Achievement and at-Risk Middle School Students’ Perspectives of Academic Support

Similar to the research findings in school climate literature, character education initiatives recognize that all students have needs for safety, belonging, and the experience of meaningfully contributing, furthermore they are more likely to internalize the values and expectations of groups that meet these needs (Berkowitz & Bier, 2003; CEP, 2003). Thus, one of the common strands in character education and school climate is community (Berkowitz, Vincent, & McKay, 2002). As such, an effective character education initiative creates a caring school community (CEP, 2003). Although character education is intended to promote student character development (Berkowitz & Bier, 2006), there has been relatively little research on the role of school climate and sense of community and its impact on student achievement. Still, the possible linkages among the social context of the school, character education, students’ problem behaviors, and subsequently academic achievement have begun to be explored (Battistich & Hom, 1997).

Research has shown that students who experience school as a community enjoy school more, engage in less disruptive behavior, and have higher academic achievement (Battistich, Solomon, Kim, Watson, & Schaps, 1995). Furthermore, research findings demonstrated that higher levels of schools’ sense of community were associated with less delinquent behavior and better academic achievement (Battistich & Hom, 1997). Thus, it is likely that there would be higher academic achievement among students who perceive their school as a caring community.

According to the findings reported by Berkowitz and Bier (2006), good character education supports and enhances the academic goals of schools, that is, student achievement. Berkowitz and Bier (2006) reviewed 33 scientifically supported programs in an effort to synthesize existing scientific research on the effects of K-12 character education. One approach to their task was to observe the percentage of variables for a specific outcome category for each program and determine the ones that were most consistently positively impacted by the effort. What they found was that attachment to school (61%) and academic achievement (59%) were among the top ten outcomes that most consistently had a positive impact.

Both literature found within the school climate movement and stemming from character education initiatives suggest that in a caring school community, the daily life of classrooms and all other parts of the school environment are imbued with a climate of concern and respect for others (CEP, 2003; Berkowitz et al., 2002; Doyle & Doyle, 2003; Levine & Lezotte, 1990; Marzano, 2003). Students report more positive forms of motivation and academic engagement when they perceive their school implements learner-centered practices that involve caring (Berkowitz & Grych, 2000; Doyle & Doyle, 2003; Meece, 2003, Saunders & Saunders, 2002). Implicit in these findings is the idea that developing a strong sense of community, one built on support and caring, occurs for the students when stronger and more positive relationships are built with teachers, school staff, peers, and the community at-large (Berkowitz et al., 2002; Doyle & Doyle, 2003; Saunders & Saunders, 2002). These relationships, in turn, are likely to facilitate educational resilience and improve academic achievement.

Although the community at-large plays a valuable role in the character development and academic achievement of young people, the students’ parents are the most influential in this regard (Berkowitz, 1998; CEP, 2003). Therefore, in addition to developing caring school communities, schools that reach out to families and include them in all aspects of educating a child greatly enhance the student’s chances for success (Berkowitz, Schaeffer & Bier, 2001; CEP, 2003; Deiro, 1996; Levine & Lezotte, 1990; Marzano, 2003). However, Farkas, Johnston, and Duffet (1999) reported that only 27% of teachers were satisfied with the level of parental involvement in their schools; likewise, 70% of parents indicated that they should be more involved in the education and development of their child.

Schools that empower parents, as well as other community organizations, enhance structures that lead to improved student success. Many character education programs understand that families and communities play a vital role in school based initiatives and student achievement (Berkowitz & Bier, 2006). In caring school communities, parents are empowered in the process of educating their child when allowed to share in authority and decision making in schools (Doyle & Doyle, 2003). When parents are empowered in this manner, they become part of the solution and provide support for more than just their own students (Novick, Kress, & Elias, 2002).

