be associated with interestin the short term, as well as continued interest in psychology. Incontrast, we predicted that performance goals, but not masterygoals, would be associated with performance in the short term (asin Harackiewicz et al., 1997, and A. J. Elliot & Church, 1997), butthat both mastery and performance goals would have positiveeffects on subsequent performance in psychology courses.

Method

Overview

The initial component of the study took place over the course of asemester in introductory psychology classes at a large, Midwestern uni-versity and consisted of three assessment waves during the semester. Wemeasured students* achievement goals for the class 2 to 3 weeks into thesemester, their study strategies midway through the semester, and theirinterest in the class near the end of the semester. We also obtained theirfinal grades in the course. Three semesters later, we obtained students'academic records to measure their subsequent course choices and grades.

Participants and Setting

Students were recruited from four sections of Introductory Psychology.This is a general survey course, typically taken by freshmen and sopho-mores before they have selected a major. Departmental records indicatethat approximately 10% of the students who take Introductory Psychologygo on to major in psychology. Sections of the course were taught by oneof two instructors (average enrollment of 355 students per section). Thesize of the classes and the lecture format necessarily limited opportunitiesfor instructor-student interactions. Students' grades were determined by


their performance on multiple-choice exams, and grades were assignedaccording to normative curves recommended by the Psychology Depart-ment. A sample of 648 introductory psychology students (218 men and 430women) participated in the initial study and received extra credit forparticipating.

Goals Wave

Two to three weeks into the term, after students became familiar with thecourse and their instructor but before they had taken any quizzes or exams,they completed an 18-item questionnaire. The questionnaire was designedto assess students' self-reported adoption of mastery, performance, andwork avoidance goals in their introductory psychology class. Items werebased on those used in Harackiewicz et al. (1997), which were originallyadapted from Meece et al. (1988) and Pintrich (Pintrich & De Groot, 1990;Pintrich, Smith, Garcia, & McKeachie, 1993). These scales differedslightly from those used by Harackiewicz et al. (1997). The major differ-ence was that items in this study emphasized goals adopted in the specificclass, as opposed to more general goal orientations. Because our scaleswere developed for use in college classes, they cannot be directly comparedwith other scales developed for elementary, middle-school, and high-school populations. However, our measures of mastery and performancegoals appear quite comparable to the mastery and relative ability scalesdeveloped by Midgley et al. (1998).

In addition, three items assessed students' performance expectations forthe class. Students were instructed to indicate the extent to which theybelieved each item to be true of them on a 1 (not at all) to 7 (very) scale.The mastery, performance, work avoidance, and performance expectationsitems appear in Table 1.

Study Strategies Wave

At mid-semester (7 to 8 weeks into the term), students completed a16-item questionnaire. The questionnaire was designed to measure stu-

Table 1Goals Wave

dents' strategies for studying their psychology course material and toobtain another measure of performance expectations. Study strategy itemswere adapted from Pintrich and De Groot (1990) and Pintrich et al. (1993)to reflect recent theorizing on cognitive styles for learning. In particular,items measured the extent to which students used elaboration and rehearsallearning strategies. Elaboration strategies reflect a deep-level approach tolearning (e.g., trying to connect and integrate information that is beinglearned with what one already knows), whereas rehearsal strategies reflecta more superficial and surface-level approach to learning (e.g., simplytrying to memorize important information). In addition, we assessedwhether students lacked a study strategy. Study strategy and mid-semesterperformance expectations items appear in Table 2.

Interest Wave

Near the end of the semester, but before taking their last exam andreceiving their final course grades, students' interest in the class wasmeasured. Students indicated their agreement with each of 13 items on a 1(strongly disagree) to 7 (strongty agree) scale. Items were derived from ourprevious work (Harackiewicz et al., 1997), but we also added new items totest for a distinction between enjoyment of the class (catch) and moresubstantive interest in psychology (hold). Interest items are presented inTable 3.

Grades

Final grade in psychology. Students' final course grades were obtainedfrom departmental records. Students could receive 1 of 8 possible grades,based on the university's 4-point scale. The average grade for students inour study was 2.82 (SD = .92). Grades were distributed as follows:A = 20.7%, AB = 17.9%, B = 14.6%, BC = 16.1%, C = 23.2%,D = 6.8%, and F = 0.8%. The average grade for students in our samplewas higher than the average grade for all students enrolled in Introductory

Scale Items

Mastery

Performance

Work Avoidance

Initial Performance Expectations

1. I want to learn as much as possible in this class.2. In a class like this, I prefer course material that really challenges

me so I can learn new things.3. The most important thing for me in this course is trying to

understand the content as thoroughly as possible.4. Understanding psychology is important to me.5. I like it best when something I learn makes me want to find out

more.6. In a class like this, I prefer course material that arouses my

curiosity, even if it is difficult to learn.1. It is important for me to do better than other students.2. My goal in this class is to get a better grade than most of the other

students.3. It is important for me to do well compared to others in this class.4. I want to do well in this class to show my ability to my family,

friends, advisors, or others.5. Getting a good grade in this class is the most important thing for

me right now.6. It is important for me to establish a good overall grade-point

average, so my main concern in this class is getting a good grade.1. I want to do as little work as possible in this class.2. I just want to do as much as I have to in order to get by in this

class.3. I like my classes best when the tests are easy.1. Considering the difficulty of this course and my skills, I think I

will do well in this class.2. I expect to do well in this class.3. I think I will do poorly in this class (reversed).


Table 2Study Strategy Wave Items and Scales

Scale Items

Elaboration

Rehearsal

Lack of Strategy

Mid-Semester Performance Expectations

1. When I study, I try to explain the key concepts in my ownwords.

2. When reading, I try to connect the things I am readingabout with what I already know.

3. When I study for a test, I try to put together theinformation from class and from the book.

4. When I am studying a topic, I try to make everything fittogether.

5. When reading the text, I am interested in the "big picture"and how concepts relate to each other.

1. When I study for this course, I go through the readingsand my class notes and try to find the most importantideas to memorize.

2. When I study for this class, I practice saying importantfacts to myself over and over.

3. When I study, I try to memorize as many facts as I can.4. When studying for this class, I read my class material to

myself over and over again.5. When studying, I make up memory cues that I can use

during tests to help me recall the material.1. I'm not sure how to study for this course.2. I often find that I don't know what to study or where to

start.3. I really don't have a strategy for studying for this class.1. Considering the difficulty of this course and my skills, I

think I will do well in this class.2. I expect to do well in this class.3. I think I will do poorly in this class (reversed).

Psychology (M = 2.58; N = 1,542), but the distribution of grades acrossall sections was comparable to that of our sample.

Semester GPA. We also obtained students' grades for all other coursestaken that semester and computed a GPA for the semester (semester GPA).

Long-Term Follow-Up

Three semesters later, we obtained students' permission to access theiracademic records. We used these records to count the number of coursecredits taken in psychology over the three semesters, which provided abehavioral measure of subsequent interest in psychology (courses taken).

At this university, psychology courses range from 1 to 5 credits each, andthe typical lecture course is taken for 3 credits. We also computed anoverall GPA for all courses taken in the three semesters following thesemester of the initial study (subsequent GPA) and one specifically forsubsequent psychology courses (psychology GPA).

