Linking Teachers’ Perceptions of Educational Value Discontinuity to Low- Income Middle School Students’ Academic Engagement & Self-Efficacy

Within the last forty years, particularly with the publication of Rosenthal and Jacobsen’s classic, Pygmalion in the Classroom (1968), there has been increased examination of teacher and classroom-based characteristics and their association with middle school students’ achievement and achievement correlates. For example, many education researchers believe that the expectations teachers hold for students are associated with students’ attitudes towards learning and also their academic performance (Alvidrez & Weinstein, 1999; Gill & Reynolds, 1999; Good & Nichols, 2000; Jussim & Harber, 2005; Kuklinski & Weinstein, 2001; Rubie-Davies, 2006).

While this line of research has contributed to a greater understanding of the role teacher expectations play in student academic performance, the empirical literature linking teachers’ perceptions of students’ home environments to performance is limited (Arunkumar, Midgley, & Urdan, 1999; Hauser-Cram, Sirin, & Stipek, 2003). Even less is known about the association between teachers’ perceptions of students’ home and personal characteristics and the psychological and behavioral antecedents of academic performance. In particular, no study has examined whether teachers’ perceptions of educational values salient in the home are linked to low-income middle grade students’ reports of academic engagement or academic self-efficacy. Anecdotal data suggest that teachers’ expectations for students are premised, to some degree, on their perceptions of students’ demographic backgrounds and their out-of-school socialization experiences (Bakari, 2003; Maccoby, 1992; Neal, McCray, Webb-Johnson, & Bridgest, 2002; Seyfried, 1998). Thus, it is important to determine whether such perceptions of this out-of-school socialization—as it pertains to educational values—are linked to students’ reports of their academic efficacy and motivation. More importantly, research should seek to determine whether teachers’ perceptions of these experiences—and the values that inform them—are linked to these academic performance antecedents.

The purpose of the current study, therefore, seeks to examine the predictive relationships between teachers’ reports of perceived educational value discontinuity (between themselves and parents of middle school students) and academic task engagement and academic self-efficacy among middle grade school students (grades 6-8) in Language Arts classes. One reason for choosing Language Arts classrooms in this study is the given importance of reading and writing as a way to enhance overall student performance. That is, even though some researchers have found that self-efficacy, for many elementary and middle grade students, is domain specific (Eccles, Midgley, Wigfield, Buchanan, Reuman, & Flanagan, 1993), central to optimal performance in each academic domain is the student’s ability to read and comprehend written material. The major research question driving the current study asks whether teachers’ reports of perceived educational value discontinuity are statistically significant predictors of student reports of emotional, behavioral, and cognitive engagement and academic self-efficacy.

It has been shown that teachers, often times, have concepts of learning that may differ from those held by parents and students, especially those from ethnically or culturally diverse backgrounds (Gay, 2000; Rogoff, 2003). This notion, referred to here as value discontinuity, has been considered a factor that complicates the school achievement, adjustment and attainment of many ethnic minority students, namely African American, Native American and Latino students (Gay, 2000; Phelan, Davidson, & Cao, 1991; Rogoff, 2003; Tyler, Haines, & Anderman, 2006). Cultural value discontinuity, a more common form of value discontinuity, views the perceived disconnect between the in-school and out-of-school cultural values of ethnic minority children as a source of academic complexity for them (Neal, et al., 2003; Rogoff, 2003).

Similar to cultural value discontinuity, perceived educational value discontinuity is described as a mismatch between educational values espoused at home and those manifested in the classroom (Hauser-Cram et al., 2003). An example of an educational value would be the emphasis both parents and teachers place on the importance of reading (Hauser-Cram et al., 2003). This mismatch of educational values and corresponding activities is believed to impact student performance outcomes and their antecedents as well (Phelan, Davidson, & Coe, 1991). Some research has corroborated this claim.

