tracking, students' effort, and academic achievement

Tracking, Students' Effort, and Academic AchievementAuthor(s): William CarbonaroSource: Sociology of Education, Vol. 78, No. 1 (Jan., 2005), pp. 27-49Published by: American Sociological AssociationStable URL: http://www.jstor.org/stable/4148909 .

Accessed: 21/12/2014 02:26

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

American Sociological Association is collaborating with JSTOR to digitize, preserve and extend access toSociology of Education.

http://www.jstor.org

This content downloaded from 128.235.251.160 on Sun, 21 Dec 2014 02:26:25 AMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=asa

http://www.jstor.org/stable/4148909?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


Tracking, Students' Effort, and Academic Achievement

William Carbonaro University of Notre Dame

This study examined the links among students' effort, tracking, and students' achievement. It found that students in higher tracks exert substantially more effort than do students in lower tracks. These differences in effort are largely explained by differences in prior effort and

achievement, as well as students' experiences in their classes. Students' effort is strongly related to students' learning, and track differences in students' effort account for a modest portion of track differences in students' learning. Finally, the effect of students' effort on students'

learning is roughly the same, regardless of the track in which a student is placed.

ociologists of education have focused heavily on how structural factors-the systemic organizational and institutional

characteristics of schools-shape academic outcomes. Curricular tracking is perhaps the most prominent structural aspect of schools that researchers have studied. Research has strongly suggested that students in higher "tracks" and ability groups tend to learn more than do comparable students in lower tracks and ability groups (Barr and Dreeben 1983; Gamoran 1986, 1987; Gamoran and Mare 1989; Hoffer 1992).1 Much research has focused on differences in learning opportunities across ability groups and tracks as a possible explanation (see, e.g., Gamoran 1986; Pallas et al. 1994). Typically, higher- ability groups and higher-track classes are characterized by higher-quality instruction (Gamoran and Carbonaro 2002-03; Oakes 1985; Page 1991), more time spent on instruction (Barr and Dreeben 1983; Oakes 1985), and greater curricular coverage (Barr and Dreeben 1983; Brophy and Good 1986; Rowan and Miracle 1983). Since each of these aspects of students' learning opportunities is related to students' learning (Wang 1998), it is not surprising that students in higher-ability groups and tracks enjoy greater

gains in learning than do those in lower-ability groups and tracks.

Although learning opportunities are certainly important in determining how much students learn in school, other factors that influence learning have received less attention. Sociologists of education have focused heavily on the importance of social structure but have been less attentive to the importance of human agency in shaping students' outcomes. For example, curricular tracking is a social structure that differentially provides opportunities and imposes constraints upon what students have the potential to learn. A massive array of studies have described curricular tracking as a practice and have examined its effects. In contrast, a much smaller number of studies have focused on human agency-whether a student chooses to engage himself or herself in the learning process-even though it also plays a critical role in explaining why some students learn more than do others.2 Serensen and Hallinan (1977) argued that differences in achievement among students can be explained by three factors: learning opportunities, effort, and ability. By focusing on learning opportunities and effort, they highlighted the importance of both social structure and human agency in explaining differences in learning.

Sociology of Education 2005, Vol. 78 (January): 27-49 27



28 Carbonaro

Despite Sorensen and Hallinan's important conceptual contribution, few studies have successfully accounted for the role that both structure and agency play in determining students' outcomes. The main goal of this study is to examine how curricular tracking and effort are related to gain a better understanding of how structure and agency shape students' outcomes. In so doing, I hope to render a more complete understanding of how unequal outcomes emerge from structural factors within the school that are imposed on students and choices that students make on being placed in such structures.

BACKGROUND

What Is Students' Effort?

Before I discuss how effort is related to outcomes and curricular tracking, it is necessary to devise a working definition of effort. Effort is the amount of time and energy that students expend in meeting the formal academic requirements established by their teacher and/or school. It is goal specific, and different students may exert the same level of effort in meeting some goals but different levels of effort in meeting others. Often these goals are hierarchical, and some require little more than tacit compliance, while others demand greater commitments of time and/or thought.

It is possible to distinguish among three different types of effort: rule oriented, procedural, and intellectual. Rule-oriented effort entails students' compliance with the most basic rules and norms required by the school, such as showing up for class regularly and refraining from misbehavior. Two students who attend class regularly are exerting the same level of rule-oriented effort. Procedural effort requires students to try to meet the specific demands set forth by a teacher in a particular class, including completing required assignments, turning in assignments on time, and participating in class discussions. Procedural effort places higher demands on students than does rule-oriented effort: Two students who attend class regularly are equal in terms of rule-oriented effort, but one stu-

dent may exert more procedural effort by turning in homework assignments more con- sistently than the other student. Finally, students exert intellectual effort when they apply their cognitive facilities toward understanding the intellectual challenges posed by the curriculum. Two students exert the same amount of procedural effort if they both submit the same number of homework assignments, but if Student A devotes more time and thought to answering all the questions in the assignment correctly while Student B is simply con- cerned with handing in the assignment (regardless of quality), Student A expends more intellectual effort than does Student B. As these examples illustrate, effort is a multi- dimensional concept, and a good indicator of effort should include measures of a broad range of tasks and expectations.

This definition of effort can be contrasted with three concepts that are often associated with it. First, Willis (1977) popularized the concept of resistance in his study of working- class youths in England. Resistance clearly connotes students' withdrawal of effort. However, this concept is limited because it fails to differentiate among the different levels of effort exerted by students who have not rejected the school culture.

Second, psychologists focus heavily on motivation and self-efficacy (see, e.g., Bong and Clark 1999). Motivation and self-efficacy are clearly related to effort because they explain why some students exert more effort than do others. However, neither is equiva- lent to effort because two students may exert the same level of effort and have different motives and/or levels of self-efficacy.

Finally, it is important to emphasize how the concept of effort differs from that of engagement. Engagement has been defined and operationalized in numerous ways by different researchers. Typically, researchers have argued that effort, as represented by behaviors like attending class and time spent on homework, is a key component of engagement (see Johnson, Crosnoe, and Elder 2001; Smerdon 1999). However, some researchers have also argued that engagement includes an affective or psychological component that focuses on students' enthusiasm about, interest in, and attachment to their school and/or



Tracking, Effort, and Achievement 29

the schooling process (see, e.g., Newmann 1992). I argue that effort can be studied apart from this affective component and that it is beneficial to do so because effort can affect outcomes, regardless of whether students are enthusiastic about, interested in, and/or attached to their school.

Students' Effort and Learning Numerous studies have found that students' effort is related to academic achievement. Studies of "engagement" have typically relied heavily on measures of effort, such as the completion of homework, attentiveness, and preparedness. Generally, the findings of such studies have indicated that students who are more engaged learn more in school (Johnson et al. 2001; Marks 2000; Smerdon 1999; Yair 2000). Farkas et al. (1990) found that students' "work habits," as measured by teachers' reports of homework, class participation, effort, and organization, were positively related to students' mastery of courses and grade point averages (GPAs). Rosenbaum (2001) also found that students' "preparedness" and absenteeism was related to their GPAs. Time spent on homework by students is a com- monly used measure of effort (see, e.g., Natriello and McDill 1986), and numerous studies have indicated that it is related to higher achievement (Alexander and Cook 1982; Carbonaro and Gamoran 2002; Natriello and McDill 1986; but see Bryk, Lee, and Holland 1993 for an exception). In short, although the labels and measures used have varied across studies, effort has been found to be positively related to achievement.

Although research on students' effort has yielded interesting insights, researchers have paid little attention to the possible connec- tion between effort and curricular tracking. This omission is surprising, given the sheer volume of research on curricular tracking. Most studies of tracking have focused on the importance of learning opportunities in explaining differences in learning across ability groups and tracks, and only a few have examined the links between either ability grouping or curricular tracking and effort (Eder 1981; Felmlee and Eder 1983; Natriello and McDill 1986). Hence, an examination of

the relationships between effort and curricular tracking will render a more complete account of how curricular tracking affects academic outcomes.

RESEARCH AGENDA

The analyses presented here focus on four main questions regarding effort and curricular tracking: (1) Does effort vary across tracks? (2) What explains variation in effort across tracks? (3) Does variation in effort explain differences in learning across tracks? and (4) Does effort have the same effect on learning across tracks, or does effort matter more for learning in some tracks than in others? Each research question links effort, curricular differentiation, and learning in ways that will enhance researchers' understanding of inequality in students' outcomes.