A total of 648 students participated in the initial study, but 33 studentscould not be reached for permission to access their subsequent records. Ofthe 615 students who participated in the initial study and gave permissionfor access to their academic records, 11 took no more college courses afterIntroductory Psychology (9 dropped out of college and 2 graduated). Thus,

Table 3Interest Wave

Scale Items

Interest in Psychology

Enjoyment of Lectures

1. I think what we are learning in this class is interesting.2. I think I will be able to use what I learn in this course in other courses.3. I would recommend this class to others.4. I am enjoying this psychology class very much.5. I think the field of psychology is very interesting.6. This class has been a waste of my time (reversed).7. I'm glad I took this class.8. I think the course material in this class is useful for me to learn.9. I would like to take more psychology classes after this one.

10. I am more likely to register for another psychology class because ofmy experience in introductory psychology.

1. The lectures in this class really seem to drag on forever (reversed).2. I don't like the lectures very much (reversed).3. I like my professor.


the final sample (used in all data analyses, unless otherwise noted) included604 students (204 men and 400 women).

Results

Table 4Descriptive Statistics for Goals, Study Strategy,and Outcome Variables

Scale Construction

To verify whether the items assessed at each wave conformed toour a priori classification, a series of factor analyses was con-ducted. Scales were constructed with unweighted linear combina-tions of variables with sufficiently high loadings on primary fac-tors and sufficiently low cross-loadings. For all analyses, itemsloaded at least .50 on the primary factor and below .30 on otherfactors. For the goals wave questionnaire, principal componentsfactor analysis with varimax rotation yielded a four-factor solutionreflecting our a priori identification of Mastery, Performance,Work Avoidance, and Performance Expectations factors. The eig-envalues for Mastery, Performance, Work Avoidance, and Perfor-mance Expectations were 3.72, 3.42, 1.30, and 1.70, respectively.Cronbach's alphas reached acceptable levels for the six-item Mas-tery (a = .79), the six-item Performance (alpha = .80), and thethree-item Performance Expectations (a = .79) scales. The three-item Work Avoidance scale was not as internally consistent (a =.51) and was therefore not included in the primary analyses. Wedid however conduct supplemental analyses including work avoid-ance goals for comparisons with previous research that will bereported separately as ancillary analyses.

For the study strategies wave questionnaire, principal compo-nents factor analysis with varimax rotation yielded a four-factorsolution mirroring our a prior classification of elaboration, re-hearsal, lack of strategy, and performance expectations items. Theeigenvalues for Elaboration, Rehearsal, Lack of Strategy, andPerformance Expectations were 2.48,1.40,1.10, and 4.75, respec-tively. Cronbach's alphas reached acceptable levels for the five-item Elaboration (a = .76), the five-item Rehearsal (a = .61), thethree-item Lack of Strategy (a = .72), and three-item PerformanceExpectations (a = .82) scales.

For the 13-item interest wave questionnaire, we used a principalcomponents factor analysis with obiinun rotation because weexpected factors to be correlated. This analysis yielded two factors,one with 10 items (eigenvalue — 7.05; a — .93) reflecting generalinterest in psychology (Interest in Psychology), and one with threeitems (eigenvalue = 1.22; a = .72) reflecting specific enjoymentof the class (Enjoyment of Lectures).

Descriptive and Correlational Analyses

The means and standard deviations for all variables are reportedin Table 4; zero-order intercorrelations are reported in Table 5,Although the structure, content, and grading distributions of thefour class sections were comparable, we tested for instructor andsection differences in all variables. The only significant instructoreffect was on Enjoyment of Lectures, indicating that studentsenjoyed sections taught by one instructor more than the other.Therefore, we included a contrast code term to test and control fordifferences between the two instructors in all subsequent analyses(Cohen & Cohen, 1983).

Regression Models

Overview. We conducted a series of multiple regression anal-yses to examine the effects of achievement goals on outcome

Variable

Mastery goalsPerformance goalsWork avoidance goalsInitial performance expectationsElaborationRehearsalLack of strategyMid-semester performanceExpectationsInterest in psychologyEnjoyment of lecturesFinal gradeSemester GPASubsequent GPAPsychology GPA (n = 148)Courses taken (no. of credits)

Range

1-71-71-71-71-71-71-71-7

1-71-7( M0-40-40-40-27

M

5.365.213.785.505.284.923.145.57

5.024.932.822.973.072.S82.28

SD

0.841.061.070.950.890.961.321,11

1.191.190.920.640.550.814.41

Note. GPA = grade-point average.

variables measured during the semester in which students tookIntroductory Psychology (the short term) and on measures col-lected in the follow-up assessment (the long term). This dataanalytic strategy allowed us to retain the continuous nature of thevariables and to test the independent effects of each goal variableas well as the interactions between them. For the following anal-yses, all main effect terms were standardized, and multiplicativetwo- and three-way interaction terms were created with thesevariables (Aiken & West, 1991). We set alpha at .01 as ourcriterion for significance because our large sample size yieldedconsiderable statistical power.

Regression model for short-term analyses. To investigate thedirect effects of goals on the short-term outcomes, interest, enjoy-ment, final grade, and semester GPA were each regressed on apreliminary model that included the goals measures and alsocontrolled for gender and instructor effects. Initial regression mod-els included the main effects of mastery and performance goals,gender (coded +1 for females and - 1 for males), instructor (coded+1 for one instructor and — 1 for the other), and all two- andthree-way interactions among these five variables. Preliminaryanalyses revealed no significant interactions with gender or in-structor on any measures, and thus the only interaction term thatwe retained in our final model was that between mastery andperformance goals. We therefore tested a basic model that simul-taneously tested four main effects and one interaction.

Short-Term Effects of Goals

Interest in psychology. The overall model was significant, F(5,598) = 27.10, p < .001, R2 ~ .18. A main effect was found formastery goals, F(l, 598) = 121.86, p < .001, such that studentswho adopted mastery goals reported higher levels of interest inpsychology than students who did not endorse mastery goals (/3 =.41). Neither the main effect of performance goals (/3 = -.06) northe Performance X Mastery goal interaction (/3 = .00) was sig-nificant, and no other effects were significant.


Table 5Zero-Order Correlations for Goals and Outcome Variables

Variable 1 10 11 12 13 14 15

1. Mastery goals2. Performance goals3. Work avoidance goals4. Initial performance expectations5. Elaboration6. Rehearsal7. Lack of strategy8. Mid-semester performance

expectations9. Interest in psychology

10. Enjoyment of lectures11. Final grade12. Semester GPA13. Subsequent GPA14. Psychology GPA (n = 148)15. Courses taken

0529*30*37*13*16*23*

42*23*0301050518*

—.24*.15*.10.22*.04.17*

-.04.01.14*.12*.12*.21*

-.01

—-.09-.12*

.00

.19*-.13*

- .23*- . 2 1 *- .11*-.10-.06

.03-.06

—.13*.01

-.30*.48*

.14*

.14*

.16*

.10

.06

.20

.05

—.37*

-.30*.29*

.23*

.14*

.16*

.12*

.12*

.13

.09

—-.13*

.13*

.19*

.13*

.06

.09

.13*

.18

.09

—-.55*

-.26*-.20*- . 3 1 *-.24*-.18*-.13-.02

—

3 1 *.19*.37*.25*.20*.28*.06

.58*

.26*

.09

.06

.06

.27*

—.19*.12*.09.09.06

—.80*.60*.55*.11*

—.68* —.57* .75* —.07 .04 .04

Note. GPA = grade-point average.*p< .01.