Educational resilience is defined as “the heightened likelihood of success in school and in other life accomplishments, despite environmental adversity” (Wang, Haertel, & Walberg, 1993, p. 3). Research on educational resilience has reported both risk factors and protective factors associated with better academic outcomes (e.g., Aronson, 2001; Morales, 2000; Rouse, Bacama-Gomez, Newman, & Newman, 2001). Research observations have consistently indicated factors that place students at-risk for academic failure or underachievement: students from urban or inner-city communities (McMillan & Reed, 1994; Reed, McMillan, & McBee, 1995; Rouse et al., 2001; Wang, Haertal, & Walberg, 1997), from rural communities (Elder & Conger, 2000; McMillan & Reed, 1994; Murry & Brody, 1999; Reed et al., 1995), have families that are identified as low-income or have a low socioeconomic status (SES) (Aronson, 2001; Elder & Conger, 2000; Horn & Chen, 1998; Murry & Brody, 1999), have low self-esteem (Devlin, 1995), have problem behaviors (Elder & Conger, 2000; Hudley, Daoud, Hershberg, Wright-Castro, & Polanco, 2002; Malecki & Elliott, 2002), and have had poor academic performance in the past (Cappella & Weinstein, 2001; Catterall, 1998).

However, several protective factors have also been examined and these include internal locus of control (Borman & Rachuba, 2001; Cappella & Weinstein, 2001; Grossman, Bein-ashowitz, Anderson, Sakurai, Finnin, & Flaherty, 1992; Grotberg, 1995; Wang, Haertel, & Walberg, 1998; Westfall & Pisapia, 1994), aspirations (Cappella & Weinstein, 2001; Hudley et al., 2002; Rouse et al., 2001; Westfall & Pisapia, 1994), high expectations (Finley, 1994; Horn & Chen, 1998), academic and social behaviors or skills (Borman & Rachuba, 2001; Devlin, 1995; Elder & Conger, 2000; Gibbs, 1998; Grotberg, 1995; Malecki & Elliott, 2002; Rouse et al., 2001), involvement with school (Borman & Overman, 2004; Catterall, 1998; Chavkin & Gonzalez, 2000; Elder & Conger, 2000; Hawkins & Mulkey, 2005; Horn & Chen, 1998; Westfall & Pisapia, 1994), high self-esteem, self-efficacy or self-confidence (Borman & Overman, 2004; Borman & Rachuba, 2001; Chavkin & Gonzalez, 2000; Finley, 1993; Westfall & Pisapia, 1994), and support (Aronson, 2001; Cappella & Weinstein, 2001; Catterall, 1998; Chavkin & Gonzalez, 2000; Elder & Conger, 2000; Finley, 1994; Grossman et al., 1992; Grotberg, 1995; Horn & Chen, 1998; Hudley et al., 2002; Kenny, Gallagher, Alvarez-Salvat & Silsby, 2002; Nettles, Mucherah & Jones, 2000; Rouse et al., 2001; Sanacore, 2000; Wang et al., 1997; Wang et al., 1998; Westfall & Pisapia, 1994). While there are several factors that can place students at-risk, the protective factors facilitate academic achievement.

In terms of psychological resilience, support is one of the most widely examined protective factors (e.g., Garmezy, 1985; Masten, Best, & Garmezy, 1990; Werner, 1990). Research on academic resilience has focused on three major sources of support: peer, family, and school. Additionally, the types of support were divided into social support and academic support. While the results for peer and family support have been mixed with positive and neutral findings, the school level research has been predominantly positive. For example, studies reporting specifically on mathematics or reading achievement found that peer support was not a significant factor (Cappella & Weinstein, 2001; Nettles et al., 2002) while studies examining general achievement noted that peers were important (Horn & Chen, 1998; Westfall & Pisapia, 1994); furthermore, Rouse et al. (2001) observed mixed results. In their examination of reading achievement, Cappella and Weinstein (2001) found that peer attitudes about school did not significantly contribute to high school students’ resilience. Similarly, Nettles et al. (2000) found that peer social support was not related to reading or mathematics achievement for their sample of elementary school students. In contrast, Horn and Chen (1998) found that friends’ college plans had a strong impact on college enrollment; furthermore, the students in Westfall and Pisapia’s (1994) study indicated that peers were important contributors to their resilience. Perhaps the findings from Rouse et al. (2001) illuminate why there are discrepancies, as the students in their qualitative study also reported mixed perceptions about the facilitative nature of peer support. While a few students indicated that peer support was beneficial, most of their participants perceived that friends were detrimental to their academic success, as they could succumb to negative peer pressure.