For example, Hauser-Cram, et al. (2003) found that teachers’ perceptions of dissimilar educational values between themselves and parents resulted in the teacher-held belief that students were inadequately supported for learning. Hauser-Cram et al. (2003) also found that teachers rated students less competent and had lowered expectations for students’ future success when they believed parents’ education-related values differed from their own. Also, Hauser-Cram et al. (2003) found that perceived educational value dissonance was significantly linked to perceptions of poor literacy and mathematics skills.

Similar findings were reported in a longitudinal study by Arunkumar et al. (1999), where African American and European American fifth grade students reporting high home-school dissonance also reported lower levels of academic and emotional well-being, efficacy, self-esteem and GPA, along with higher levels of anger and self-deprecation. While Arunkumar and colleagues (1999) did not specify the type of dissonance being experienced by students—that is, whether the discontinuity was based on cultural or educational values—the stated effects of home-school dissonance, did have implications for student achievement, both directly (i.e., significant association with GPA) and indirectly (i.e., significant association with antecedents of GPA such as efficacy, esteem and academic and emotional well-being).

Grade level homogeneity, however, limits the generalizability of value discontinuity findings occurring in these past studies. Arunkumar and colleagues (1999) included only a 5^th grade sample, while Hauser-Cram and colleagues (2003) utilized only a kindergarten sample for their study. Also, the current literature does not contain evidence to support whether perceptions of educational value discontinuity between parents and teachers are related to academic self-efficacy and task engagement for middle grade students. Some description of the academic self-efficacy and the various forms of task engagement are presented below.

Self-efficacy refers to one’s beliefs about his or her ability to learn and/or carry out a task or set of behaviors at an identified, optimal level of performance (Bandura, 1997; Pajares, 1996). Some research has noted the unique contribution of self-efficacy over other self or personcentered variables such as self-esteem and self-concept. For example, Pajares (1996) argued that self-efficacy beliefs are better predictors of individual behavior than self-concept and self-esteem. Also, selfefficacy has been found to be more closely associated with academic engagement and classroom learning than self-concept (Linnenbrink & Pintrich, 2003).

Indeed, self-efficacy is considered to be one of the person-centered variables associated with student performance (Bandura, 1997). That is, the belief in one’s ability to carry out a task effectively shapes, to a large degree, the actual task or tasks to be carried out (task choice), how hard and how long he or she will work to complete the task at an optimal level of performance (effort and persistence), and finally, whether the task is performed at an optimal level (achievement) (Linnenbrink & Pintrich, 2003). According to Bandura and colleague Dale Schunk (1981; Bandura, 1997; Schunk, 2003) and others (Pajares, 1996; Wigfield & Eccles, 2000), individuals high in self-efficacy participate in the task more often, work harder, evidence greater persistence in performing the task and ultimately, complete the task far more successfully than students who do not feel as efficacious.

Regarding public school student performance, several studies on self-efficacy among elementary and middle school students have shown the positive association between perceived or self-reported self-efficacy and academic motivation, engagement, and performance (Bandura, 1997; Bandura, Barbaranelli, Caprana, & Pastorelli, 2001; Bandura and Schunk, 1981; Linnenbrink & Pintrich, 2003; Pajares, 1996; Pintrich & DeGroot, 1990; Schunk, 2003; Wigfield & Eccles, 2000). In addition to the outcomes predicted by academic self-efficacy, predictors of selfefficacy have also been identified in the empirical literature, including classroom behavior modeling (Schunk & Rice, 1989) and teacher based social persuasions (Usher & Pajares, 2006). The empirical literature on self-efficacy has, indeed, helped to shape a greater understanding of its role in student academic performance. Yet, the relationship between teachers’ perceptions of students’ educational values—as determined by reported perceptions of parents’ educational values—and students’ academic self-efficacy is unknown.

Strong academic/school engagement has been consistently linked to lower dropout rates and higher levels of academic success and is a strong predictor of optimal student achievement and appropriate school behavior, irrespective of socioeconomic status (Finn & Rock, 1997; Voelkl, 1995). Students with lower levels of academic/school engagement are at risk for outcomes such as disruptive behavior in class, absenteeism, and school drop out (Steinberg, Brown, & Dornbusch, 1996).

