Abstract
Academic failure and exclusionary discipline practices are key elements in what has been described as the “school to prison pipeline.” While there exists a strong body of research on the risks for delinquency, few studies have addressed the variables within schools that exacerbate or counteract these risks. We examined the academic pathways in English/Language Arts (E/LA) of a group of students recommended for expulsion in one school district, including students with and without disabilities. We obtained the letter grades that students received and the statewide assessment (SWA) results from elementary through high school for 81 participants who were recommended for expulsion in high school. Multiple t tests demonstrated that overall, students’ performance decreased as they progressed through school and students’ SWA scores differed considerably from course performance grades in elementary school. We discuss the results in terms of implications and recommendations for improving the educational pathways of students at risk of school expulsion.
The U.S. Department of Education estimated that more than 100,000 students were expelled in the 2005–2006 school year, which represented a 15% increase in expulsions since the 2001–2002 school year (Brownstein, 2009). Because students have historically been expelled from schools for violent behavior, this increase may suggest that students are becoming more violent in schools. However, changes in school discipline policies may offer a better explanation. For example, in 1994, zero tolerance became a federal policy with the Gun-Free Schools Act (20 U.S.C. §§ 8921-23), which mandated a 1-year expulsion from school for a student who brought a weapon to school. Over time, zero tolerance policies have been extended to include other, less serious behavioral infractions such as bringing drugs or cigarettes to school, and causing school disruption (Skiba & Peterson, 1999).
For example, South Carolina’s regulations for the standards of student conduct and disciplinary enforcement procedures for schools describe three levels: Level I includes disorderly conduct, defined as activities that impede orderly classroom and school procedures (e.g., tardiness, abusive language, and failure to carry out directions). Level I sanctions may entail such responses as verbal reprimands, detention, and in-school suspension. Level II includes disruptive conduct, defined as activities directed against persons or property (e.g., threats, fighting, abusive language to staff). Level II sanctions may involve such responses as out-of-school suspension, alternative education program, and expulsion. Level III includes criminal conduct, defined as activities that result in violence to persons or property or that poses a direct and serious threat to the safety of persons in the school (e.g., assault and battery, bomb threat, possession of weapons). Level III sanctions may consist of out-of-school suspension, alternative education program, expulsion, and restitution of damages. The regulations also state that Level I behaviors may be reclassified as Level II if the behavior occurs 3 or more times. Thus, a student could be expelled for repeating behaviors that are not directed against persons or property. In fact, the reason given for more than half of the expulsion hearings in this study was consistent problems. In other words, zero tolerance policies have evolved to the point of excluding students from school for many types of misbehavior.
The problem with zero tolerance policies in schools is that it does not address any of the circumstances that may have contributed to the misbehavior, such as economic disparities, prejudice, inequitable school structures and practices, and unjust hierarchies in our culture that sanction discriminatory punishment, and forceful control over youth (Casella, 2001). In a report on the effects of zero tolerance policies, the Civil Rights Project at Harvard University concluded that such policies are contrary to children’s developmental needs, deny children educational opportunities, and often lead to criminal behavior (Civil Rights Project, 2000). Unfortunately, there is little empirical evidence documenting the consequences of zero tolerance, exclusionary discipline practices in our public schools (Skiba, 2002). However, we do know that school expulsion is a risk factor that sets the student on a trajectory that can lead to delinquency and failure to graduate from school (Christle, Jolivette, & Nelson, 2005; Morrison et al., 2002).
Failing to graduate from high school not only results in grave consequences for the individuals, it places burdens on society as well. Students who do not graduate from high school average 27% less income per year than high school graduates, which translates to approximately US$250,000 in reduced income over a lifetime. Students who do not graduate are more likely to experience health problems, engage in criminal activities, and become dependent on welfare and other government programs (Martin, Tobin, & Sugai, 2002; Stanard, 2003). An estimated 25% of freshmen do not graduate from high school within 4 years in the United States (Stillwell, 2009), and this represents a tremendous waste of human potential and productivity. Research has suggested that students who do not graduate from high school comprise 52% of welfare recipients, 82% of the prison population, and 85% of juvenile justice cases (Stanard, 2003). Thus, it is imperative that we work toward identifying those students who are at risk of failing to graduate from high school and find solutions to increase graduation rates.