First, does effort vary across curricular tracks? Although numerous studies have doc- umented differences in learning opportunities across tracks (Gamoran and Carbonaro 2002-03; Oakes 1985, 1990; Page 1991), only a handful of studies have examined whether effort varies across tracks. Eder (1981) and Felmlee and Eder (1983) found that first graders in lower-ability groups were less attentive than their peers in higher-ability groups. If "attentiveness" is considered a measure of effort, this research suggested that ability grouping is related to effort. Natriello and McDill (1986) found that students in the college track spent more time on their homework than did students who were not in the college track. While Natriello and McDill used "time spent on homework" as a proxy for effort, this measure is flawed because other research on tracking has indicated that teachers assign more homework to students who are enrolled in higher-track classes (Oakes 1985). Thus, it is not clear whether higher-track students are actually exerting more effort or whether they are simply responding to the greater demands that their teachers place on them. Finally, Smerdon (1999) found that engagement, as measured by attendance, preparedness, and time spent on homework, was positively related to track placement. Smerdon's measures



30 Carbonaro

of track were reported by students and were not subject specific. Some scholars have raised questions regarding the appropriate- ness and quality of such measures (e.g., Lucas 1999; Lucas and Gamoran 2002). Clearly, additional research, using nationally repre- sentative data with better measures of effort and track, is needed to determine the extent to which students' track placements are related to the amount of effort they exert in their classes.

The next research question focuses on why effort varies across curricular tracks. Prior research has suggested numerous possible explanations. These explanations, along with the hypothesized relationships between effort and learning, are displayed in Figure 1.

First, students' background characteristics may explain track differences in effort. Numerous scholars have argued that "oppositional cultures" emerge from and are sus- tained by cultural differences across racial/ethnic (Farkas, Lleras, and Maczuga 2002; Fordham and Ogbu 1986; Mickelson 1990; Ogbu 1978, 2003; Suarez-Orozco 1987) and class boundaries (Cook and Ludwig 1998; MacLeod 1995; Weis 1990; Willis 1977). The antischool norms fostered by these subcultures disengage students from

the learning process, sap their desire to strive for academic success, and ultimately under- mine their levels of academic achievement. The conclusions from research on black-white differences in effort have been mixed. Qualitative studies have suggested that antischool norms and low effort by students are an important source of underachievement by minority students (Fordham and Ogbu 1986; Ogbu 1978, 2003; Suarez-Orozco 1987). Quantitative research that has analyzed teachers' reports of students' effort (Ainsworth-Darnell and Downey 1998; Tach and Farkas 2003) has found that whites exert more effort than do blacks.3 However, studies that have examined students' reports of effort have found that black students and white students generally do not differ in the effort they exert in school (Cook and Ludwig 1998; Ferguson 2001; Marks 2000; Smerdon 1999).

The findings on class and gender differences in effort have been more consistent: Students of higher socioeconomic status (SES) and females are more likely to exert more effort than are lower-SES and male students (Cook and Ludwig 1998; Marks 2000; Smerdon 1999). SES, race/ethnicity, and gender are important factors that shape how students are sorted into different tracks

Beliefs about Self and

EffortFuue t

Future

Track SStudent

Placement - - - - - - > Effort Student

Achievement

Prior

Achieve-

ment Intellectual Engagement

OTL,

Background Factors

Figure 1. Conceptual Model for Understanding the Relationship Among Effort, Tracking, and Achievement




(Gamoran and Mare 1989; Jones, Vanfossen, and Ensminger 1995). The overrepresenta- tion of racial/ethnic minority, lower-SES, and male students in lower-track classes may partly explain track differences in students' effort.

Second, it is possible that prior effort is an important criterion that is used to sort students into curricular tracks. Students who exert more effort in the 8th grade may be more likely to be placed in higher tracks in the 10th grade and, consequently, may be predisposed to exert greater effort in the 10th grade. Alternatively, it is well established that academic achievement is one of the strongest predictors of track placement (Gamoran and Mare 1989; Hallinan 1992; Jones et al. 1995; Lucas 1999). If effort is related to academic achievement and students are largely sorted into tracks on the basis of their prior achievement, it is possible that differences in effort across tracks may simply reflect this sorting process of high-achieving and effort-exerting students into higher-track classes.

Third, it is possible that higher-track students have beliefs about themselves and their future that lead them to exert more effort in their classes. Higher-track students may feel more efficacious because of teachers' different expectations across tracks (see, e.g., Oakes 1985). If students internalize such expectations, students in different tracks may hold different beliefs about their own chances of academic success. Students who believe they can succeed and expect to succeed in school will try harder because they anticipate that there will be a distinct "payoff" to their efforts. In contrast, students who do not believe that academic success is likely or even possible are unlikely to try hard in school because they view such efforts as a waste of time. While I argue that beliefs evoke effort, it should be noted that effort may alter students' beliefs in response to teachers' praise and/or higher grades.

Finally, track placement may be related to the cognitive demands of and stimulation from the curriculum, which may ultimately shape how much effort students exert in a given class. Oakes (1985) found that high- track students had greater opportunities for critical thinking and were typically given more challenging material to study. Students

will exert more effort when there is a greater academic challenge and when they find meaning in the daily tasks that are required of them. Marks (2000) found that "authentic instructional work," a loosely related concept that focuses on whether instruction taps into students' interests and requires that they "dig deeply" in their studies, was positively related to engagement. Yair (2000) also found that students displayed greater engagement when they are exposed to instruction that is characterized by greater relevance, challenge, and academic demand. If higher-track classes provide material and require tasks that are more intellectually stimulating, it is possible that students in higher-track classes may respond with greater effort.

It is important to note that dashed arrows point from track placement to effort and achievement in Figure 1. These dashed arrows are meant to denote that these relationships are estimated in the analyses, but the expectation is that these paths should be insignificant if all the factors in Figure 1 are fully accounted for in the model. In other words, the effects of track placement on effort and achievement should be indirect, working through the four main factors described earlier. Any residual track differences in effort and achievement after these factors are controlled for may be due to poor measurement of the mediating variables or some other variable that is unaccounted for in the conceptual model.

The next two questions focus on how effort and tracking are related to achievement outcomes for students. Smerdon (1999) found that engagement and track placement had independent effects on reading and math achievement, but she did not examine whether any of the track-achievement relationship was explained by engagement. The analyses presented here examine this issue and reveal whether differences in effort across tracks partly explain why students in higher- track classes tend to learn more than do those in lower-track classes.

Second, I examine whether the relationship between effort and achievement varies across curricular tracks. Since higher-track classes typically have higher-quality instruction and more-experienced teachers (Oakes



32 Carbonaro

1985, 1990), students in these classes may have to expend less effort to learn the material than may students in lower-track classes who are exposed to low-quality instruction by less-experienced teachers. Lectures that lack clarity and excitement are likely to require students to ask more questions in class, seek help outside class, and do more work at home to learn the material. In short, students in lower-track classes may need to exert more effort to do as well as students in higher-track classes.4 Thus, effort may be a critical predictor of students' success in lower-track classes but may be less important for learning in higher-track classes. Examining these questions will illuminate how the interplay between structure (learning opportunities provided by different track classes) and agency (students' efforts to learn) shapes academic outcomes for students.

DATA

The analyses reported here used data from the 8th- to 10th-grade cohort of the National Education Longitudinal Survey of 1988 (NELS:88). This data set is well suited for studying the relationship between tracking and effort for several reasons. First, the longitudinal design of the data makes it possible to control for 8th-grade differences in students' attitudes, behaviors, and achievements, thereby reducing the likelihood that differences in effort and learning simply reflect pre- existing differences prior to the 10th-grade track placement. Second, the NELS:88 data provide a vast array of student- and teacher- reported indicators that serve as important controls in the models.

Finally, and most important, the NELS:88 data provide two teacher reports per student in the 8th and 10th grades. In both the 8th and 10th grades, two teachers were surveyed per student; one teacher taught either math or science, and the other taught either English or history. Subject-specific reports by teachers have an important advantage over data that are not subject specific: Both students' effort and students' track placement are allowed to vary across classes. To capital- ize on this important feature of the NELS

data, samples of students in four subjects- math, English, history, and science-were created. This design allows students to appear in multiple samples, thereby allowing their track status and effort to vary by subject. Consequently, the design of this study is superior to the designs of other studies of tracking and effort in which measures of track status and effort/engagement were reported without reference to specific academic subjects (e.g., Smerdon 1999). The results reported in this article focus on math, since most of the research in the area of high school tracking has focused on math achievement (e.g., Gamoran and Mare 1989; Hoffer 1992). However, the analyses were per- formed in all four subjects, and the overall findings were consistent across academic subjects (results available on request).