These inconsistencies were also observed for studies examining family support, whether academic (Catterall, 1998; Horn & Chen, 1998; LaForett, Watt, Diaz, McCullough, & Barrueco, 2000; Rouse et al., 2001) or social (Grossman et al., 1992; Kenny et al., 2002; LaForett et al., 2000; Nettles et al., 2000; Westfall & Pisapia, 1994). LaForett et al. (2000) found that a more autonomous parenting style had a positive impact on reading achievement for headstart students; however, parental involvement in school was not related to achievement. Similarly, Nettles et al. (2000) reported that family social support did not have a significant impact on elementary school students’ achievement in reading or mathematics. However, others have found that perceptions of family cohesiveness (Grossman et al., 1992), perceptions about family communication (Grossman et al., 1992), and supportive maternal relationships (Kenny et al., 2002) were important factors for older students.

In terms of academic support, Catterall (1998) reported that when families encouraged educational endeavors it had a positive impact on the mathematics achievement of their nationally representative sample of high school students. Furthermore, Horn and Chen (1998) found that family discussions of academic expectations had a strong impact on college enrollment. However, when interviewed, resilient adolescents rarely nominated their parents (less than 1%) as significantly contributing to their academic success (Rouse et al., 2001). Perhaps these discrepancies could be linked back to the findings of Farkas et al. (1999), such that parents are not as involved as they would like to be in their children’s education.

The importance of school support is one area that consistently received positive appraisal. From school culture (Borman & Overman, 2004; Catterall, 1998; Cappella & Weinstein, 2001; Horn & Chen, 1998; Rouse et al., 2001) to teacher responsiveness (Catterall, 1998; Hudley et al., 2002; Rouse et al., 2001) and counselor support (Westfall & Pisapia, 1994), schools have been a supportive source for students. For example, in their national study of elementary school students, Borman and Overman (2004) found that schools with supportive climates were the most successful in facilitating resilience. This model included three components: safe and orderly environment, positive student-teacher relations, and commitment to parental involvement; thus, illustrating the importance of the relationship among school climate, resilience, and support. Similarly, in their small study examining elementary school students’ perceptions of social support, Nettles et al. (2000) found that teachers’ support positively and significantly impacted mathematics achievement (but not reading achievement). While both studies only examined the achievement of elementary school students, other studies examining reading/English achievement (Cappella & Weinstein, 2001), college enrollment (Horn & Chen, 1998), and school engagement (Hudley et al., 2002) of older students yielded positive findings.

Interestingly, despite the rich history of character education (Berkowitz, 1998), school climate, and the increased examination of educational resilience (e.g., Aronson, 2001; Chavkin & Gonzalez, 2000; Elder & Conger, 2000; Rouse et al., 2001; Sanacore, 2000; Wang et al., 1997; Wang et al., 1998), research linking these areas is extremely sparse. Yet, a review of the literature within these fields demonstrates a link worthy of examination (see Table 1). For example, literature found within the field of character education suggests that social support from peers and family are important components of educating a child (Berkowitz & Grych, 2000; CEP, 2003, Lickona, 1991). Similarly, examinations of social support research reported in the educational resilience literature have also been positive (Aronson, 2001; Chavkin & Gonzalez, 2000; Elder & Conger, 2000; Finley, 1994; Rouse et al., 2001; Sanacore, 2000; Wang et al., 1997; Wang et al., 1998).

Research has focused on support from parents (Bean, Bush, McKenry, & Wilson, 2003; Gadeyne, Ghesquiere, & Onghena, 2004; Rutter, 1979; Sukon & Jawahir, 2005), peers (Azmitia & Cooper, 2001), and teachers (Barber & Olsen, 2004; Hudley et al., 2002; Turner & Patrick, 2004) with little information available about school counselors (Gordon, 1992; Westfall, & Pisapia, 1994). In general, research has revealed that supportive relationships are associated with positive school outcomes such as increased achievement (Azmitia & Cooper, 2001; Kenny et al., 2002; Sukon & Jawahir, 2005), decreased delinquency (Gordon, 1992; Rutter, 1979) or discipline problems (Gibbs, 1998). Like Bronfenbrenner (1979, 1986), Epstein (1987) suggests that these relationships are interconnected; likewise, the Character Education Partnership (2003) stresses the importance of these relationships.