Most research on expelled students has focused on demographic characteristics such as age, ethnicity, gender, and disability status, whereas a few researchers have examined developmental and contextual characteristics of expelled students such as behavioral difficulties and family problems (Hayden, Shepherd, & Ward, 1996; Imich, 1994). Few studies have examined school-related factors such as academic achievement even though academic failure has been identified as a strong risk factor for school failure and delinquency (Lerner & Galambos, 1998; Wald & Losen, 2003). In one study, Morrison and D’Incau (1997) examined the files of 158 students (elementary, middle, and high school) who were recommended for expulsion and found that most of these students were failing in English and Math. Kirkwood and Richardson (2006) examined the files of 104 expelled students of ages 15 to 18, and of the 68% who had academic achievement data, 67% were identified as “below average.” In the present study, we tracked the academic achievement in English/Language Arts (E/LA) from elementary school through high school of students recommended for expulsion in high school.
The Academic-Behavior Connection
Research has suggested a statistically significant relationship exists between academic skills and problem behavior in schools. Several researchers have identified that academic problems often foster behavior problems, and behavior problems negatively affect academic achievement; a relationship that can lead to subsequent delinquency (Brunner, 1993; Coleman & Vaughn, 2000; Drakeford, 2002; Leone, Krezmien, Mason, & Meisel, 2005; Malmgren & Leone, 2000). Maguin and Loeber (1996), in a meta-analysis of studies that had investigated the relation between academic deficits and youth delinquency, found that low school achievement predicted delinquency. They identified three major relations between academic performance and delinquent behavior. First, the onset, frequency, persistence, and seriousness of delinquent behaviors are related to poor academic performance. The opposite also is true; the absence of delinquent behaviors is related to higher academic performance. Second, attention problems and cognitive deficits are related to poor academic performance and delinquent behavior. Third, increases in academic performance are associated with decreases in rates of delinquency. Strategies that promote academic success, therefore, are protective factors against delinquency.
Reading and language arts skills are particularly critical for students to succeed in school. Brunner (1993) suggested that frustration resulting from reading failure may lead to delinquent behavior. He concluded that by the end of elementary school a snowball effect occurs, increasing frustration and vulnerability to later risks such as maladaptive and aggressive behaviors. Inevitably, these students develop negative attitudes and patterns of behavior in reaction to years of poor reading and academic failure (Denti & Guerin, 2004). This relationship between reading failure and misbehavior is described in a well-documented failure cycle that school personnel may inadvertently perpetuate. It begins with students who may have low preacademic skills as a consequence of little exposure to print material, and little verbal interaction with caregivers. Due to their lack of readiness skills, schoolwork is difficult for these students and academics may become aversive. Disruptive behaviors provide a means of escape from academic tasks and the disruptive student usually is sent out of the classroom (e.g., timeout, office referral, and suspension), which results in the student falling further behind academically. The cycle continues until the student is either expelled or drops out of school (Scott, Nelson, & Liaupsin, 2001).
Because poor academic skill is a strong predictor of school exclusion, examining the academic histories of expelled students could lead to strategies and programs aimed at reducing the risks for expulsion. Much of the research regarding high school students who have been expelled has focused on factors related to secondary education. For example, Morrison and D’Incau (1997) examined academic performance of expelled students the semester of expulsion and the semester before expulsion. They found that the average grade point average of the 158 students who were recommended for expulsion was 1.45 or a D+. However, as the failure cycle described above suggests, students’ outcomes in high school are built on the educational foundations they developed prior to high school.