Students' Effort Variables

The main variables of interest measure students' effort. It is important to recognize that measuring effort is problematic in several respects. First, it may be difficult for students to report accurately the effort they exert in school for several reasons. One problem is that "exerting high effort" may mean something different for different students. Some students may think they are exerting high levels of effort if they do everything the teacher asks; others may consider such effort only adequate. Such different definitions of effort decrease the reliability of self-reported effort. Social desirability bias is another problem with self-reports of effort. Brint, Contreras, and Matthews (2001) found that elementary schools transmitted many messages to students about the value of hard work as part of the hidden curriculum. If students internalize such messages, it may be difficult for them to admit that they are not working hard. Alternatively, some evidence has suggested that high school students may downplay their effort in explaining either their academic success or failure (Bishop 1999).

Teachers serve as an alternative source of information about students' effort. Just as with students, there are advantages and dis- advantages to using teachers' reports of effort. On the one hand, teachers are limited




in that they must indirectly assess effort. For example, a student may try hard and spend a great deal of time on a homework assignment, but still do a poor job because he or she does not have the skills or resources needed to succeed. A teacher may incorrectly attribute the student's poor performance on the assignment to insufficient effort and thus provide an unreliable estimate of effort. On the other hand, a teacher is not susceptible to social desirability bias in reporting students' effort and hence may provide less-biased estimates of effort.

Despite their potential limitations, the teachers' reports of students' effort in the NELS:88 data are preferable to the students' reports for two reasons. First, the range of items on effort reported by the teachers is more extensive than that reported by the students. In addition, the teacher measures of students' effort relied on both subjective assessments of students' effort and students' behaviors that are more tangible, easily observed, and reported. Again, when different types of information about effort are used, the limitations of any given item will be less important. Two separate measures of 8th- and 10th-grade effort were created for the analyses that follow.

First, the measure of 8th-grade (or "prior") effort was created from teacher-reported items of effort that were collected in the 8th grade. By controlling for 8th-grade effort, it is possible to determine whether differences in 10th-grade effort across tracks reflect the placement of students who expended low effort into lower-track classes, rather than a response by students to their placement in a given track. The seven items used in the "prior effort" scale (displayed in Appendix Table A) cover both subjective assessments of effort (e.g., "this student performs below his or her ability") and reports of concrete, observable student behaviors (e.g., "The student is frequently absent"). The various items in the scale tap the different types of effort mentioned previously: rule oriented (student is absent, tardy, and disruptive), procedural (student rarely completes homework), and intellectual (student performs below ability, is withdrawn, and is inattentive).

Each student had teachers in two subjects

(math or science and history or English) who reported these items on the basis of their particular perceptions of effort in a given subject. It would have been preferable to match reports of 8th- and 10th-grade effort by academic subject (i.e., 8th- and 10th-grade effort as reported by a student's 8th- and 10th- grade math teachers). Unfortunately, it was not possible to do so for many students because their math teachers in the 8th and 10th grades were not sampled; instead, these students had a science teacher sampled in the 8th grade and a math teacher sampled in the 10th grade. To deal with this problem, I included the reports from both 8th-grade teachers for each student in a single scale, which made it possible to get an average level of effort exerted across the two classes, which serves as a proxy for effort in a given subject. The reliability for this scale was high (Cronbach's alpha = .85).

A 10th-grade measure of effort was created using three items from the 10th-grade reports from teachers (see Appendix Table A). While the scale includes an item that is a subjective report of effort, it also uses two items that are based on students' behaviors: attentiveness and turning in homework.5 Turning in homework is an indicator of procedural effort, and attentiveness is a measure of intellectual effort. Ideally, separate measures for each of the three types of effort-rule oriented, procedural, and intellectual-could be included in the analyses, but the three available measures in NELS are best suited to be combined in a scale, to maximize the reliability and validity of the effort measure.

These measures are subject specific for each student, and most students had separate reports from two of their teachers. The major advantage of using subject-specific samples is that effort is allowed to vary across students' classes. Students' effort may vary across their classes because of differences in either intrinsic or extrinsic motivation (e.g., students preferring or valuing one academic subject over another). In addition, if (as I argue later) students' track placements vary across subjects and effort is expected to vary across tracks, then a subject-specific measure of effort is crucial for the analyses.

In her study, Smerdon (1999) used stu-



34 Carbonaro

dent-reported engagement measures that did not vary across subjects, arguing that these general reports of engagement are preferable because they provide a more complete description of students' engagement over the duration of a school day. However, since students' track placements, academic experiences, and effort likely vary across classes, the class-specific reports of effort used in this study should more accurately reflect how effort is related to the instructional and curricular context of a given class.6

Overall, the 10th-grade effort scales were reliable, with alphas ranging from .83 to .86 across academic subjects. It should be noted that these reliabilities are much higher than the .61 alpha reliability reported by Smerdon (1999). Consequently, the analyses presented here are an improvement on Smerdon's because they are less susceptible to possible problems associated with random measurement error in the regression analyses.

It is important to note that the 8th- and 10th-grade measures of effort are not identical. First, the wording of the questions and response categories are not identical in the 8th and 10th grades. The differences in wording are similar enough that they are only of minor concern. The 10th-grade items have the advantage of five response categories, as opposed to two for the 8th-grade measure, thereby making it likely that the 10th-grade estimates are somewhat more precise than the 8th-grade estimates. Second, the 8th- grade scale has four additional items that are not included in the 10th-grade measure (see Appendix Table A). Adding these additional items to the 8th-grade effort scale enhanced the reliability of the scale (.81 versus .85), and the validity was probably improved as well. The "reduced" 8th-grade measure without the additional items is highly correlated (r = .925) with the measure used in the analyses. Given this high correlation, the regression results were virtually identical, regardless of which measure of 8th-grade effort was used.

Track Indicators There are three sources of information about a student's track in NELS:88: students' self- reports, teachers' reports, and information

derived from transcript data. Lucas (1999) discussed the strengths and limitations of each method of operationalizing students' track placements and concluded that researchers may legitimately use any of the three measures, depending on the particular goals of a given analysis. I decided to use the teachers' reports of track for four main reasons: (1) Teachers' reports are subject specific and therefore allow track placement to vary across subjects, (2) teachers are at least as likely to identify important distinctions between classes correctly as are researchers in examining the transcript data, (3) it is unclear whether viable course-based indicators can be constructed in subjects other than math from the NELS:88 transcript data, and (4) the results do not differ when the teachers' reports of track are replaced with track indicators that were derived from the transcript data.7

NELS:88 provides two subject-specific teachers' reports of track for each student. Teachers were asked, "Which of the following best describes the 'track' this class is considered to be?" and were given five response categories: honors or advanced, academic, general, vocational-technical-business, and other. Dummy variables to represent the honors/advanced, academic, and vocational- technical-business/other categories were created for the analyses; the general-track class served as the reference category. Hence, the coefficients for the track dummy variables indicate the average difference in effort and learning between students in a given track and students in general-track classes.

Additional Variables In addition to the measures of effort and track, numerous other variables were used in and created for the analyses. Descriptions of all variables used in the analyses are reported in Table 1. While the measure of 10th-grade effort in math serves as the dependent variable in the first set of analyses, 10th-grade math achievement serves as the dependent variable in the second set. IRT (item-response theory) scores were used because they are easily interpretable (one point equates with one item correct on the examination) and are




Table 1. Means and Standard Deviations for the Variables in the Analyses (n = 6,911)a

Variable Mean SD Low High

Background Female .50 0 1 Race/ethnicity

White, non-Hispanic (reference) .75 - 0 1 Black .09 - 0 1 Hispanic .10 - 0 1 Asian .06 - 0 1

SES -.013 0.775 -2.97 2.56

8th-Grade Achievement Math 37.31 12.09 16.18 66.81 Reading 27.60 8.67 10.61 43.83 Science 19.21 4.87 9.46 32.88 History 29.96 4.54 19.23 41.30

10th-Grade Achievement Math 37.31 12.08 16.37 72.76

Track (Math) Honors/advanced .13 - 0 1 Academic .53 - 0 1 General (reference) .29 - 0 1 Vocational/other .05 - 0 1

Students' Effort Effort scale (G8) 12.45 2.69 0 14 Effort scale (G10) (math) 6.49 2.07 0 9 Effort scale (G10) (English)b 6.47 2.07 0 9 Effort scale (G10) (history)b 6.61 2.05 0 9 Effort scale (G10) (science)b 6.48 2.05 0 9

Students' Self-Efficacy Locus of control (G8) 0.049 0.595 -2.89 1.52 Self-concept (G8) 0.028 0.649 -3.61 1.89 Educational expectations (G8) 4.62 1.27 1 6 Locus of control (G10) 0.0340 .619 -2.66 1.44 Self-concept (G10) -0.0060 .681 -2.95 1.66 Educational expectations (G10) 6.40 2.06 1 9

Students' Intellectual Stimulation Student feels challenged 12.62 7.74 0 20 Asked to show understanding 10.43 7.93 0 20

a The 8-10 panel weight (F1 PNLWVVT) was used to calculate the means and standard deviations in this table.

b The mean and standard deviation for this variable were based on a subject-specific sample.