The literature finds student engagement to consist of three components, behavioral, cognitive, and affective engagement (Linnenbrink & Pintrich, 2003; Skinner & Belmont, 1993). Behavioral engagement involves observable behaviors. In a classroom setting, this can include observations of effort, persistence, and help-seeking (Birch & Ladd, 1997; Linnenbrink & Pintrich, 2003). Along with involvement in extracurricular activities, behavioral engagement focuses on positive behavior in school, such as following rules and classroom norms, attending school and staying out of trouble (Finn, Pannozzo, & Voelkl, 1995; Finn & Rock, 1997). Behavioral engagement has been linked to positive academic outcomes and preventing students from dropping out (Fredricks, Blumenfeld, & Paris, 2004). According to these researchers, students who exhibit more effort, persist longer at tasks, and seek help with academic tasks are more likely to learn and achieve at higher levels.

Cognitive engagement encompasses deep thinking about the material to be learned, thinking about what one knows and does not know about the material (Weinstein & Mayer, 1986) as well as critical and creative thinking about the material (Linnenbrink & Pintrich, 2003). Cognitively engaged students use meta-cognitive practices to plan, monitor, and evaluate their progress and effort (Pintrich & De Groot, 1990). Further, engaged students rehearse, summarize, and elaborate material in order to better remember, organize, and understand it (Corno & Madanach, 1983; Weinstein & Mayer, 1986). Higher levels of cognitive engagement have been linked to increased learning and higher levels of achievement (Weinstein & Mayer, 1986).

According to Fredricks et al. (2004), affective or emotional engagement refers to students’ positive and negative reactions to teachers, classmates, academic content and curricula, and school. Allan Wigfield and Jacquelynne Eccles (Wigfield & Eccles, 2000) have proposed and researched an expectancy-value model of achievement motivation containing emotion-based factors said to precede student performance. Referred to as achievement values, the most recent articulation of the expectancy-value model of achievement motivation has identified five such values that, for some researchers, (Linnenbrink & Pintrich, 2003) fit conceptually under the emotional engagement construct. The first achievement value component is the perceived importance or attainment value of the task (e.g., How important is it for me to do well on this task). Following attainment value is intrinsic interest that the student holds for the material (e.g., Do I enjoy doing this task?), utility value (e.g., Does this task help me to achieve a future goal I hold?), and cost of doing the task (e.g., What must I sacrifice to complete this task optimally?) (Wigfield & Eccles, 2000).

Along with these identified achievement values is affect, which includes emotions like anger and anxiety, as well as happiness and excitement. Affect is typically associated with the students’ perception of how important it is to complete a task successfully, the interest in the task, the perceived utility of the task and the experienced cost of completing the tasks (Linnenbrink & Pintrich, 2003). Thus, if a student in intrinsically interested in completing a homework assignment, believes it taps into his or her preexisting interests and will be useful to him or her in the future and the cost of carrying out such a task is minimized, then he or she will likely feel happy about completing the task. Opposing achievement values may result in feelings of sadness or discontent.

Emotional engagement factors (i.e., interest, importance, utility and cost), as articulated within the expectancy-value process model, do not have a direct link to learning and achievement. Rather, these affective factors—when they are positive—have been found to increase cognitive engagement and the use of metacognitive strategies (Pintrich & DeGroot, 1990).

The purpose of this study is to investigate the association between perceived educational value discontinuity, as reported by middle grade Language Arts teachers, and student reports of academic engagement and academic self-efficacy. Significant relationships are expected to emerge between teachers’ perceptions of educational value discontinuity and academic self-efficacy and academic task engagement.

The middle school in the present study was randomly selected from a list of low-income middle schools located in a municipality in the Southeastern region of the United States. Specifically, a random numbers table was used to identify one middle school located within 15 miles of the institution hosting the research. Low-income status was determined by a majority percentage of students receiving free or reduced lunch rates. At the participating school site, the percentage of students receiving free or reduced lunch was 87%. As an incentive for participation, the school principal was offered $250 to purchase supplies for faculty and staff.