Although the importance of students’ perceptions of social support has been widely acknowledged (e.g., Demaray & Malecki, 2002a, 2002b; Malecki & Elliott, 2002; Malecki & Demaray, 2002; Malecki & Demaray, 2003), little research has specifically examined the impact of academic support on achievement. Therefore, this paper extends previous work by investigating the impact of middle school students’ perceptions of academic support on their academic achievement. Consistent with Ford’s (1992) Motivational Systems Theory, Rouse (2003) notes, “students’ beliefs about their academic environment are important for motivation and achievement” (p. 7). Thus, this paper examines the impact of students’ perceptions of parent, peer, teacher, and counselor academic support on their academic achievement for eighth grade students identified as at-risk for academic failure. Furthermore, this study examines how these different sources of academic support impact the achievement of students experiencing different levels of risk.

1. How do students’ perceptions of academic support (i.e., peer, parent, teacher, and counselor) predict their achievement?

2. How do students’ perceptions of academic support operate for students with different levels of risk?

Participants were selected from a larger pool of students who participated in the NELS (U.S. Department of Education, 1993). For NELS, 24,559 students were surveyed in Grades 8, 10, 12, and postcollege. From those 24,559 students, 11,004 reported experiencing at least one factor that placed them at-risk for academic failure. These risk factors included: single parent family, low parent education, sibling had dropped out, 3 or more hours home alone, low English proficiency, and/or low SES. Of those students reporting at least one risk factor, 7,813 (3,699 boys, 4114 girls) had complete data for the variables of interest in this study. Their racial background varied: 1.7% American Indian/Alaskan Native, 6.2% Asian/Pacific Islander, 16.5% African American, 18.1% Latino/a, and 57.5% Caucasian. They were from urban (n = 2,513), suburban (n = 2,936), and rural (n = 2,364) schools. The students experienced between one and five risk factors: 59.3% one risk factor, 27.9% two risk factors, 9.8% three risk factors, 2.6% four risk factors, and 0.3% five risk factors.

There were several variables of interest in this study: demographics, achievement test scores, risk factors, and academic support. Demographics, achievement test scores, and risk factor data were exported directly from the NELS data set (U.S. Department of Education, 1993).

The achievement variables included: reading (NELS variable name: by2xrirr), mathematics (NELS variable name: by2xmirr), and science (NELS variable name: by2xsirr) achievement. The three achievement scores were combined into one general achievement score. The scores were highly correlated and combined have an internal consistency of .81. A maximum likelihood factor analysis (oblique rotation) was conducted and the results indicate one factor (i.e. achievement) accounting for 79% of the variance (see Table 2 for the factor loadings and communalities).

The risk factor variable was a composite directly exported from NELS (NELS variable name: byrisk). Students were then divided into risk groups based on the number of risk factors that they experienced: low risk (n = 4,636), middle risk (n = 2,182), and high risk (n = 995). The lowest risk group experienced one risk factor, the middle risk group experienced two risk factors, and the highest risk group experienced three or more risk factors.

The academic support variables were created from individual items in the NELS data set: parent (11 items), peer (2 items), teacher (11 items), and counselor (6 items) (see Table 3 for an explanation of these composite variables). To create these variables, two of the authors independently content analyzed the Base Year NELS student questionnaire (National Opinion Research Center, 1988). Discrepancies were discussed until agreement was reached. Cronbach’s alpha was calculated for the support composites: parent α = .68 (11 items), teacher α = .74 (11 items), counselor α = .76 (6 items). Because there were only two items examining peer support, Cronbach’s α was not calculated.

A random intercept model using the MIXED (multilevel) component of SPSS 12.0 was used. A random intercept model was selected because multistage cluster sampling was implemented when collecting the NELS data (i.e., schools were first sampled and then students were selected from those schools) and multiple students were selected from the same school; thus, these analyses allowed for the exploration of potential differences due to the nesting of students within schools (see Bryk & Raudenbush, 1992) and interclass correlations ranged from .19 to .23. As Barcikowski (1981), noted when the interclass correlations are positive the Type 1 error is inflated. Alpha was set at .01 to preserve an experimental wise error rate of .05, as three analyses were conducted. See Table 4 for a summary of the descriptive statistics for the support variables and achievement. See Table 5 for a summary of the descriptive statistics for achievement by risk group. See Table 6, Table 7, and Table 8 for the impact of academic support on general achievement by risk group.