36 Carbonaro

scaled so that 8th- and 10th-grade test scores are in the same metric.

Four categories of variables were used to explain differences in effort and achievement across tracks. First, since background factors, such as race/ethnicity, sex, and social class, may be related to effort, track placement, and achievement, controls for these variables were included in the analyses. Race/ethnicity was based on students' 8th-grade reports and was represented by a series of dummy variables (black, Hispanic, and Asian), which were contrasted with non-Hispanic whites (the reference category). SES was a composite variable that was composed of five different factors taken from the base year: family income, mother's and father's education, and mother's and father's occupations. A dummy variable ("female") was created to denote each students' sex.

Second, measures of prior achievement consisted of 8th-grade IRT test scores in reading, math, science, and history. Following Jencks's (1985) suggestion, all four test scores were included separately as independent variables in the regression analyses to control for prior achievement. The use of four test scores minimized possible reliability problems in the analysis.8

Third, multiple measures of students' beliefs about themselves and their future were used in the analysis: students' (1) 8th- and 10th-grade locus of control,9 (2) 8th- and 10th-grade self-concept,10 and (3) 8th- and 10th-grade expectations regarding how far students think they will go in school.11 By including the 8th- and 10th-grade measures as separate predictors in the regression models, I was able to establish whether a change in students' locus of control and/or self-concept affected effort. Finally, two separate measures of intellectual stimulation were included: whether students felt challenged in a given subject and whether they were asked to show understanding in a given subject. Although items that measure whether a student was interested in or stimulated by a class may be preferable, such measures are unavailable in NELS. Regardless, the measures used here are likely to be strongly correlated with these more-direct measures.

It is important to note that although

opportunities to learn (OTL) are present in the conceptual model (see Figure 1), direct measures of learning opportunities were not included in the analysis. Although it would be preferable to include such measures in the models, valid OTL measures are unavailable in NELS.12 Consequently, after adjustments for the other mediating variables, the track dummy variables probably represent differences in exposure to learning opportunities, as well as other unmeasured factors.

METHODS

Ordinary least-squares (OLS) regression tech- niques were used to analyze the data. The dis- tribution for the effort scale is clearly negatively skewed. Consequently, when effort is used as a dependent variable, the estimates of the coefficients will be less efficient than they would be if effort were normally distributed. The models presented here were rerun with a normalized version of the effort scale as the dependent variable. Comparisons of the two sets of analyses indicated that the magnitude and levels of statistical significance for the coefficients were nearly identical. The results for the analyses using the untransformed effort scale are presented in the tables, since the interpretation of coefficients is more straightforward in these analyses.

As with any quantitative study, missing data also presented problems for the analyses. Although there was no specific variable that had a high degree of item nonresponse (percentages of missing cases ranged from 0 to 20 percent), when included together in the "full" regression models, 35 percent to 42 percent of the cases were lost with listwise deletion. If the data are "missing completely at random," listwise deletion provides unbiased estimates, but the smaller sample size decreases the statistical power of the models (Alison 2002).

To avoid this limitation of listwise deletion, multiple imputation (using AMELIA software) was used to deal with missing data that were due to item nonresponse. Multiple imputation provides larger sample sizes than does listwise deletion and requires only the weaker "missing at random" assumption to produce




unbiased estimates. Like all imputation proce- dures, King et al.'s (2001) multiple imputation technique uses information from other variables in the data set to generate predicted values for cases with missing data. In this case, all the variables listed in Table 1 were used to impute missing data in each academic subject. However, multiple imputation is unique in that it generates multiple data sets-in this case, five-with different imputed values in each data set. By imputing missing values five different times, it is possible to account for the uncertainty inherent in the imputation process. The same models are then run on each data set, and the final results are then averaged across the five analyses. Since the imputed data sets have no missing data, the sample sizes remain the same in each of the regression models (unlike listwise deletion). The sample sizes in English, math, science, and history were 8,518, 6,911, 5,896, and 4,351, respectively.13

Since NELS is a multistage cluster sample, the true standard errors are actually larger than the standard errors reported by most statistical software packages. The "survey" command in Stata was used to calculate the correct standard errors and account for the design effects in NELS. By using information about the sampling strata and primary sampling unit, Stata is able to generate weighted point estimates that are then used to create a

first-order matrix Taylor series expansion that generates the correct standard errors (Statacorp 1999).

RESULTS

Does Effort Vary Across Tracks? Do students in different track classes exert different levels of effort? Table 2 displays the means for 10th-grade effort by track. The higher the track of the class, the more effort students exerted. The differences in effort across tracks are sizable and statistically significant in all four subjects. For example, the difference in effort between students in the honors and academic tracks is roughly a third of a standard deviation. Even more striking, the difference in effort between students in the honors and general track is between 60 percent and 85 percent of a standard deviation. The teachers reported that the students in vocational classes exerted the least effort in all four subjects.

Explaining Track Differences in Effort

The results presented in Table 2 suggest that students in different tracks exert different levels of effort. What explains these differences

Table 2. Mean Differences in 10th-Grade Effort, by Track, for English, Math, Science, and History (standard deviations in parentheses)

Track English Math Science History

Honors 9.378b,c,d 9.400b,c,d 9.275b,c,d 9.358b,c,d (1.693) (1.775) (1.756) (1.985)

Academic 8.636a,c,d 8.647a,c,d 8.701 a,c,d 8.796b,c,d (1.946) (1.965) (1.960) (2.055)

General 7.986a,b,d 7.845a,b 8.020a,b,d 8.1 76a,b,d (2.121) (2.157) (2.106) (2.087)

Vocational 7.481 b,c,d 8.007a,b 7.71 6a,b,c 7.41 3a,b,c (2.191) (2.122) (2.167) (1.877)

Total 8.470 8.487 8.483 8.490 (2.065) (2.068) (2.052) (2.053)

a Significantly different from the honors track at the .05 level. b Significantly different from the academic track at the .05 level. c Significantly different from the general track at the .05 level. d Significantly different from the vocational track at the .05 level.



38 Carbonaro

in effort across tracks? The next set of analyses (displayed in Table 3) examine several possible explanations of these differences in math. As I pointed out earlier, the same analyses were run in the other subjects, and the results were similar.

Model 1, the "base" model, indicates the unadjusted differences in teacher-reported effort, thereby simply restating the mean differences in effort by track displayed in Table 2 in a regression framework. (Note that the track coefficients are relative differences between a given track and general-track classes, the reference category.) Model 2 examines whether differences in background characteristics explain track differences in effort. Prior research has suggested that effort is related to social class (Cook and Ludwig 1998; Willis 1977), gender (Marks 2000; Smerdon 1999), and possibly race/ethnicity (Fordham 1996; Fordham and Ogbu 1986; Ogbu 1978, 2003). Since SES, gender, and race/ethnicity are related to students' track placements (see Gamoran and Mare 1989, Oakes 1985), it is possible that track differences in effort reflect these differences in students' characteristics across tracks. Consistent with prior research, female students exerted more effort than did male students, and higher-SES students exerted more effort than did low-SES students. The results revealed significant racial/ethnic differences in 10th-grade effort: black students exerted less effort (on average) than did white students, while Asian students exerted more effort than did students in any other racial/ethnic group. The effort exerted by Hispanic students was not significantly different from the effort exerted by white students.

After controls for race/ethnicity, gender, and SES were added to the model, the track differences in effort were reduced. For example, the differences in effort for students in the honors and academic tracks were reduced by about 20 percent from Model 1 to Model 2. Overall, background differences explain some of the track differences in effort, but sizable, statistically significant, differences remain.

The next set of models examine whether the criteria used to sort students into tracks explains track differences in effort. Model 3 adds a con-

trol for 8th-grade effort, and Model 4 adds the four 8th-grade achievement scores as predictors of 10th-grade effort.14 If students who are predisposed to exert greater effort are sorted into higher tracks, Models 3 and 4 should eliminate most of the track differences in effort. Both 8th-grade effort and 8th-grade achievement are significant predictors of 10th-grade effort. In addition, these variables explain a great deal of the track differences in effort. Controlling for 8th-grade effort reduces the honors coefficient in Model 2 by 28 percent and the academic coefficient by about 40 percent. The addition of controls for prior achievement in Model 4 reduces the coefficients observed in Model 2 even more: The effect of honors track is reduced by 60 percent, and the effect of academic track is reduced by almost 70 percent when both prior effort and achievement are controlled. A surprising finding is that net of background, prior effort, and achievement, vocational-track students exert more (not less) effort than do general-track students (see Model 4). Although it is tempting to con- struct a post hoc explanation for this finding, it should be noted that the finding is anomalous: Vocational-track students did not differ in the effort they exerted in Model 4 (or any subsequent model) in any of the three other subjects examined (results not shown).