Convenience sampling technique was used to gather students for participation in the study. Researchers were given permission from both the school principal and Language Arts teachers to collect data in all the Language Arts classrooms located in the school. In total, there were twelve Language Arts classrooms. The study sample consisted of 271 low-income middle school students selected from these twelve Language Arts classrooms.

Fifty-five percent of the sample was male. Grade level distribution for students included 28% in the sixth grade, 43% in the seventh grade, and 29% in the eighth grade. Fifty-seven percent of the students were African American, 24% European American and 18% were Latino. Current demographic data for the school district show that African American students make up roughly 24% of all middle school students, European Americans make up 62%, and Latinos make up roughly 8% of middle school students.

Of the 324 students given an informed consent form to have signed by their parents, 271 returned the forms and participated in the research, yielding a return rate of 83%. Participating students were provided one dollar for returning informed consent forms and a $5 gift card to a local department store for completing the survey.

In addition, six general education Language Arts teachers also participated in the study. The teacher sample included two African- American females and four European-American females. Each teacher was the instructor of record for two of the Language Arts courses where student participants completed surveys. Range of teaching experience in terms of years went from less than one year to 22 years. Data from all teachers were included in analyses. Participating Language Arts teachers were offered $35 for completing surveys on students in each participating class.

The Hauser-Cram Scale consisted of five questions that determine the extent to which teachers believe that parents endorse the same education-related values as themselves (Hauser-Cram, Sirin, & Stipek, 2003). The five areas addressed are discipline, parental role in education, and the teaching of math, literacy, and writing. Teachers responded to each item using a three-point scale (1 = no difference, 2 = some difference, 3 = great difference). The higher the score, the greater are teachers’ perceptions of differences in education-related values between themselves and students’ parents. Hauser-Cram et al. (2003) reported a reliability coefficient of .92, indicating strong internal consistency for this scale. However, because teachers in the current study did not instruct students in math, the item addressing math was eliminated. Additionally, the responses were changed to reflect a four- point scale. The scale values for the revised Hauser-Cram scale included 1= no difference, 2 very little difference, 3 = some difference, and 4 = complete difference. A pilot study was conducted to provide a reliability coefficient for the resulting four-item, four-response scale. The pilot results yielded an reliability coefficient of .88. A sample item from the Hauser-Cram scale included: “Are there differences between parents’ values or preferences and your values with respect to the educational program in discipline?”

The Student Engagement scale was developed by Fredricks, Blumenfeld, & Paris (2004). This instrument is comprised of three subscales: Emotional Engagement, Behavioral Engagement, and Cognitive Engagement, each of which was given to third to fifth grade students across five elementary schools in previous work (Fredericks, et al., 2004). The emotional engagement subscale responses were on a five-point, Likert-type scale, with responses ranging from 1 = Strongly Disagree to 5 = Strongly Agree. A sample item for this subscale is “My classroom is a fun place to be”. The behavioral engagement subscale had responses on a five-point scale, ranging from 1 = “Never” to 5 = “Always”. An example of an item in this subscale is “I pay attention in class”. The cognitive engagement subscale employed the same set of responses as the Behavioral Engagement subscale. A sample item for this scale read “I study at home even when I don’t have a test.”

The Academic Efficacy measure is a subscale of the Patterns of Adaptive Learning Scales (PALS) developed by (Midgley, Maehr, Hruda, Anderman, Anderman, Freeman, Gheen, Kaplan, Kumar, Middleton, Nelson, Roeser, & Urdan, 2000). The entire instrument measures multiple dimensions of the learning environment and student characteristics. The academic efficacy subscale contains five questions and uses a five-point, Likert-type scale. Responses range from “Not at all true = 1” to “Very true = 5”. The samples used in the validation of the scales represent nine school districts in three Midwestern states (Midgley, et al., 2000). The reliability coefficient derived for previous work from the Academic Efficacy measure was. 78. Reliability coefficients for each instrument used in the current study are located in Table 1.