Enjoyment of the lectures. The overall model was significant,F(5, 598) - 26.35, p < .001, Z?2 = .18. The main effects ofmastery goals, gender, and instructor were significant, F(l,598) = 33.25,11.76, 82.42, respectively, allps < .001. Individualswho adopted mastery goals enjoyed the lectures more (/3 = .22).Women enjoyed the class more (/3 = .13), and students of oneinstructor enjoyed their class more than students of the otherinstructor (0 = .34). Neither the main effect of performance goals(fi ~ -.01) nor the Performance X Mastery goal interaction (0 =.01) was significant.

Final grade. The overall model was significant, F(5,598) = 4.46, p < .001, R2 = .04, and a main effect was found forperformance goals, F(l, 598) = 12.97, p < .001. Students whoadopted performance goals achieved higher grades in the coursethan those who did not endorse performance goals (j3 = .15).Neither the main effect of mastery goals (0 = .02) nor thePerformance X Mastery goal interaction (0 = -.07) was signif-icant, and no other effects were significant.

Semester GPA. The overall model was significant, f(5,598) = 5.91, p < .001, J?2 = .05, and the main effects ofperformance goals and gender were significant, F(l, 598) = 10.77and 13.89, respectively, ps < .01. Students who adopted perfor-mance goals in their psychology class attained higher grades in alltheir classes that semester (j3 = .13). In addition, women attainedhigher grades than men (£ = .15). No other effects were signifi-cant. Figure 1 displays a path model illustrating the short-termconsequences of goals.

The Effects of Interest and Enjoyment on Performance

Because we measured interest and enjoyment before the end ofthe semester, it was also possible to test these variables as predic-tors of graded performance in the class. However, students hadalready received feedback on two exams by the point in thesemester that we measured interest, and it is therefore possible thattheir knowledge about their performance to data affected theirreports of interest and enjoyment. Thus, the causal ordering be-

tween interest, enjoyment, and final grade is more ambiguous thanfor other variables tested in our process models.

Nonetheless, we added the main effects of interest and enjoy-ment to the basis model on final grade. The overall model wassignificant, F(7,596) = 10.48, p < .001, R2 = .11, and accountedfor significantly more variance than the basic model, p < .001. Themain effect of interest was significant, F{\, 596) = 31.58, p <.001, but the main effect of enjoyment was not, F < 1. Studentswho expressed more interest in psychology performed better in theclass (j3 = .29). In this model, the only effect that had beensignificant in the basic model remained significant. Figure 2 dis-plays a path model illustrating these effects.

Mediators and Moderators of Short-Term Goal Effects

Study strategies are one hypothesized link between achievementgoals and performance (Pintrich & Schrauben, 1992). We there-

MasteryGoal

Instructor

Gender

PerformanceGoal

• Interest

Enjoyment

Final Grade

Semester GPA

Figure 1. Path model of the effects of mastery and performance goals,gender, and instructor on interest, enjoyment, final grade, and semestergrade-point average (GPA). Paths represent significant (p < .01) effects,and path coefficients are standardized regression coefficients.


MasteryGoal .41 • Interest- .29- ~* Final Grade

.22\

Instructor —.34—* Enjoyment

.13Gender

PerformanceGoal

Figure 2. Path model of interest and enjoyment as predictors of finalgrade. Paths represent significant (p < .01) effects, and path coefficientsare standardized regression coefficients.

fore conducted analyses, following the procedures outlined byJudd and Kenny (1981), to test whether the direct goal effects onfinal grade established in the previous analyses were mediated orpartially mediated by students' self-reported study strategies. Toreveal mediatorial effects, three criteria must be satisfied. First, asdocumented previously, a direct effect between a predictor and anoutcome must be documented. Next, a link between the predictorand the hypothesized mediator must be established. Finally, whenthe original predictor and mediator are tested simultaneously aspredictors of the outcome, the mediator should significantly affectthe outcome variable. Moreover, the effect of the original predictorshould be substantially or completely reduced to document partialor complete mediation, respectively. Therefore, to establish thatachievement goals were linked to the hypothesized mediators, weregressed each of the study strategy variables on the basic regres-sion model.

Elaboration. The overall model was significant, F(5,598) = 20.54, p < .001, R* = . 15, and the main effect of masterygoals was significant, F(l, 598) = 86.43, p < .001. Students whoadopted mastery goals were more likely to report using elaborationstrategies than students who did not endorse mastery goals O =.36). No other effects were significant.

Rehearsal The overall model was significant, F(5,598) = 11.82, p < .001, R1 = .09. Main effects were found forperformance goals and gender, F(l, 598) = 30.44 and 15.99,respectively, p < .001. Students who adopted performance goalswere more likely to use rehearsal strategies ((3 = .22), and womenreported more rehearsal use than men (/3 = .16). No other effectswere significant.

Lack of strategy. The overall model was significant, F(5,598) = 5.01, p < .001, R2 = .04. A main effect was found formastery goals, F(l, 598) = 17.68, p < .01. Students who adoptedmastery goals were less likely to report being confused about howto study for their psychology course (0 = -.17). No other effectswere significant.

Mediational effects on final grade. A direct effect of perfor-mance goals on final grade was documented earlier (see Figure 1),so the three study strategy variables were added to the basic modelto test for possible mediation. The overall model was significant,F(8,595) = 11.90, p < .001, R2 = .14. Although a main effect wasfound for lack of strategy, F(l, 595) - 56.71,/? < .001,0 = - .30,neither elaboration nor rehearsal strategies had significant effects.

Moreover, the performance goal effect remained significant, F(l,595) = 17.78, p < .001, )3 = .17. Study strategies therefore failedto mediate this effect. Figure 3 displays a path model illustratingthe predictors and correlates of study strategies.

Performance expectations. Dweck and Leggett (1988) arguedthat performance goals should have their most detrimental conse-quences at lower levels of perceived competence. We thereforeconducted analyses to test whether the effects of performancegoals were moderated by performance expectations. We measuredperformance expectations at two points in time: in the goals waveat the outset of the class and in the study strategies wave, midwaythrough the semester, when students had received some perfor-mance feedback in the class. We regressed interest, enjoyment,final grade, and semester GPA on two models, each of whichincluded one measure of performance expectations and its inter-actions with the two goals measures. First, we added the maineffect of initial performance expectations and the two-way inter-actions with mastery and performance goals to the basic modeltested earlier. There was a significant main effect of initial perfor-mance expectations on final grade, F(l, 595) = 11.64, p < .001,0 = .15, and on semester GPA, F(l, 595) = 7.71,/? < .01, ̂ =.12, but there were no significant interactive effects of perfor-mance expectations on any measure. Moreover, the basic patternof findings reported earlier remained the same.

Next, we added the main effect of mid-semester performanceexpectations and the two-way interactions with mastery and per-formance goals to the basic model tested earlier. In these regres-sions, the main effect of mid-semester performance expectationswas significant on interest, F(i, 595) = 41.69, p < .001, j3 = .25,enjoyment, F(l, 595) = 17.77, p < .001, j3 = .17, final grade, F(l,595) = 90.60, p < .001, 0 = .38, and semester GPA, F(l,595) = 45.05, p < .001, ji = .28. Students who believed theywould perform well in their psychology course reported higherlevels of interest and enjoyment and obtained higher grades in thatcourse and all other courses that semester. However, there were nosignificant interactions of performance expectations with goals on

MasteryGoal

.36

-.17

Elaboration

Lack ofStrategy

Gender .16 -Rehearsal-.30

.22

Performance.,Goal

-.17- Final Grade

Figure 3. Path model of the predictors and consequences of study strat-egies. Paths represent significant (p < .01) effects, and path coefficientsare standardized regression coefficients.


any measure, and the pattern of findings reported earlier was notaltered by the inclusion of these mid-semester expectations terms.