Overall, Models 3 and 4 indicate that much, although not all, of the track differences in effort are due to the sorting of students who are more predisposed to exert greater effort-those who exert more effort and have higher achievement in the 8th grade-into higher-track classes in the 10th grade. By including these measures of prior effort and achievement in subsequent regression models, I was able eliminate these selection factors as threats to internal validity.

When prior effort and achievement are controlled, black-white differences in effort are no longer statistically significant, and the SES effect decreases dramatically (by two thirds). This finding suggests that black-white and SES differences in effort are explained mainly by differences in effort and achievement that exist before entry into high school. In contrast, the female and Asian coefficients do not decrease much (about 15 percent) when prior effort and achievement are con-




Table 3. Effects of Track Placement, Background, Prior Effort, Achievement, Students' Beliefs, and Intellectual Stimulation on Students' Effort

(1) (2) (3) (4) (5) (6) Variable Base + Background + Prior Effort + Prior + Students' + Intellectual

Achievement Beliefs Stimulation

Track Honors 1.556*** 1.273*** .91 7*** .496*** .370*** .330***

(.256) (.210) (.151) (.081) (.061) (.054) Academic .803*** .642*** .394*** .207** .117 .098

(.193) (.155) (.095) (.050) (.028) (.023) Vocational .168 .207 .242 .293* .368* .388**

(.013) (.01 7) (.020) (.024) (.030) (.032) Background

Female .767*** .610*** .662*** .647*** .643*** (.185) (.147) (.160) (.156) (.155)

Black -.353** -.251* -.059 -.215* -.248** (-.050) (-.035) (-.008) (-.030) (.034)

Hispanic -.129 -.055 .059 .003 .002 (-.019) (-.008) (.008) (.001) (.000)

Asian .532*** .494*** .454*** .41 2*** .411 *** . 6nn n)(.4f6 (.051) . n40) (.0 46)

SES .300*** .244*** .099** .001 .006 (.1 12) (.091) (.037) (.000) (.002)

Prior Effort 8th-grade effort .1 84*** .1 84*** .1 68*** .1 66***

(.269) (.240) (.219) (.216) Prior Achievement

Math .027*** .025*** .026*** (.I61) (.148) (.154)

Reading -.001 -.004 -.004 (-.006) (-.018) (-.01 7)

History .014 .007 .006 (.031) (.016) (.014)

Science .004 .002 .002 (.011) (.003) (.005)

Students' Beliefs Educational expectations .147*** .140***

(.147) (.140) Locus of control (Gi10) .232*** .213***

(.070) (.063) Self-concept (G1 0) .051 .045

(.016) (.015) Intellectual Stimulation

Challenge .011** (.040)

Show understanding .013*** (.050)

Adjusted R2 .060 .114 .179 .201 .221 .226

Note: Coefficients are unstandardized, and numbers in parentheses are standardized coefficients. * p < .05, ** p < .01, *** p < .001 (two-tailed tests).

trolled. Thus, female and Asian students appear to increase their levels of effort in high school, regardless of their previous history of working hard and academic achievement.

The next two models examine whether

factors related to students' different experiences within different track classes explain the remaining track differences in effort. In Model 5, 8th- and 10th-grade measures of students' beliefs about themselves and their future



40 Carbonaro

were added as controls in the regression. The inclusion of 8th-grade measures accounts for any prior differences in students' beliefs before their placement in their 10th-grade track. Hence, Model 5 examines whether changes in beliefs that are due to track placement are related to students' effort. Tenth- grade expectations and locus of control are both positively related to students' effort, while students' self-concept is not.15 Controlling for beliefs decreases the track differences in effort. The honors-track effect is reduced by 30 percent from Model 4. For the academic track, the coefficient becomes statistically insignificant. While the track differences in effort decrease when students' beliefs about themselves and the future are controlled, it should be noted that causality remains ambiguous: It is possible that greater effort leads to higher grades, which, in turn, raise students' self-concept, locus of control, and expectations, so these results must be interpreted with caution.

Finally, Model 6 adds controls for intellectual stimulation to the model. One way in which track placement may be related to effort is through greater intellectual stimulation in response to the more-demanding curriculum and stimulating instruction in higher-track classes. Both the level of challenge and the degree to which students are required to "show understanding" in math are positively related to effort, although the magnitude of the association is fairly modest. The results indicate that intellectual stimulation explains a modest portion (11 percent) of the honors effect on effort, but the coefficient remains statistically significant. If better measures of intellectual stimulation, such as indicators of a student's interest in or stimulation by a given class, were included in the model, more of the track differences in effort may be explained.

Thus, the overall conclusion drawn from Table 3 is that track differences in effort in the 10th grade are explained mostly by the process by which students are sorted into tracks: Students who were predisposed to exert more effort in the 10th grade (as evi- denced by their 8th-grade effort and achievement) were more likely to be placed in higher tracks. However, when combined, beliefs and intellectual stimulation explain a sizable

portion of the differences in effort across tracks: When the coefficients in Models 4 and 6 are compared, the honors-track effect on effort is reduced by 33 percent, and the academic coefficient becomes statistically insignificant. Thus, the results suggest that differences in effort across tracks reflect more than simply the types of students who are sorted into different tracks; rather, students' experiences in their classes also partly explain why higher-track students tend to exert more effort than do lower-track students.

It should be noted that the effects of the honors and vocational tracks, although small, remain significant in Model 6.16 The conceptual model displayed in Figure 1 suggests that after mediating variables are added as controls in the model, no track differences in effort should remain significant. There are several possible explanations for why the honors- and vocational-track coefficients remain significant in the final model. First, the mediating variables may be poorly measured in the models and hence do not serve as adequate controls for the concepts in the analyses. Second, there may have been unob- served variables that were omitted from the model that created track "effects" that are spurious. Finally, it is possible that teachers are susceptible to a "halo effect," whereby students' track placements affect teachers' judgments about their effort, independent of the actual levels of effort the students exert. Since the effort measure relies on some student behaviors that are observable by teachers (i.e., completion of homework and paying attention), this effect is probably minor. However, it is still possible that missed homework assignments and incidences of inatten- tion by lower-track students are perceived more readily and negatively by teachers, thereby artificially strengthening the track- effort relationship.

Effort and Track Effects on Achievement

The first two sets of analyses suggest that there are important differences in effort across tracks and that although these differences are largely the result of the different types of students who are placed in different




tracks, they are also partly due to students' experiences within these tracks. The next set of questions link effort with track differences in achievement. In particular, are track differences in learning partly explained by the greater effort exerted by students in higher tracks? The results of the analyses examining this question are displayed in Table 4.

The first model (Model 1) estimates the track differences in gains in math after background characteristics and 8th-grade achievement are controlled. These "adjusted track effects" indicate that regardless of students' prior math achievement, students gain more when they are placed in a higher math track. Controls for 10th-grade effort were added to the next model (Model 2) to establish whether differences in effort partly explain some of the track differences in math gains. Effort has a significant, positive effect on math gains. These effects are relatively large compared with the differences in math gains across tracks: A standard deviation increase in effort produces an average gain in math achievement that is two fifths and one quarter of the academic and honors effects (respectively).17

Does the positive relationship between effort and math gains partly explain the effects of track on learning? When the track coefficients in Model 1 are compared with those in Model 2, a fairly modest reduction in the track effects on learning is observed: The effect of honors track is reduced by roughly 10 percent, and the effect of academic track is reduced by 7 percent. In the other three subjects, effort explains slightly more of the track effects (generally from 10 percent to 20 percent), but most of the track effects remain unexplained (results not shown). Hence, although effort plays an important role in predicting learning, it explains little of the track effects on learning. If effort does not explain much of the track differences in achievement, what does? As Figure 1 (and prior research) suggests, differences in learning opportunities across tracks likely account for track differences in achievement. Unfortunately, this interpretation of the results cannot be con- firmed through further analysis because of the absence of adequate direct measures of learning opportunities in the NELS data.

How does controlling for students' effort and track placement affect the relationship between background characteristics and achievement? Controlling for effort does not change the racial/ethnic coefficients and only slightly reduces the SES effect on achievement. It is interesting that although the base- line model (Model 1) did not reveal gender differences in achievement, the addition of the effort measure revealed an advantage for male students. Hence, female students are closing the achievement gap in math with male students by exerting more effort in their math classes (see Table 3).