All six Language Arts teachers agreed to participate in the study. A schedule for distributing and collecting parental consent from students was devised. Informed consent forms were provided to teachers and were sent home to parents by students. Informed consent forms were collected from teachers prior to data collection. During data collection, student assent, privacy issues, and procedural information were provided to the students. During their regular Language Arts class period, in their assigned classroom, participating students were administered a stapled packet in a coded envelope that contained: a demographic sheet, the Academic Engagement scales, and the Academic Efficacy scale. Non-participating students were directed by the classroom teacher to engage in a non-academic task. Data collection required approximately thirty to forty minutes to complete per classroom.

At the end of the school day, the teacher was given an envelope containing the Hauser-Cram instrument, which was to be completed for each student participating in the study. Teachers also received a demographic form and were given one week to complete the survey packet. All teachers returned completed survey packets.

Cronbach’s alpha reliability coefficients, descriptive and correlation analyses were calculated for each of the instruments in the study and are presented in Table 1. The data analytic plan for this study include one multivariate analysis of variance to determine the effects of student race, gender, grade, teacher race and years of teaching experiences on the five dependent variables of interest, namely academic self-efficacy, cognitive, behavioral and emotional engagement, and perceived educational value discontinuity.

Also, a bi-variate correlation matrix was computed to examine statistically significant associations among study variables and final, hierarchical linear regression analyses were computed to determine whether perceived educational value discontinuity was a statistically significant predictor of academic self-efficacy and emotional, cognitive and behavioral engagement scores.

One multivariate analysis of variance was computed to determine significant differences in reports of perceived educational value discontinuity, academic self-efficacy, and emotional, cognitive and behavioral engagement as a function of student race (African American, European American, and Latino), gender (male and female), grade (sixth, seventh and eighth), teacher race (African American and European American) and years of teaching experience (less than one year, one to two years, three to four years and five or more years). Multivariate analyses showed significant F-statistics for the student race [F (2, 263) = 3.16, p < .01, ή² = .05] and grade [F (2, 263) = 6.76, p < .01, ή²= .06]. Uni-variate analyses revealed significant effects of race on emotional engagement [F (2, 263) = 5.67, p < .01, ή² = .06] and behavioral engagement [F (2, 263) = 5.90, p =.02, ή²= .06] and of the grade variable on emotional engagement [F (2, 263) = 6.30, p =.01, ή² =.08], self-efficacy [F (2, 263) = 9.03, p < .01, ή²= .11] and educational value discontinuity [F (2, 263) = 26.38, p < .01, ή²= .18].

Post hoc analyses revealed that Latino students (3.86) reported significantly higher emotional engagement than either African American (3.40) or European American (3.50) students. Also, Latino students (4.07) had significantly higher behavioral engagement than African American students (3.74). Also revealed were significant differences in emotional engagement and self-efficacy as a function of grade. Specifically, sixth (3.78) and seventh (3.64) grade students were more emotionally engaged than eighth grade students (3.34). Also, sixth (4.27) and seventh (4.08) grade students had significantly higher reported academic self-efficacy than eighth grade Language Arts students (3.81). Regarding educational value discontinuity, sixth grade students (1.51) had significantly lower educational value discontinuity reports than seventh and eight grade students (2.43 and 2.42, respectively).

Several statistically significant, negative associations emerged between educational value discontinuity and academic engagement and efficacy. Educational value discontinuity was also negatively associated with student grade and teacher race. Here, lower educational value discontinuity scores—as reported by Language Arts teachers—were associated with higher academic engagement and self-efficacy scores, as reported by middle grade students. Also, students in the higher grades were perceived by teachers to have parents whose educational values differed from their own. Also, based on the nominal coding for African American and European American teachers (1 = African American and 2 = European American), it appears that European American teachers had higher perceptions of educational value discontinuity than did their African American counterparts.

Also, several positive correlations emerged among the three engagement subscales, indicating their fidelity as engagement constructs. Finally, significant associations were found between selfefficacy and all engagement subscales, demonstrating stated associations among these factors (Bandura et al., 2001; Linnenbrink & Pintrich, 2003; Wigfield & Eccles, 2000). See Table 1 for all correlation coefficients.