Long-Term Effects of Goals

Regression model for long-term analyses. To investigate thedirect effects of goals on the long-term outcomes, courses taken,subsequent GPA, and psychology GPA were regressed on thebasic five-term model established in the short-term analyses. Next,to examine the long-term effects in the context of the short-termeffects documented earlier, we added interest, enjoyment, and ameasure of graded performance to the basic model. This secondmodel allowed us to test for mediated and/or indirect effects ofgoals on the long-term consequences. When the long-term out-come variable was a general GPA measure (subsequent GPA), weused a general short-term measure (semester GPA) to control forinitial performance; when the long-term outcome was specific topsychology (courses taken or psychology GPA), we used a specificshort-term measure (final grade) to control for performance. Al-though the regression models tested were identical for the threelong-term outcome measures, the psychology GPA analyses wereconducted with the subset of the sample that actually took addi-tional psychology courses, and these analyses will be reportedseparately.

Direct effects on long-term outcomes. The overall model oncourses taken was significant, F(5, 598) = 5.10, p < .001, R2 =.04, and the only significant predictor was mastery goals, F(l,598) = 16.8, p < .001, 0 = .16. Students who adopted masterygoals in Introductory Psychology were more likely to take morepsychology courses over the next three semesters. The overallmodel on subsequent GPA was significant, F(5, 598) — 5.51, p <.001, i?2 = .04, and two terms were significant. There was a maineffect of performance goals, F(l, 598) = 10.68, p< .01,j3 = .13,and a main effect of gender, F(1,598) = 15.03, p< .001,0 = .16.Students who adopted performance goals in Introductory Psychol-ogy had higher grades over the course of three subsequent semes-ters, and women obtained higher grades over the following semes-ters. Figure 4 displays a path model illustrating the direct effects ofgoals on the measures of long-term consequences.

Indirect and mediated effects on courses taken. The additionof interest, enjoyment, and final grade to the model on courses

MasteryGoal .16 -*• Courses

Taken

Gender

PerformanceGoal

^ SubsequentGPA

Figure 4. Path model of the direct effects of mastery and performancegoals and gender on courses taken and subsequent grade-point average(GPA). Paths represent significant (p < .01) effects, and path coefficientsare standardized regression coefficients.

taken resulted in a significant overall model, F(8,595) = 8.57, p <.001, R2 = .10, and it accounted for significantly more variancethan the direct effects model, p < .001. In this model, the maineffect of interest was significant, F(l, 595) = 33.91, p < .001, J3 =.31. We found no evidence of indirect performance goal effects (pfor final grade = .05, ns). The fact that the mastery goal effect wasreduced in size (J3 = .06, ns, from .16 in the direct effects model)indicates that the direct mastery goal effect was partially mediatedthrough interest (Judd & Kenny, 1981).

It is interesting to note that we also found a significant negativeeffect of enjoyment, F(\, 595) = 8.75, p < .01, 0 = - .15 ,reflecting a suppression effect (Cohen & Cohen, 1975; Conger,1974). Suppression occurs when one variable increases the pre-dictive validity of another variable when both are tested in aregression equation, revealing a beta for one variable that is largerthan its zero-order correlation with the dependent variable. Inaddition, the beta for the suppressor variable frequently changessign from its zero-order correlation. When interest and enjoymentwere tested simultaneously as predictors of courses taken, thevalue of the beta for interest (.31) was greater than the zero-ordercorrelation between interest and courses taken, r(604) = .27, p <.05. In addition, the beta for enjoyment (—.13) did indeed changesign from its zero-order correlation with courses taken, r(604) =.06, p > .05. Thus, controlling for enjoyment (the suppressorvariable) increased the association between interest and coursestaken. No other effects were significant.

We also tested this model with a revised version of our interestmeasure. Because two of the items referred to taking more psy-chology classes, we removed them from our Interest scale andtested a revised eight-item interest measure as a predictor ofcourses taken. In this model, the eight-item Interest scale was asignificant predictor of courses taken, F(l, 595) = 20.65, p <.001, 0 = .25, but enjoyment was not (j3 = -.12).

Indirect and mediated effects on subsequent GPA. The addi-tion of interest, enjoyment, and semester GPA to the model onsubsequent GPA resulted in a significant overall model, F(8,595) = 65.39, p < .001, R2 = .47, and accounted for significantlymore variance than the direct effects model, p < .001. The maineffect of semester GPA was the only significant predictor, F(l,595) = 471.21, p < .001, 0 = .67. Students who received highergrades in the semester in which they took Introductory Psychologyreceived higher grades over the next three semesters. Thus, wefound no evidence of indirect mastery goal effects (/} for interest =- .05 , ns; j3 for enjoyment = .01, ns). The fact that the perfor-mance goal effect was reduced in size (j8 = .04, ns, from .13 in thedirect effects model) indicates that the direct performance goaleffect was partially mediated through semester GPA (Judd &Kenny, 1981). Figure 5 displays a path model illustrating thelong-term effects of goals, as mediated through the short-termconsequences.

Grades in subsequent psychology courses. For the psychologyGPA analysis, our sample was limited to the group of 148 studentswho actually enrolled in additional psychology classes and re-ceived grades in them. Because of the reduction in power due tothis smaller sample size, we set our criterion for significance to .05for analyses conducted with this subset of 148 students. To exam-ine the direct effects of goals on subsequent psychology grades, weregressed psychology GPA on the basic model. This model wasnot significant, F(5, 142) - 2.10, p < .07, if = .07, but the main


MasteryGoal Interest - .31

CoursesTaken

Instructor — .34

Gender

Performance—.13Goal

Enjoyment

SemesterGPA

-.67-Subsequent

Figure 5. Path model of the mediated effects of mastery and performancegoals on courses taken and subsequent grade-point average (GPA). Pathsrepresent significant (p < .01) effects, and path coefficients are standard-ized regression coefficients. For courses taken, final grade was used tocontrol for performance. For subsequent GPA, semester GPA was used tocontrol for performance.

—.15). There were no effects of work avoidance goals on long-term measures.

Selection bias and restriction of range. Because approxi-mately half the students in the course participated in the study, andbecause they obtained somewhat higher grades than students whodid not participate, restriction of range (in grades or goals) mayhave attenuated the findings reported here. Moreover, these resultsmay be affected by selection bias and may not generalize to allstudents in Introductory Psychology. To address these issues, weanalyzed data collected as part of another project in a differentsemester. We obtained goals measures collected on the first day ofclass from all Introductory Psychology students (N = 1,592) aswell as their final grades in the course. We were therefore able totest the same model in a sample with no selection bias and the fullrange of grades and goals. We tested the same five-term basicmodel on final grades in this sample, and the overall model wassignificant, F(5, 1586) = 15.13, p < .001, K2 = .05. The onlysignificant predictor was performance goals, F(i, 1586) = 23.79,p < .001, £ = .12. These findings therefore replicate thosereported in the primary analyses and are evidence for the gener-alizability of the findings.

effect of performance goals was significant, F(l, 142) = 7.62, p <.01, 3 = .23. No other effects were significant.