While 10th-grade effort has a strong effect on learning, this effect could be due largely to the fact that students who exert more effort are more likely to be sorted into higher tracks. To eliminate this possibility, Model 3 adds the control for 8th-grade effort as a predictor of 10th-grade learning gains. The addition of this variable to the model slightly reduces changes in the track differences in learning or the effects of effort on students' learning gains, but both remain sizable and significant. When 10th-grade effort is removed as a predictor of achievement in Model 3, the results indicate that 8th-grade effort explains some of the track differences in achievement, but only about half as much as 10th-grade effort does in Model 2.18 Thus, it appears that the effects of 10th-grade effort do not simply reflect the fact that students who try hard are more likely to be sorted into higher-track classes. Rather, students who exert greater effort in the 10th grade learn more, regardless of how much effort they exerted in the 8th grade.

Effects of Effort on Learning Across Tracks The last question of interest focuses on whether the effect of effort on learning varies across tracks. Model 4 in Table 4 examines this issue by adding interaction terms between effort and track placement to the model. None of the interaction terms is statistically significant at the .05 level. Hence, the results suggest that the effects of effort on learning are the same for all students, regardless of their track.



42 Carbonaro

Table 4. Effects of Track Placement, Effort, and the Track-Effort Interaction on Scores on the 10th-Grade Math Test

(1) (2) (3) (4) Adjusted + Effort + Prior Effort + Effort x

Variable Track Effects Track Interactions

Track Honors 4.866*** 4.451 *** 4.275*** 3.960**

(.119) (.109) (.104) (.097) Academic 3.474*** 3.252*** 3.090*** 3.515"***

(.124) (.117) (.110) (.126) Vocational -.670 -.836 -.796 .499

(-.008) (-.010) (-.010) (.006) Effort

10th-grade effort .602*** .534*** .563*** (.089) (.079) (.083)

8th-grade effort .213*** .212*** (.041) (.041)

Track x Effort Honors x Effort -.030

(.006) Academic x Effort -.051

(-.01 7) Vocational x Effort -.162

(-.016) Background

Female -.129 -.732*** -.733*** -.730*** (-.005) (-.026) (-.026) (-.026)

Black -1.573*** -1.513*** -1.479*** -1.477*** (-.033) (-.031) (-.031) (-.031)

Hispanic -.577* -.592* -.555* -.558* (-.012) (-.013) (-.122) (-.012)

Asian .425 .136 .139 .130 (.007) (.002) (.002) (.002)

SES .920*** .858*** .859*** .859*** (.051) (.047) (.047) (.047)

Prior Achievement Math .727*** .707*** .702*** .702***

(.632) (.615) (.611) (.611) Reading .114*** .115** .113*** .113***

(.071) (.071) (.070) (.070) History .209*** .1 94*** .186*** .1 85***

(.068) (.063) (.060) (.060) Science .1 75*** .1 74*** .1 76*** .1 76***

(.061) (.061) (.061) (.061)

Adjusted R2 .788 .794 .796 .796

Note: Coefficients are unstandardized, and numbers in parentheses are standardized coefficients. * p < .05, ** p < .01, *** p < .001 (two-tailed tests).




DISCUSSION

The main goal of this study was to explore the complex and multifaceted ways in which effort, curricular tracking, and learning are related. The major substantive conclusions of the analyses are as follows: (1) The higher a student's track, the more effort she or he exerts; (2) most of the differences in effort across tracks are explained by differences in prior effort and achievement across tracks, but factors related to students' experiences within tracks also explain track differences in effort; (3) effort is an important predictor of achievement, but effort does not account for a large proportion of the track effect on gains in achievement; and (4) the effect of effort on achievement gains does not vary across tracks.

Although the findings were mixed overall, they still suggest that examining effort in the context of curricular tracking is a fruitful direction to pursue. Better measures of effort that capture a wider range of behaviors may indicate that effort plays a greater role in explaining track differences in achievement and may reveal track-by-effort interactions. In particular, future studies should distinguish between the three types of effort identified here (rule oriented, procedural, and intellectual) and measure them separately. Researchers must recognize that each type of effort may be related to different outcomes. Rule-oriented effort (e.g., showing up for class regularly) is likely to be related to outcomes like high school graduation, but in the absence of greater procedural and intellectual effort, academic outcomes may not be drastically improved. Procedural effort (e.g., handing in homework regularly) may be important for grades, but increased learning and achievement are most likely to be related to students' level of intellectual effort (e.g., students' attempts to use their cognitive skills to comprehend the material). These different types of effort may also be emphasized and rewarded differentially across tracks. For example, teachers in low-track classes may emphasize and reward rule-oriented effort, whereas teachers in high-track classes may expect and encourage greater intellectual effort from students. Hence, track differences

in effort may be more pronounced if separate measures of the three types of effort identified here are analyzed. By extension, if higher-track classes produce more intellectual effort than do lower-track classes, higher- track students may experience greater gains in learning because intellectual effort is most likely to improve that specific outcome. While the measure of effort that was used in this article tries to capture rule-oriented, procedural, and intellectual effort, it is admittedly limited in scope. Future research that uses more-expansive measures of rule-oriented, procedural, and intellectual effort could greatly improve our knowledge of how effort, tracking, and learning are related.

Although the findings reveal some interesting insights regarding the relationships among tracking, effort, and achievement, it is necessary to consider some possible alternative explanations for the findings. One possible problem with the teachers' reports of effort is that they are formed in reaction to students' performance, not vice versa. In other words, teachers may believe that high- achieving students are diligent and attentive and complete homework assignments while low-achieving students are and do not; if teachers assess effort accordingly, then the relationship between effort and achievement may be biased because of measurement error.

While it is certainly possible that the observed relationship between effort and achievement is biased owing to errors in teachers' reports, there are two reasons to believe that this is not the case. First, the items focusing on attentiveness and the completion of homework are based, to some extent, on behaviors that teachers can recall, rather than simply subjective impressions (such as whether the student "tries hard"). Second, it is not clear that teachers could accurately predict students' learning gains from the 8th to the 10th grade. Since teachers themselves assign grades, they, of course, know them, but the same is not true for gains in test scores. Hence, while gains in test scores and grades are correlated, it is unlikely that teachers' reports of effort simply reflect students' gains in test scores.

What conclusion can be drawn regarding the importance of agency versus structure for



44 Carbonaro

learning? If agency is of paramount importance, track effects should disappear when effort is controlled. If structure is the most crucial, differences across tracks should remain when effort is controlled. Ultimately, it appears that both agency and structure are important factors that contribute to learning: Effort has important effects on learning even after track placement is controlled, but track placements are still related to gains in learning when effort is taken into account. In other words, when comparable students in lower- track classes try as hard as students in higher- track classes, they still learn less than they would in the higher track; however, when comparable students in the same track exert different levels of effort, students who exert more effort learn more. Thus, it appears that higher-track placements and greater effort are not mutually exclusive paths to higher achievement; academic rewards inhere to both the structural positions that students occupy in the curricular hierarchy and to their actions within these positions.

I hope that this study will stimulate future research by sociologists of education on effort and its role in creating unequal outcomes. This line of research will provide an important means for understanding how structure is related to agency and, ultimately, to academic outcomes. Curricular differentiation is just one example of an aspect of social structure within schools that may influence effort; others include status hierarchies between students, academic and normative school climates, ped- agogical practices within the classroom, and rule governing students' ability to choose their own courses. By linking school attributes with effort, it is possible to understand how structure shapes agency and, ultimately, how it shapes outcomes. In addition, research on these topics could deepen our insights into how "oppositional cultures" function. For example, do oppositional cultures overwhelm the school culture? Can school cultures effec- tively offset norms and beliefs that are created by groups of peers inside and outside the school? Questions such as these can help link different aspects of students' experiences into a larger gestalt that will deepen our under-

standing of class, racial, and gender inequalities in academic outcomes.

Finally, some important implications for policy and future research emerged from the findings. Generally, both policy makers and researchers pay greater attention to differences in learning opportunities among students than to differences in effort. For sociologists of education, this impulse to emphasize structure more heavily than agency may partly reflect a tendency for sociologists to resist explanations that may appear to "blame the victim" and discount the importance of social structure. This is a healthy response to the societal impulse to reduce all problems to the level of the individual and to overlook the influence of social structure on human action.

However, by de-emphasizing effort, researchers and policy makers overlook an important potential resource that all students have and can use to improve their academic outcomes. Researchers and policy makers need to consider how to create classroom environments that encourage all students to try hard in school. In exploring how school and classroom conditions are linked with effort, greater emphasis is actually placed on social structure and its relationship with human agency. Those who argue for increased effort by way of moral exhortation, couched in the language of "students' responsibility," ignore such important link- ages and fail to recognize that some peda- gogical practices are more likely than are others to motivate students to work hard. For example, the findings indicated that intellectual stimulation was positively related to effort, regardless of the track in which students were placed. Fortunately, there is a rich literature on motivation that can point both practitioners and policy makers in promising directions (for examples, see Corbett, Wilson, and Williams 2002; Ginsberg and Wlodkowski 2000). I hope that this research will stimulate new policy-relevant ideas about how to use effort as a means of redressing inequalities in students' outcomes.