To address the major research question, several hierarchical regression analyses were computed to determine which variables would be statistically significant predictors of emotional, behavioral, and cognitive engagement, and academic efficacy. Examination of the variance inflation factor (VIF) values and tolerance levels indicated that multicollinearity was not a factor with the data as values were found within the acceptable range. Predictor variables included in the hierarchical regression analyses were: student race, gender, and grade, along with teacher-reported perceptions of educational value discontinuity. The race and gender variables were dummy coded to allow for the categorical variables to be included in the regression analyses. Student grade was considered an ordinal variable in the regression analyses. Demographic variables (e.g., student race, grade) were entered first, followed by educational value discontinuity. Interaction terms between race and grade and between race and perceived educational value discontinuity were also created and included in the final step of each regression analysis.

Educational value discontinuity (β = -.09, t = -1.04, p = .30) was not a significant predictor of emotional engagement (mean = 3.52, N = 264). Student grade did, however, emerge as a statistically significant predictor of emotional engagement (β = -.23, t = -3.22, p = < .01). Here, students in the lower middle grades had higher emotional engagement scores than students in the higher grades. Variables in the final regression model [F (9, 262) = 3.10, p < .01] accounted for 11% of the variance in emotional engagement scores. Student and teacher demographic variables accounted for 10% of the variance in emotional engagement scores, collectively.

On the behavioral engagement subscale (mean = 3.85, N = 264), hierarchical regression analyses revealed that teacher perceptions of educational value discontinuity (β = -.33, t = -4.22, p < .01) was a statistically significant predictor of behavioral engagement. Being a female student also emerged as a significant predictor of behavioral engagement (β = .20, t = 2.65, p < .01). Variables in the final regression model [F (9, 262) = 7.16, p < .01] accounted for 22% of the variance in behavioral engagement scores. Educational value discontinuity uniquely accounted for 11% of the variance in behavioral engagement scores.

Educational value discontinuity (β = -.06, t = -.74, p = .46) was not a significant predictor of cognitive engagement (mean = 2.96, N = 263). Student grade did, however, emerge as a statistically significant predictor of emotional engagement (β = -.25, t = -3.47, p = < .01). Here, students in the lower middle grades had higher cognitive engagement scores than students in the higher grades. Variables in the final regression model [F (9, 262) = 2.67, p < .01] accounted for 10% of the variance in behavioral engagement scores. Student and teacher demographic variables—including student grade—accounted for 9% of the variance in behavioral engagement scores, collectively.

For self-efficacy (mean = 4.09, N = 265), teacher reports of educational value differences between themselves and parents (β = - .35, t = -4.29, p < .01) was predictive of academic self-efficacy scores. Variables in the final significant regression model [F (9, 264) = 5.10, p < .01] accounted for 17% of the variance in self-efficacy scores. Educational value discontinuity uniquely accounted for 16% of the variance in academic self-efficacy scores.

The purpose of this study was to examine the predictive relationships between perceived educational value discontinuity and academic engagement and academic self-efficacy. The literature is replete with examples of how varying teacher characteristics are predictive of students’ performance and psychological outcomes (Alvidrez & Weinstein, 1999; Hudley, 1997; Sirin & Rogers-Sirin, 2004; Tucker, Zayco, Herman, Reinke, Trujillo, Carraway, Wallack, & Ivery, 2003). Yet, very few have examined teachers’ perceptions of parents’ educational values and whether these perceptions are linked to students’ academic performance antecedents and actual classroom behavior. At the middle grade level, this research is virtually nonexistent.

The findings from the present study support the existence of statistically significant relationships between educational value discontinuity between school and home and several teacher and student-centered variables and outcomes. To begin, initial bi-variate correlation coefficients showed positive relationships between teacher race and educational value discontinuity and student grade and educational value discontinuity. Here, students in the higher grades were perceived by teachers to have parents whose educational values differed from their own.