When the effects of interest, enjoyment, and final grade wereadded to this direct effects model, the overall model becamesignificant, F(8, 139) = 9.64, p < .001, Z?2 = .36, and accountedfor significantly more variance than the direct effects model, p <.001. The main effect of final grade was significant, F( l ,139) = 61.52, p < .001. Students who received higher grades inIntroductory Psychology received higher grades in their later psy-chology courses (0 = .57). The performance goal effect was nolonger significant and it was reduced in size (/3 = .13, ns, from .23in the direct effects model), suggesting that the direct performancegoal effect was partially mediated through final grade. We foundno evidence of indirect mastery goal effects (j3 for interest = —.11,ns). No other effects were significant. Figure 6 displays a pathmodel illustrating the predictors of grades in subsequent psychol-ogy courses.

Ancillary Analyses

Work avoidance goals. Despite its low reliability, we con-ducted supplemental analyses including work avoidance goals forcomparisons with previous research (Harackiewicz et al., 1997;Meece et al., 1988). We added work avoidance goals to our basicmodel and tested for short-term and long-term effects. The inclu-sion of work avoidance goals in these models did not alter thepattern of significant effects reported in the primary analyses.There were significant main effects of work avoidance goals oninterest, F(l , 597) = 7.23, p < .01, and on enjoyment, F(l,597) = 17.61, p < .001. Students who adopted work avoidantgoals were less interested in psychology (JS = —.11) and enjoyedthe class less (0 = —.17). There were also main effects of workavoidance goals on final grade, F(l, 597) = 14.98, p < .001, andon semester GPA, F(l, 597) = 12.66, p < .001. Students whoadopted work avoidance goals received lower grades in the classthan those not endorsing work avoidance goals (0 = —.17), aswell as lower grades in all of their classes that semester (0 =

Discussion

Our primary objectives in this research were to examine theshort-term consequences of achievement goals adopted in an in-troductory psychology course for interest and performance in thatclass and to track the long-term consequences of these goals forcontinued interest in psychology and subsequent performance inpsychology classes. Goal theorists have argued that achievementgoals play an important role in shaping how students approach,experience, and react to achievement situations, and we found thatstudents' goals, measured early in the semester, were related totheir interest in psychology, enjoyment of lectures, and finalgrades in the course. Specifically, we found that mastery goalswere predictive of interest in psychology and enjoyment of thecourse, but failed to predict academic performance. In contrast, wefound that performance goals were unrelated to interest and en-joyment, but predicted graded performance in two different sam-ples. These short-term results, observed over the course of a

Mastery _Goal

Instructor-

-.35-

.37-

• Interest

• Enjoyment Psychology,GPA

.57

Performance.Goal

-.19-Final

* Grade

Figure 6. Path model of the mediated effects of performance goals onpsychology grade-point average (GPA; n = 148). Paths represent signifi-cant (p < .05) effects, and path coefficients are standardized regressioncoefficients.


college semester, replicate those found by Harackiewicz et al.(1997) and extend our previous findings in several ways. Weobtained these results with a much larger and more representativesample of introductory students, developed differentiated measuresof interest in the course, measured general academic performance(in addition to performance in the psychology class), and includedmeasures of performance expectations and study strategies to testfor potential moderators and mediators of goal effects. Most im-portant, however, we extended the scope of this study beyond thesemester in which achievement goals were adopted.

To track the long-term consequences of achievement goals, weobtained a behavioral measure of continued interest (the number ofadditional psychology credits students chose to take) and assessedacademic performance over the three semesters following theintroductory course. Our measure of long-term interest reflectscontinuing motivation in the field of psychology and is conceptu-ally similar to the behavioral measures of intrinsic motivation usedin laboratory research (e.g., Deci, 1971; Lepper, Greene, & Nis-bett, 1973). For example, Harackiewicz and Elliot (1993) collectedself-report measures of interest in pinball and also recorded howmuch time participants spent playing pinball in a subsequentfree-choice period. To our knowledge, ours is the first study to usesuch a behavioral measure of continuing motivation in an educa-tional context, and we believe that it is an important outcomemeasure to include in educational research (Maehr, 1976).

Over the long term, we found that the goals adopted in anintroductory course continued to predict students' interest in psy-chology and academic performance, although the effects wereweaker over time. Specifically, mastery goals predicted continuedinterest in psychology, and performance goals were associatedwith subsequent grades in psychology courses (for those studentswho actually enrolled in additional psychology courses), as well assubsequent academic performance (for all students in our sample).Although we had hypothesized that mastery goals might havepositive effects on performance in later psychology classes, we didnot find evidence of direct mastery goal effects on performance insubsequent psychology classes, nor did we find mastery goaleffects on the more general measure of overall academic perfor-mance. Thus, the same pattern of goal effects obtained in the shortterm was also observed on measures of continued interest andperformance collected over the following three semesters. In sum,the goals adopted by students in an introductory class may haveimplications for their interest and academic work in subsequentcourses.

Mediated and Indirect Effects on Long-Term Outcomes

In addition to documenting these direct effects of achievementgoals on long-term outcome measures, we also examined whethergoals had mediated or indirect effects on these long-term out-comes. We hypothesized that performance goals should promotesubsequent academic performance by facilitating performance inan introductory course that provides foundational knowledge in aparticular discipline. In addition, we hypothesized that masterygoals might indirectly affect subsequent performance in psychol-ogy courses by fostering interest in psychology. When goals,interest, enjoyment, and final grade in Introductory Psychologywere tested simultaneously as predictors of subsequent psychologygrades, however, final grade emerged as the only significant pre-

dictor, suggesting that final grades were indeed partial mediatorsof the direct effect of performance goals on subsequent psychologygrades. We found no evidence of indirect mastery goal effects,Similarly, when goals, interest, enjoyment, and semester GPAwere tested simultaneously as predictors of overall GPA in sub-sequent courses, semester GPA emerged as the only significantpredictor, indicating that it partially mediated the direct effect ofperformance goals.

We also hypothesized that the long-term effect of mastery goalson continued interest would be mediated through their short-termeffect on interest in the introductory course. In addition, we con-sidered the possibility that performance goals could have indirecteffects on continued interest by promoting higher grades in Intro-ductory Psychology that might, in turn, predict taking additionalpsychology courses. Indeed, we did find a small zero-order corre-lation between final course grades and continued interest, r(604) =.11, p < .01, indicating some degree of support for this position.However, when goals, interest, enjoyment, and final grade weretested simultaneously as predictors of subsequent course choices,interest emerged as the only significant predictor, suggesting thatit partially mediated the direct effect of mastery goals. We foundno evidence of indirect performance goal effects. Thus, masterygoals influenced subsequent course choices through their directeffects on interest. Specifically, mastery goals predicted interest inpsychology, and, in turn, interest predicted the number of addi-tional psychology courses taken. The fact that students' self-reported interest during the introductory course predicted theiractual course enrollment behavior over an ensuing three-semesterperiod is evidence of the crucial role that interest can play incollege education.

In sum, we tested the simultaneous effects of goals and short-term outcome variables as predictors of our long-term outcomevariables and found no evidence for any type of indirect goaleffects. Rather, these analyses suggested that long-term goal ef-fects were mediated through their short-term consequences. Thus,the achievement goals adopted in an introductory course maycontinue to have long-term effects because of their important rolein shaping students' experiences in the introductory course thatprovides the foundation for later work in that discipline.