Appendix Table A

Items Included in the 8th- and 10th-Grade Effort Scales

Survey Items Response Categories

8th-Grade Effort Student performs below his or her ability [BYTI_2, BYT42] 0 = no, 1 = yes Student rarely completes homework [BYT1_3, BYT4_3] Student is frequently absent [BYTI_4, BYT4_4] Student is frequently tardy [BYT1_5, BYT4_5] Student is inattentive in class [BYT1_6, BYT4_6] Student is exceptionally passive/withdrawn [BYT1_7, BYT4_7] Student is frequently disruptive [BYT1_8, BYT4_8]

10th-Grade Effort Does this student usually work hard in your class? [F1T1_2, F1T5_2] 0 = no, 1 = yes How often is this student attentive in class? [F1T1_18, F1T5_18] 0 = never, 1 = rarely,

2 = some of the time, 3 = most of the time,

4 = all of the time How often does this student complete homework assignments in your class? [F1T1_15, F1T5_15] Same as above

Note: The brackets contain variable names in the NELS:88 codebook.

NOTES

1. In this article, ability grouping refers to the practice of placing elementary school students in the same class into different groups for the purposes of instruction. In contrast, curricular tracking (or tracking for short) refers to the practice in high school by which students are placed in different classes for instruction in a given subject. For the sake of clarity, the two terms are meant to be mutually exclusive, although the term ability grouping is sometimes used more broadly to refer to practices like tracking. It should be noted that tracking is a somewhat outdated term because it suggests that a student's track is the same across academic subjects. However, research has suggested that this is no longer the case in American high schools, and students occupy different track positions in different academic subjects (Lucas 1999).

2. It should be noted that the field has not always focused so heavily on structure at the expense of agency. For example, Kerckhoff (1976) argued that research on the status attainment process was limited by its focus on socialization processes. He contended that researchers should pay more attention to

"structural limitations and selection criteria" in accounting for differences in attainment outcomes (p. 369).

3. One exception is Farkas et al. (1990), who used teachers' reports of "work habits" and found that differences between the work habits of black students and white students were not statistically significant.

4. Other interactions are plausible as well. For example, it is possible that the effects of effort vary by ability. Also, a three-way interaction among effort, ability, and learning opportunities may exist as well. Sorensen and Hallinan (1977) developed a sophisticated model for examining such interactions, but their approach and the questions it addresses are beyond the scope of this study.

5. It should be noted that Ainsworth- Darnell and Downey (1998) used the same scale in their study, but they used both teachers' reports in the same scale. The measure used in this study keeps the teachers' reports separate, so that effort is subject specific.

6. In a related paper, Carbonaro (2003) analyzed the NELS:88 data to examine how the characteristics of parents, peers, and teachers affect effort and achievement. In contrast to the analyses presented here, the analy-



46 Carbonaro

ses in that paper relied on combined teachers' reports of effort and global measures of track that were derived from students' transcripts. Although not the main focus of the paper, the findings regarding the relationships among tracking, effort, and achievement reported in that study were consistent with those reported here. However, since this article focuses specifically on tracking, effort, and achievement, subject-specific measures of effort and track are the most appropriate, both concep- tually and empirically.

7. Points 2 and 3 require some elaboration. First, teachers have a better understanding of the local context and the meaning attached to course labels than the transcript data can convey. Second, the transcript data provide course labels, but in subjects other than math, it is not possible to create course-based indicators of track because (1) the labels are not specific enough to distinguish between track levels (high versus low), and (2) no clear course-taking sequence is apparent. Further explanation and details of the analyses using the transcript data are available on request.

8. Jencks (1985) argued that when controlling for prior achievement, it is preferable to add controls for test scores in different academic subjects (e.g., including 8th-grade math, reading, history, and science scores as predictors of 10th-grade math scores), rather than simply controlling for a single test score in one academic subject (e.g., 8th-grade math scores as a control predicting 10th- grade math scores). He contended that prior achievement will be measured more reliably when multiple test scores are added to the regression because the additional scores will help correct any measurement error in one particular test score.

9. The 8th- and 10th-grade scales were created using students' 8th- and 10th-grade reports on the same items. The items are as follows: (1) "I don't have enough control over my life"; (2) "Good luck is more important than hard work"; (3) "Every time I try to get ahead, something or somebody stops me"; (4) "My plans hardly ever work out, so plan- ning only makes me unhappy"; (5) "When I make plans, I can almost always make them work"; and (6) "Chance and luck are very important in what happens in my life." The

respondents were asked, "how they felt about" the foregoing items and could choose from a scale that ranged from "strongly agree to strongly disagree."

10. The 8th- and 10th-grade scales were created using students' 8th- and 10th-grade reports on the same items. The items are as follows: (1) "I feel good about myself"; (2) "1 am a person of worth, the equal of others"; (3) "1 am able to do things as well as most other people"; (4) "On the whole, I am satis- fied with myself"; (5) "I certainly feel useless at times"; and (6) "At times, I think I am no good at all." The respondents were asked "how they felt about" the foregoing items and could choose from a scale that ranged from "strongly agree to strongly disagree."

11. The response categories for this variable were as follows: for 8th-grade expectations, (1) "won't finish high school," (2) "will finish high school," (3) "will attend vocational/trade/business school after high school," (4) "will attend college," (5) "will finish high school," and (6) "will attend a higher school after college"; for 10th-grade expectations, (1) "less than high school graduation," (2) "high school graduation only," (3) "less than two years of trade school," (4) "more than two years of trade school," (5) "less than two years of college," (6) "two or more years of college," (7) "finish college," (8) "master's degree," and (9) "Ph.D. or M.D."

12. Ideally, adequate measures of learning opportunities would include variables, such as instructional time, curricular coverage, instructional quality, and teacher quality. Although NELS includes some information on these aspects of students' schooling experiences, the measures are generally crude. However, it should be noted that Carbonaro and Gamoran (2002) had some success in using these measures in their examination of achievement in English. It is unclear whether similar measures could be created in math or other subjects.

1 3. The sample sizes differ across academic subjects because some students did not take courses in each of the four subjects in their junior year.

14. The bivariate correlation between 8th- and 10th-grade effort is .346. Eighth-grade achievement is also positively correlated with




10th-grade effort, with correlations ranging from a high of .306 (effort and math) to a low of .224 (effort and science).

15. Neither the 8th-grade measure of locus of control nor students' self-concept was statistically significant at the .05 level in Model 5 or 6. Eighth-grade expectations had a small, negative, statistically significant relationship with effort (p = .01).

16. The honors coefficient is statistically significant in Model 6 for English and science as well.

17. In the other three subjects, the effect of effort on achievement is larger relative to the track effects than is the case in math. For example, in English, a standard-deviation increase in effort produces an increase in reading gains that is larger than the effect of academic track and two thirds that of the effect of the honors track.

18. in the other three subjects, the results indicate that 8th-grade effort (apart from 10th-grade effort) explains almost none of the track differences in achievement. Hence, math is unique because it is the only subject in which 8th-grade effort seems to contribute to track differences in achievement.

REFERENCES

Ainsworth-Darnell, James, and Douglas Downey. 1998. "Assessing Racial/Ethnic Differences in School Performance." American Sociological Review 63:536-53.

Alexander, Karl, and Martha Cook. 1982. "Curricula and Coursework: A Surprise Ending to a Familiar Story." American Sociological Review 47:626-40.

Allison, Paul. 2002. Missing Data. Thousand Oaks, CA: Sage.

Barr, Rebecca, and Robert Dreeben. 1983. How Schools Work. Chicago: University of Chicago Press.

Bishop, John. 1999. "Nerd Harassment, Incentives, School Priorities, and Learning." Pp. 231-80 in Learning and Earning: How Schools Matter, edited by Susan Mayer and Paul Peterson. Washington, DC: Brookings Institution Press.

Bong, Mimi, and Richard Clark. 1999. "Comparison Between Self-Concept and Self- Efficacy in Academic Motivation Research." Educational Psychologist 34:139-53.

Brint, Steven, Mary Contreras, and Michael

Matthews. 2001. "Socialization Messages in Primary Schools: An Organizational Analysis." Sociology of Education 74:157-80.

Brophy, jere, and Thomas Good. 1986. "Teacher Behavior and Student Achievement." Pp. 328-75 in Handbook of Research on Teaching, edited by Merlin Wittrock. New York: Macmillan.