How do students’ perceptions of academic support predict their achievement? Across all three risk groups, the more parent and peer support a student received the higher the general achievement score and interestingly, the more counselor support received the lower the general achievement score (b = -.16, SE =.18, p < .001).

How do students’ perceptions of academic support operate for students with different levels of risk? Teacher support was only statistically significant for the lowest risk group. Given the nature of the questions for teacher and counselor support, it seems that talking about career and high school plans is not associated with higher scores; however, it appears that talking with parents and peers is beneficial. This may be due to the fact that both parents and peer statements had specific homework and academic domain items. Lastly, a statistically significant difference between boys and girls on general achievement was only evident with the lowest risk group (b = 1.77, SE =.60, p = .003) where boys on average had a higher score than girls.

Consistent with previous research, it appears that parent (Berkowitz, 1998; Berkowitz & Bier, 2006; Catterall, 1998; Horn & Chen, 1998; Sukon & Jawahir, 2005) and peer support (Azmitia & Cooper, 2001; Horn & Chen, 1998; Westfall & Pisapia, 1994) are important factors in the achievement of these students. Nevertheless, teacher support was only significant for the students experiencing one risk factor; furthermore, counselor support appears to have a negative impact. These findings for teacher and counselor support are somewhat surprising with a 25-year history noting the importance of support in schools (e.g., Chambers et al., 2005; Edmonds, 1979); however, there is limited research available examining the impact of school counselors’ academic support on students’ academic resilience (e.g., Gordon, 1992; Westfall & Pisapia, 1994). Furthermore, Catterall (1998) suggests that when examining students placed at-risk, researchers should use past performance and behavior rather than grouping students based on risk factors. Perhaps the results would be different if students were selected based upon their past academic performance rather than risk factors. This is an important distinction that necessitates additional research.

While the results for teacher support (e.g., Catterall, 1998; Hudley et al., 2002; Rouse et al., 2001) and counselor support (Gordon, 1992; Westfall & Pisapia, 1994) may not be consistent with previous research, this study specifically examined students’ perceptions of academic support. Much of the previous research has focused on social support (e.g., Chavkin & Gonzalez, 2000; Elder & Conger, 2000); thus, more research is needed to clarify this potential discrepancy between social and academic support. Moreover, the academic support variables were constructed from preexisting data (NELS) (U.S. Department of Education, 1993); consequently, there were limitations to the kinds of questions students were asked. In addition, the actual amount of support that students received was not measured; rather, the survey relied on students’ self-reports. One final consideration that could be taken into account about teacher and counselor support has been their changing role as a result of the new high stakes test movement. Thus, further investigation into the construction of academic support variables is warranted.

Theoretically, the importance and interconnectedness of relationships have been noted (Bronfenbrenner, 1979, 1986; CEP, 2003; Epstein, 1987); so too has the importance of perception (Ford, 1992) and students’ perceptions of support (Demaray & Malecki, 2002a, 2002b; Malecki & Demaray, 2003). Moreover, researchers such as Elliott, Malecki, and Demaray (2001) note that there is an interaction among social support, school climate, and academic achievement; thus, a framework that incorporates these important and potentially facilitative relationships is necessary.

While this paper did not examine a character education program, it did attempt to highlight the interconnectedness among character education, school climate, and educational resilience. Imbued in many of the CEP (2003) principles is the importance of support. There are different types of support (e.g., social, academic) as well as many sources of support (e.g., parent, peer, teacher, counselor). For example, the importance of community has been acknowledged (e.g., Ceballo & McLoyd, 2002; Petri, Geiger, Boling, Hartline & Powers, 2001); however, due to the limitations of the NELS data set, we were unable to explore this area of academic support. Catterall (1998) observed that high school students’ involvement with the community promoted educational resilience. Similarly, Finley (1993) noted the importance of community support in the resilience of students placed at-risk in rural communities. While the community is a significant contributor to students’ resilience, so too is the school’s environment. Consistent with CEP (2003), Gibbs (1998) notes that a caring school community also fosters resilience; furthermore, the school environment can be enhanced through social skill instruction for students as it can enhance students’ abilities to effectively contribute to the school community (Elliott et al., 2001; Gibbs, 1998). Correspondingly, Malecki and Demaray (2002) found that social skills predicted future achievement. Thus, to help inform programmatic practices, it would be beneficial for future research to examine the impact of different academic and social support systems, including contributions made by the community.