Interest: Catch Versus Hold Components

Our third objective was to examine the nature of interest in theintroductory course and to distinguish specific reactions to thelectures from more general reactions to the content of the course.Hidi and Baird (1986) noted that situational interest has a dura-tional aspect, with triggering conditions (catch) and conditions thatfacilitate the continuation of interest over time (hold). This dis-tinction is highly relevant to our consideration of continuing in-terest. Mitchell (1993), following Dewey (1913), argued thatcatching interest involves stimulating and engaging students,whereas holding interest involves making course material involv-ing and meaningful. Several of our findings are consistent with acatch-hold interpretation of our measures. First, our two-factoranalytic components of interest seem to correspond to these twoaspects of interest in a college course. The Enjoyment scale re-flected students' reactions to their professor and the lectures,whereas the Interest scale focused on meaningfulness and intrinsicinterest in the content of the course. Second, when we tested the


simultaneous effects of interest and enjoyment as predictors of ourbehavioral measure of continued interest in psychology, we foundthat interest was the only significant positive predictor. This resultis consistent with predictions advanced by Hidi and Baird (1986)and Mitchell (1993) that the hold component of interest is mostrelevant to the duration of interest over time. Finally, interest, butnot enjoyment, was associated with performance in the short term,suggesting that it reflects the component of interest most relevantto learning. In sum, our findings represent the first test of thecatch-hold model in a college setting. Although intercorrelated,our measures represent distinct components of interest, with dif-ferent implications for continued interest (cf. Hidi & Harac-kiewicz, in press; Renninger, 1998b).

Mediators and Moderators

Our fourth objective was to examine potential mediators andmoderators of achievement goal effects in the short term. Wecollected measures of study strategies hypothesized to mediate theeffects of achievement goals on performance and documentedmany of the predicted effects noted in past literature. For example,mastery goals predicted deep, elaborative processing, whereasperformance goals predicted surface level processing. However,these strategy measures did not affect performance in this contextand thus failed to emerge as mediators of the observed goal effects.It is important to note that our assessment was limited to one classof study strategy measures, specifically self-reported cognitivestrategies. Researchers have also investigated other strategy mea-sures that are more metacognitive or more motivational in naturethan those used in this study (e.g., Pintrich & Garcia, 1991;Pressley, 1986; Zimmerman & Pons, 1986). For a more extensivediscussion of mediation, see A. J. Elliot, McGregor, and Gable(1999), whose results correspond to our findings with cognitivestrategy measures and offer a more comprehensive analysis withthe inclusion of other potential mediators.

We also collected measures of performance expectations toexamine potential moderators of goal effects. Specifically, wetested Dweck and Leggett's (1988) prediction that performancegoals have negative effects for students who are lower in perceivedcompetence. Although we found that performance expectations didpositively predict enjoyment, interest, and performance in thecourse, we failed to find any evidence that performance expecta-tions (whether assessed at the outset of the course or mid-semester)moderated the positive performance goal effects observed in thisstudy, replicating some previous findings (A. J. Elliot & Church,1997; Harackiewicz et al., 1997; Kaplan & Midgley, 1997). It isimportant to note, however, that our research has focused oncollege-aged students whose responses to performance goals maydiffer from the elementary and middle-school students who havebeen the focus of many previous achievement goal studies.

Generality Versus Specificity of Achievement of Goals

Achievement goals are typically conceptualized as specific to aparticular course or educational situation, but the fact that ourcourse-specific measures of performance goals also influencedgrades in other courses taken during the same semester suggeststhat this goal orientation may be fairly general (Nicholls, 1989).Moreover, mastery and performance goals influenced continued

interest and performance in related courses taken later in time, andperformance goals influenced academic performance in unrelatedcourses taken later in time. Because performance goals concern adesire to outperform others, they have no direct connection to thecontent of a specific course. Rather, competence is defined andpursued with respect to a student's standing relative to otherstudents, and this orientation may typify a student's orientationtoward performance in all of his or her classes. In contrast, amastery goal orientation may be more discipline specific, becausethis goal orientation is inherently task based. The desire to learnand to develop competence in an introductory course may dependmore directly on the content of a course and the nature of thematerial to be learned. An important direction for future researchwill be the assessment of students' goals in several courses takenin a particular term to explore issues of generality versus coursespecificity of achievement goals.

Optimal Motivation in College Courses

Considered together, our short- and long-term results yieldsuggestive evidence for the beneficial consequences of adoptingperformance goals in college courses. Not only did performancegoals predict grades in an introductory psychology course in twodifferent samples, replicating earlier findings (A. J. Elliot &Church, 1997; Harackiewicz et al., 1997), but they also predictedstudents' GPAs for all courses taken that semester, grades inadditional psychology courses, and GPAs for all courses takenover the three-semester follow-up. Our measure of performancegoals emphasized positive strivings toward competence defined interms of relative ability and outperforming others, and this orien-tation appears to be adaptive for academic performance in acollege setting. These results are consistent with the matchinghypothesis advanced by Harackiewicz and Sansone (1991) thatgoal effects may depend on the general context in which goals arepursued (Harackiewicz & Elliot, 1998). In other words, studentswho are striving to outperform other students may be optimallymotivated in a university context in which excellence is defined interms of an individual's achievement relative to others and inwhich grades are typically assigned on normative curves.

At the same time that we documented the positive correlates ofa performance goal orientation for both short- and long-termconsequences in this context, however, we also documented itslimitations. Performance goals had no effect on students' interestin psychology, enjoyment of the lectures, or the number of addi-tional psychology courses taken. Only mastery goals had signifi-cant positive effects on enjoyment, interest in psychology, andcontinued interest, but, of course, mastery goals failed to predictperformance. Success in college and university contexts dependson both performance and interest, and we have demonstrated theindependent contributions of mastery and performance goals inpromoting these two outcome measures. Moreover, our resultssuggest that the optimal pattern of goal adoption may include bothmastery and performance goals because neither type of goal pre-dicted both outcomes. In sum, our results are consistent with amultiple-goals perspective in which mastery and performancegoals have independent, positive effects on different measures ofsuccess in college education (Barren & Harackiewicz, in press;Harackiewicz et al., 1998).


Although temporal precedence and the logic of goal theory(Pintrich & Schunk, 1996) served as the basis for the impliedcausal direction of our analyses, it is important to note that our dataare correlational and that any causal inferences are tentative. Forexample, even though we measured goals well in advance ofperformance and interest, it is possible that students who performwell in general are more likely to adopt performance goals and thatstudents who find particular academic subjects interesting aremore likely to adopt mastery goals in courses about those subjects.These alternatives are less plausible in the case of an introductorypsychology class, however, for two important reasons. First, thisclass is typically taken during the freshman year before studentshave a history of performance in college classes and certainlybefore they have a history of performance in psychology classes.Indeed, 78% of our participants were first-semester freshmen, and96% were freshmen or sophomores; none had taken any psychol-ogy courses before the introductory course. Second, unlike othercollege courses that cover topics studied in high school (e.g.,mathematics, foreign languages), Introductory Psychology is typ-ically students' first systematic exposure to the field. Althoughthey may enroll in an introductory class on the basis of initialinterest in the topic, students' interests are unlikely to be based inextensive knowledge about the topic or years of prior course work(Renninger, 1992). More research is needed to tease apart thesecausal alternatives, and it will be important to assess both interestand goals at the outset and conclusion of courses.

In sum, we have provided evidence that mastery and perfor-mance goals have positive and complementary consequences formotivation and performance in college courses in both the shortand long term. We therefore advocate a multiple-goals approach inwhich the role of educational context and short-term consequencesis recognized in an analysis of goal dynamics over time. Webelieve that a comprehensive understanding of motivation in col-lege education requires consideration of both academic perfor-mance and continuing interest in a field, and we recognize thepositive potential of both types of achievement goals in predictingthese educational outcomes.