Bryk, Anthony, Valerie Lee, and Peter Holland. 1993. Catholic Schools and the Common Good. Cambridge, MA: Harvard University Press.

Carbonaro, William. 2003. "Teacher, Peer, and Parent Social Capital: Influences on Student Effort and Academic Outcomes." Paper presented at the "Creation and Returns to Social Capital" Colloquium, Royal Netherlands Academy of Arts and Sciences, Amsterdam, the Netherlands.

Carbonaro, William, and Adam Gamoran. 2002. "The Production of Achievement Inequality in High School English." American Educational Research Journal 39:801-27.

Cook, Philip, and jens Ludwig. 1998. "The Burden of 'Acting White': Do Black Adolescents Disparage Academic Achievement?" Pp. 375-400 in The Black-White Test Score Gap, edited by Christopher Jencks and Meredith Phillips. Washington, DC: Brookings Institution Press.

Corbett, Dick, Bruce Wilson, and Belinda Williams. 2002. Effort and Excellence in Urban Classrooms: Expecting--and Getting--Success with All Students. New York: Teachers College Press.

Eder, Donna. 1981. "Ability Grouping as a Self- Fulfilling Prophecy: A Micro Analysis of Teacher-Student Interaction." Sociology of Education 54:151-62.

Farkas, George, Robert Grobe, Daniel Sheehan, and Yuan Shuan. 1990. "Cultural Resources and School Success: Gender, Ethnicity, and Poverty Groups Within an Urban School District." American Sociological Review 55:127-42.

Farkas, George, Christy Lleras, and Steve Maczuga. 2002. "Does Oppositional Culture Exist in Minority and Poverty Peer Groups? Comment on Ainsworth-Darnell and Downey." American Sociological Review 67:148-55.

Felmlee, Diane, and Donna Eder. 1983. "Contextual Effects in the Classroom: The Impact of Ability Groups on Student Attention." Sociology of Education 56:77-87.

Ferguson, Ronald. 2001. "A Diagnostic Analysis of Black-White GPA Disparities in Shaker Heights, Ohio." Pp. 347-414 in Brookings Papers on Education Policy: 2001, edited by Diane



48 Carbonaro

Ravitch. Washington, DC: Brookings Institution Press.

Fordham, Signithia. 1996. Blacked Out: Dilemmas of Race, Identity, and Success at Capital High. Chicago: University of Chicago Press.

Fordham, Signithia, and John Ogbu. 1986. "Black Students' School Success: Coping with the 'Burden of "Acting White."'" Urban Review 18:176-206.

Gamoran, Adam. 1986. "Institutional and Instructional Effects of Ability Grouping." Sociology of Education 59:85-98. 1987. "The Stratification of High School Learning Opportunities." Sociology of Education 60:135-55.

Gamoran, Adam, and William Carbonaro. 2002-03. "High School English: A National Portrait." High School Journal 86:1-13.

Gamoran, Adam, and Robert Mare. 1989. "Secondary School Tracking and Educational Inequality: Compensation, Reinforcement, or Neutrality?" American journal of Sociology 94:1146-83.

Ginsberg, Margery, and Raymond Wlodkowski. 2000. Creating Highly Motivating Classrooms for All Students. San Francisco: Jossey-Bass.

Hallinan, Maureen. 1992. "The Organization of Students for Instruction in the Middle School." Sociology of Education 65:114-27.

Hoffer, Thomas. 1992. "Middle School Ability Grouping and Student Achievement in Science and Mathematics." Education Evaluation and Policy Analysis 14:205-27.

Jencks, Christopher. 1985. "How Much Do High School Students Learn?" Sociology of Education 58:128-35.

Johnson, Monica Kirkpatrick, Robert Crosnoe, and Glenn Elder, Jr. 2001. "Students' Attachment and Academic Engagement: The Role of Race and Ethnicity." Sociology of Education 74:318-40.

Jones, James, Beth Vanfossen, and Margaret Ensminger. 1995. "Individual and Organizational Predictors of High School Track Placement." Sociology of Education 68:287-300.

Kerckhoff, Alan. 1976. "The Status Attainment Process: Socialization or Allocation?" Social Forces 55:368-81.

King, Gary, James Honaker, Anne Joseph, and Kenneth Scheve. 2001. "Analyzing Incomplete Political Science Data: An Alternative Algorithm for Multiple Imputation." American Political Science Review 95:49-69.

Lucas, Samuel. 1999. Tracking Inequality: Stratification and Mobility in American High Schools. New York: Teachers College Press.

Lucas, Samuel, and Adam Gamoran. 2002. "Tracking and the Achievement Gap." Pp. 171-98 in Bridging the Achievement Gap, edited by John Chubb and Tom Loveless. Washington, DC: Brookings Institution Press.

MacLeod, Jay. 1995. Ain't No Makin' It: Aspirations and Attainment in a Low-Income Neighborhood (2nd ed.). Boulder, CO: Westview Press.

Marks, Helen. 2000. "Student Engagement in Instructional Activity: Patterns in the Elementary, Middle, and High Schools Years." American Educational Research Journal 37:153-84.

Mickelson, Roslyn. 1990. "The Attitude- Achievement Paradox Among Black Adolescents." Sociology of Education 63:44-61.

Natriello, Gary, and Edward McDill. 1986. "Performance Standards, Student Effort on Homework, and Academic Achievement." Sociology of Education 59:18-31.

Newmann, Fred, ed. 1992. Student Engagement and Achievement in American Secondary Schools. New York: Teachers College Press.

Oakes, Jeannie. 1985. Keeping Track: How Schools Structure Inequality. New Haven, CT: Yale University Press.

-. 1990. Multiplying Inequalities. Santa Monica, CA: RAND.

Ogbu, John. 1978. Minority, Education, and Caste. New York: Academic Press.

-. 2003. Black American Students in an Affluent Suburb: A Study of Academic Disengagement. Mahwah, NJ: Lawrence Erlbaum.

Page, Reba. 1991. Lower Track Classrooms: A Curricular and Cultural Perspective. New York: Teachers College Press.

Pallas, Aaron, Doris Entwisle, Karl Alexander, and Maria Stluka. 1994. "Ability-Group Effects: Instructional, Social, or Institutional?" Sociology of Education 67:27-46.

Rosenbaum, James. 2001. Beyond College for All: Career Paths for the Forgotten Half. New York: Russell Sage Foundation.

Rowan, Brian, and Andrew Miracle, Jr. 1983. "Systems of Ability Grouping and the Stratification of Achievement in Elementary Schools." Sociology of Education 56:133-44.

Smerdon, Becky. 1999. "Engagement and Achievement: Differences Between African- American and White High School Students." Research in Sociology of Education and Socialization 12:103-34.

Serensen, Aage, and Maureen Hallinan. 1977. "A Reconceptualization of School Effects." Sociology of Education 50:273-89.

Statacorp. 1999. Stata Statistical Software: Release 6.0. College Station, TX: Stata Corporation.




Suarez-Orozco, Marcelo. 1987. "Towards a Psychosocial Understanding of Hispanic Adaptation to American Schooling." Pp. 156-68 in Learning to Learn: A Multicultural Perspective, edited by Henry Trueba. Newberry, MA: Newberry Press.

Tach, Laura, and George Farkas. 2003. "Ability Grouping and Educational Stratification in the Early School Years." Paper presented at the annual meeting of the American Sociological Association, Atlanta, GA.

Wang, Jia. 1998. "Opportunity to Learn: The

Impacts and Policy Implications." Educational Evaluation and Policy Analysis 20:1 37-56.

Weis, Lois. 1990. Working Class Without Work: High School Students in a De-industrializing Economy. New York: Routledge.

Willis, Paul. 1977. Learning to Labor. New York: Columbia University Press.

Yair, Gad. 2000. "Educational Battlefields in America: The Tug-of-War over Students' Engagement with Instruction." Sociology of Education 73:247-69.

William Carbonaro, Ph.D., is an assistant professor of sociology at the University of Notre Dame. His research interests are in the areas of education and social stratification. He is currently working on several projects that focus on how students' high school experiences affect racial/ethnic differences in college graduation.

The author thanks Adam Gamoran, Michael Olneck, Warren Kubitschek, Maureen Hallinan, Sean Kelly, and attendees of the University of Wisconsin-Madison sociology of education brownbag series for their valuable feedback on an earlier version of this article and Bridget Nicholson for her valuable research assistance on this project. Address all correspondences to William Carbonaro, Department of Sociology, University of Notre Dame, 1016 Flanner Hall, Notre Dame, IN, 46556; e-mail: [email protected].



tracking, students' effort, and academic achievement

Documents