Some literature suggests that there is a significant decrease in parent involvement as students progress through middle school (Cotton & Wikelund, 2001). This is oftentimes attributed to the difficulty of academic material which moderates parents’ ability to be more involved in their children’s schooling (Cotton & Wikelund, 2001). Such perceptions of parents’ academic involvement and competence—especially for low-income parents—may be held by middle grade teachers. As a result, these teachers may have lowered expectations regarding whether parents’ educational values match their own.

In addition, it was shown that European American teachers had higher perceptions of educational value discontinuity between home and school than did their African American counterparts. Some works have concluded that one of the possible outcomes of the interaction between a mostly mainstream/middle-class European American female teaching force and a predominantly low-income, ethnic minority student population is teachers’ misperception of students’ orientation towards education (Gay, 2000; Ladson-Billings, 2005).

Further, although educational value discontinuity was negatively associated with scores on each engagement subscale and the academic self-efficacy, it only emerged as a statistically significant predictor of low-income middle school students’ behavioral engagement and academic self-efficacy reports. The inverse relationships that emerged in the regression analyses for educational value discontinuity and behavioral engagement and educational value discontinuity and academic self-efficacy suggest that lower perceived discordance in parent and teachers’ educational values was predictive of behavioral engagement and academic self-efficacy scores reported above scale midpoints. That is, the less educational value discontinuity Language Arts teachers reported for their students, the higher the level of behavioral engagement these students exhibited in their classrooms. Also, less educational value discontinuity was linked to higher reports of academic self-efficacy among the middle school sample, irrespective of grade, ethnicity or gender.

Findings of students in the lower middle grades having higher emotional and behavioral engagement scores than students in the higher middle grades are consistent with findings in the literature (Greenwood, Horton, & Utley, 2002; Marks, 2000). Also, female middle grade students being more behaviorally engaged in classes is also consistent with research indicating that, during the elementary and middle school years, female students are more inclined to demonstrate behaviors consistent with classroom rules and learning goals than are male students (Mickelson & Greene, 2006; Van Houtte, 2004).

Overall, these findings have implications for educational policy and practice. Specifically, it has been noted that parents of middle grade students tend to be less involved in students’ academic activities than parents in primary and high school grade levels (Cotton & Wikelund, 2001). However, when parents display adequate involvement in their children’s academic lives and manifest educational values deemed consistent with their teachers, it seems to have positive benefit for their children in middle school. Thus, educational policymakers and school administrators should look to enhance the relationship—particularly the communication—between teachers and parents so that there is some alignment with respect to their educational values.

Overall, these data demonstrated that perceptions of what happens in students’ homes are extremely important to the engagement and selfefficacy of middle-school students. Similar to the teacher expectation literature, where favorable teacher expectations are linked to positive outcomes for students, these data illustrate that teachers’ perceptions of students’ home backgrounds and values are associated with how these students think of themselves as learners and how engaged they are in class.

Some limitations to the current study exist. To begin, this study focused on low-income, middle school students in one middle school located in the Southeastern region of the United States and thus, findings should be interpreted with caution as external validity to samples dissimilar to the current one is compromised. Due to an inability to retrieve Language Arts performance data from school personnel, actual Language Arts grade indices could not be examined in this study. Further, this study was conducted in general education, Language Arts classrooms. Findings may differ for students in mathematics classrooms, special education classes, and/or English as a Second Language (ESL) classes. Future research should look to extend this line of work by repeating this study in the aforementioned classrooms with greater number of student and teacher participants as the low number of teacher participants also limits generalizability of the findings. What’s more, the coefficients of determination (i.e., R- squares) were relatively low. This indicated that, given the multiple factors present in each hierarchical regression model, there remain several additional students and teacher-based factors that can meaningfully predict middle school students’ academic engagement and self-efficacy. Future studies should include these additional factors that have been shown to be significantly associated with academic engagement and self-efficacy (Bandura, 1997; Linnenbrink & Pintrich, 2003; Wigfield & Eccles, 2000). Finally, future research should include the actual performance indices of students in addition to the psychological antecedents of these.