References

Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing andinterpreting interactions. Newbury Park, CA: Sage.

Alexander, P. A. (1997). Mapping the multidimensional nature of domainlearning: The interplay of cognitive, motivational, and strategic forces.In M. L. Maehr & P. R. Pintrich (Eds.), Advances in motivation andachievement (Vol. 10, pp. 213-250). Greenwich, CT: JAI Press.

Alexander, P. A., Kulikowich, J. M , & Jetton, T. L. (1994). The role ofsubject-matter knowledge and interest in the processing of linear andnonlinear texts. Review of Educational Research, 64, 201-252.

Alexander, P. A.. Kulikowich, J. M., & Schulze, S. K. (1994). Theinfluence of topic knowledge, domain knowledge, and interest on thecomprehension of scientific exposition. Learning and Individual Differ-ences, 6, 379-397.

Ames, C. (1992). Classrooms: Goals, structures, and student motivation.Journal of Educational Psychology, 84, 261-271.

Ames, C , & Archer, J. (1988). Achievement goals in the classroom:Students' learning strategies and motivation processes. Journal of Edu-cational Psychology, 80, 260-267.

Archer, J. (1994). Achievement goals as a measure of motivation inuniversity students. Contemporary Educational Psychology, 19, 430-446.

Barron, K. E.F & Harackiewicz, J. M. (in press). Achievement goals andoptimal motivation. In C. Sansone & J. M. Harackiewicz (Eds.), Intrin-sic and extrinsic motivation: The search for optimal motivation andperformance. New York: Academic Press.

Biggs, J. (1993). What do inventories of student learning processes reallymeasure? A theoretical review and clarification. British Journal ofEducational Psychology, 63, 3-19.

Bouffard, T., Boisvert, J., Vezeau, C, & Larouche, C. (1995). The impactof goal orientation on self-regulation and performance among collegestudents. British Journal of Educational Psychology, 65, 317-329.

Brophy, J. E. (1983). Conceptualizing student motivation. EducationalPsychologist, 18. 200-215.

Cohen, J., & Cohen, P. (1975). Applied multiple regression/correlationanalysis for the behavioral sciences. Hillsdale, NJ: Erlbaum.

Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlationanalysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.

Conger, A. J. (1974). A revised definition for suppressor variables: A guideto their identification and interpretation. Educational and PsychologicalMeasurement, 34, 35-46.

Deci, E. L. (1971). Effects of externally mediated rewards on intrinsicmotivation. Journal of Personality and Social Psychology, 18, 105-115.

Deci, E. L. (1992). Interest and the intrinsic motivation of behavior. InK. A. Renninger, S. Hidi, & A. Krapp (Eds.), The role of interest inlearning and development (pp. 43-70). Hillsdale, NJ: Erlbaum.

Dewey, J. (1913). Interest and effort in education. Cambridge, MA: Riv-erside Press.

Duda, J. L., & Nicholls, J. G. (1992). Dimensions of achievement moti-vation in schoolwork and sport. Journal of Educational Psychology, 84,290-299.

Dweck, C. S. (1986). Motivational processes affecting learning. AmericanPsychologist, 41, 1040-1048.

Dweck, C. S., & Leggett, E. L- (1988). A social-cognitive approach tomotivation and personality. Psychological Review, 95, 256-273.

Elliot, A. J. (1997). Integrating the "classic" and "contemporary" ap-proaches to achievement motivation: A hierarchical model of approachand avoidance motivation. In M. L. Maehr & P. R. Pintrich (Eds.),Advances in motivation and achievement (Vol. 10, pp. 243-279). Green-wich, CT: JAI Press.

Elliot, A. J., & Church, M. A. (1997). A hierarchical model of approachand avoidance achievement motivation. Journal of Personality andSocial Psychology, 72, 218-232.

Elliot, A. J., & Harackiewicz, J. M. (1996). Approach and avoidanceachievement goals and intrinsic motivation: A mediational analysis.Journal of Personality and Social Psychology, 70, 461-475.

Elliot, A. J., McGregor, H. A., & Gable, S. L. (1999). Achievement goals,study strategies, and exam performance: A mediational analysis. Journalof Educational Psychology, 91, 549-563.

Elliot, E. S., & Dweck, C. S. (1988). Goals: An approach to motivation andachievement. Journal of Personality and Social Psychology, 54, 5-12.

Harackiewicz, J. M., Barron, K. E., Carter, S. M., Lehto, A. T., & Elliot,A. J. (1997). Predictors and consequences of achievement goals in thecollege classroom: Maintaining interest and making the grade. Journalof Personality and Social Psychology, 73, 1284-1295.

Harackiewicz, J. M., Barron, K. E., & Elliot, A. J. (1998). Rethinkingachievement goals: When are they adaptive for college students andwhy? Educational Psychologist, 33, 1-21.

Harackiewicz, J. M., & Elliot, A. J. (1993). Achievement goals andintrinsic motivation. Journal of Personality and Social Psychology. 65,904-915.

Harackiewicz, J. M., & Elliot, A. J. (1998). The joint effects of target andpurpose goals on intrinsic motivation: A mediational analysis. Person-ality and Social Psychology Bulletin, 24, 675-689.

Harackiewicz, J. M., & Sansone, C. (1991). Goals and intrinsic motivation:You can get there from here. In M. L. Maehr & P. R. Pintrich (Eds.),


Advances in motivation and achievement (Vol. 7, pp. 21-49). Green-wich, CT: JAI Press.

Hidi, S. (1990). Interest and its contribution as a mental resource forlearning. Review of Educational Research, 60, 549-571.

Hidi, S., & Baird, W. (1986). mterestingness—a neglected variable indiscourse processing. Cognitive Science, 10, 179-194.

Hidi, S., & Harackiewicz, J. M. (in press). Motivating the academicallyunmotivated: A critical issue for the 21st century. Review of EducationalResearch.

Hidi, S., & McLaren, J. (1990). The effect of topic and theme interesting-ness on the production of school expositions. In H. Mandl, E. DeCorte,N. Bennett, & H. F. Friedrich (Eds.), Learning and instruction: Euro-pean research in an international context (Vol. 11, pp. 295-308).Oxford, England: Pergamon,

Hidi, S., & McLaren, J. (1991). Motivational factors and writing: The roleof topic interestingness. European Journal of Psychology in Educa-tion, 6, 187-197.

Jackson, P. W. (1968). Life in classrooms. New York: Holt, Rinehart &Winston,

Judd, C. M., & Kenny, D. A. (1981). Process analysis: Estimating medi-ation in treatment evaluations. Evaluation Review, 5, 602-619.

Kaplan, A., & Midgley, C. (1997). The effect of achievement goals: Doeslevel of perceived academic competence make a difference? Contempo-rary Educational Psychology, 22. 415-435.

Krapp, A., Hidi, S., & Renninger, K. A. (1992). Interest, learning, anddevelopment. In K. A. Renninger, S. Hidi, & A. Krapp (Eds.), The roleof interest in learning and development (pp. 3-25). Hillsdale, NJ:Erlbaum.

Lepper, M. R., Greene, D., & Nisbett, R. E. (1973). Undermining chil-dren's intrinsic interest with extrinsic reward: A test of the overjustifi-cation hypothesis. Journal of Personality and Social Psychology, 28,129-137.

Lepper, M. R., & Hodell, M. (1989). Intrinsic motivation in the classroom.In C. Ames & R. Ames (Eds.), Research on motivation in education:Goals and cognitions (Vol. 3, pp. 73-105). New York: Academic Press.

Maehr, M. L. (1976). Continuing motivation: An analysis of a seldomconsidered educational outcome. Review of Educational Research, 46,443-462.

Malone, T. W., & Lepper, M. R. (1987). Making learning fun: A taxonomyof intrinsic motivations for learning. In R. E. Snow & M. J. Farr (Eds.),Aptitude, learning, and instruction: Cognitive and affective processanalyses (Vol. 3, pp. 223-253). Hillsdale, NJ: Erlbaum.

Marton, F., Hounsell, D. J., & Entwistle, N. J. (Eds.). (1984). The experi-ence of learning. Edinburgh, Scotland: Scottish Academic Press.

Meece, J. L. (1991). The classroom context and students' motivationalgoals. In M. L. Maehr & P. R. Pintrich (Eds.), Advances in motivationand achievement (Vol. 7, pp. 261-285). Greenwich, CT: JAI Press.

Meece, J. L., Blumenfeld, P. C , & Hoyle, R. H. (1988). Students' goalorientations and cognitive engagement in classroom activities. Journalof Educational Psychology, 80, 514-523.

Middleton, M. J., & Midgley, C. (1997). Avoiding the demonstration oflack of ability: An underexplored aspect of goal theory. Journal ofEducational Psychology, 89, 710-718.

Midgley, C , Kaplan, A., Middleton, M., Maehr, M. L., Urdan, T., Ander-man, L., Anderman, E., & Roeser, R. (1998). The development andvalidation of scales assessing students' achievement goal orientations.Contemporary Educational Psychology, 23, 113-131.

Midgley, C , & Urdan, T. (1995). Predictors of middle school students' useof self-handicapping strategies. Journal of Early Adolescence, 15, 389-411.

Mitchell, M. (1993). Situational interest: Its multifaceted structure in thesecondary school mathematics classroom. Journal of Educational Psy-chology, 85, 424-436.

Nicholls, J. G. (1979). Quality and equality in intellectual development.American Psychologist, 34, 1071-1084.

Nicholls, J. G. (1989). The competitive ethos and democratic education.Cambridge, MA: Harvard University Press.

Nicholls, J. G., Cheung, P. C , Lauer, J., & Patashnick, M. (1989). Indi-vidual differences in academic motivation: Perceived ability, goals,beliefs, and values. Learning and Individual Differences, 1, 63-84.

Nolen, S. B. (1988). Reasons for studying: Motivation orientations andstudy strategies. Cognition and Instruction, 5, 269-287.

Nolen, S. B., & Haladyna, T. M. (1990). Personal and environmentalinfluences on students' beliefs about effective study strategies. Contem-porary Educational Psychology, 15, 116-130.

Pintrich, P. R., & DeGroot, E. V. (1990). Motivational and self-regulatedlearning components of classroom academic performance. Journal ofEducational Psychology, 82, 33-40.

Pintrich, P. R., & Garcia, T. (1991). Student goal orientation and self-regulation in the college classroom. In M. L. Maehr & P. R. Pintrich(Eds.), Advances in motivation and achievement (Vol. 7, pp. 371-402).Greenwich, CT: JAI Press.

Pintrich, P. R., & Schrauben, B. (1992). Students' motivational beliefs andtheir cognitive engagement in classroom tasks. En D. H. Schunk & J.Meece (Eds.), Student perceptions in the classroom: Causes and con-sequences (pp. 149-183). Hillsdale, NJ: Erlbaum.

Pintrich, P. R., & Schunk, D. H. (1996). Motivation in education: Theory,research, and applications. Englewood Cliffs, NJ: Prentice-Hall.

Pintrich, P. R., Smith, D. A. F., Garcia, T., & McKeachie, W. J. (1993).Reliability and predictive validity of the Motivated Strategies for Learn-ing Questionnaire (MSLQ). Educational and Psychological Measure-ment, 53, 801-813.

Pressley, M. (1986). The relevance of a good strategy user model to theteaching of mathematics. Educational Psychologist, 21, 139-161.

Rathunde, K. (1993). The experience of interest: A theoretical and empir-ical look at its role in adolescent talent development. En M. L. Maehr &P. R. Pintrich (Eds.), Advances in motivation and achievement (Vol. 8,pp. 59-98). Greenwich, CT: JAI Press.

Renninger, K. A. (1992). Individual interest and development: Implicationsfor theory and practice. In K. A. Renninger, S. Hidi, & A. Krapp (Eds.),The role of interest in learning and development (pp. 361-395). Hills-dale, NJ: Erlbaum.

Renninger, K. A. (1998a). The roles of individual interests) and gender inlearning: An overview of research on preschool and elementary school-aged children/students. In L. Hoftman, A. Krapp, K. A. Renninger, & J.Baumert (Eds.), Interest and learning: Proceedings of the Seeon Con-ference on Interest and Gender (pp. 165-174). Kiel, Germany: IPN.

Renninger, K. A. (1998b). What are the roles of individual interest, taskdifficulty, and gender in student comprehension? In L. Hoffmann, A.Krapp, K. A. Renninger, & J. Baumert (Eds.), Interest and learning:Proceedings of the Seeon Conference on Interest and Gender (pp.228-238). Kiel, Germany: IPN.

Sansone, C, & Harackiewicz, J. M. (1996). "I don't feel like it": Thefunction of interest in self-regulation. In L. L. Martin & A. Tesser (Eds.),Striving and feeling: Interactions among goals, affect, and self-regulation (pp. 203-228). Hillsdale, NJ: Erlbaum.

Schiefele, U. (1991). Interest, learning, and motivation. Educational Psy-chologist, 26, 299-323.

Schiefele, U., & Krapp, A. (1996). Topic interest and free recall ofexpository text. Learning and Individual Differences, 8, 141-160.

Schmeck, R. R. (1983). Learning styles of college students. In R. F. Dillon& R. R. Schmeck (Eds.), Individual differences in cognition (pp. 233-279). New York: Academic Press.

Skaalvik, E. M. (1997). Self-enhancing and self-defeating ego orientation:Relations with task and avoidance orientation, achievement, self-perceptions, and anxiety. Journal of Educational Psychology, 89, 7 1 -81.

330 HARACKIEWICZ. BARRON, TAUER, CARTER, AND ELLIOT

Urdan, T. (1997). Achievement goal theory: Past results, future directions. interview for assessing student use of self-regulated learning strategies.In M. L. Maehr & P. R. Pintrich (Eds.), Advances in motivation and American Educational Research Journal, 23, 614-628.achievement (Vol. 10, pp. 99-141). Greenwich, CT: JAI Press.

Wolteis, C. A., Yu, S. L., & Pintrich, P. R. (1996). The relation betweengoal orientation and students' motivational beliefs and self-regulated Received November 22, 1998learning. Learning and Individual Differences, 8, 211-238. Revision received August 31, 1999

Zimmerman, B. J., & Pons, M. M. (1986). Development of structured Accepted August 31, 1999 •

Documents

Short-Term and Long-Term Consequences of Achievement …siegler/418-Guthrie.pdf · 318 HARACKIEWICZ, BARRON, TAUER, CARTER, AND ELLIOT effects will be mediated through their short-term