Brown (2007) studied the academic, social, and emotional experiences of 29 expelled high school students through survey data. She found that most were not performing on grade level and their academic skills in reading, writing, and math were significantly underdeveloped. The students reported having had several absences from school, many of which were due to suspensions and expulsion, and that these absences caused them to fall behind in school.
McIntosh, Flannery, Sugai, Braun, and Cochrane (2008) conducted a study focused on the academic and behavioral pathways from middle school to high school to determine the significance of academic scores in eighth grade on problem behavior in ninth grade and the significance of problem behavior in eighth grade on the academic scores in ninth grade. The authors collected data from a district’s archival databases and examined the relationship between students’ grade point averages, a group-administered state reading test, and the number of office discipline referrals. As suspected, the results from their regression analyses showed that Grade 8 behavior predicted Grade 9 academic performance and Grade 8 academic scores predicted Grade 9 behavior. These researchers found that 18% of the sample exhibited challenges in academics and not behavior, while only 5% had challenges in behavior and not academics. They concluded that although low academic skills often adversely affect social behavior, problem behaviors almost always impede academic achievement. These authors also noted that 35% of the 330 students in the study needed additional support in either academics, behavior, or both in ninth grade.
To understand how these students reach high school academically behind their peers and at risk for expulsion, we need to look more closely at the cumulative nature of risk factors for school failure. Risk factors may be found in every life domain (individual, family, school, community, and peer group) and they may interact to effect student outcomes adversely. These factors include disability, adult responsibilities, low achievement, retention, poor attendance, family disruption, family mobility, low socioeconomic status, school discipline policies and practices, high-crime neighborhoods, and high-risk peer groups (Hammond, Linton, Smink, & Drew, 2007). The number, types, duration, timing, and severity of risks create a spiral of decline that increases in intensity. For example, a child from a single-parent family with low income may experience poor nutrition, poor family management practices, and high residential mobility (Hawkins et al., 2000). These factors will likely negatively affect the child’s school experiences.
Schools are generally thought of as places where children are cared for, supported, and nurtured that may offer protection against risk factors. However, researchers have documented contextual risk factors for school failure that may be perpetrated by school personnel. For example, Christle et al. (2005) found that teacher behaviors are influential on student outcomes. Teachers’ instructional styles and use of academic curriculum that does not match the students’ ability level often leads to frustration and behavioral problems (Scott et al., 2001). Teachers’ interactions with students also affect success in school. Teachers tend to interact less often with disruptive students (Carr, Taylor, & Robinson, 1991; Gunter, Jack, DePaepe, Reed, & Harrison, 1994), and are more likely to exclude students with problem behavior from the classroom for disciplinary measures (Skiba & Peterson, 2000). The present study examined student outcomes based on the contextual factors of assessing and evaluating students’ academic achievement.
Teacher Grading Practices
One of a teacher’s most important tasks is to assess student learning to provide crucial information about student progress for teachers, students, and families. For example, information on student performance helps teachers to (a) adjust instruction to match student needs, (b) evaluate instructional effectiveness, (c) share information with students and families, and (d) make critical decisions such as referral to special programs or promotion. Information on progress helps students to recognize their strengths and needs. Moreover, knowing they will be evaluated also provides motivation for many students (Green & Johnson, 2010). Families consider information on their child’s performance important so that they may track their child’s progress and provide encouragement or assistance if needed. They also can make judgments about teacher and school effectiveness (Kuhs, Johnson, Agruso, & Monrad, 2001).
Unfortunately, this important task of assessing student learning is fraught with chronic problems. According to Marzano (2000), there are three major problems with general education teachers’ grading systems. First, teachers often include factors unrelated to academic achievement. Second, teachers weigh assignments differently, which creates inconsistency across classes and participants. Third, teachers emphasize single scores to reflect student performance rather than several scores or combinations of scores.
Researchers have examined the literature and conducted studies on the factors that teachers consider when determining grades. For example, McMillan, Myran, and Workman (2002) conducted a study of more than 900 elementary schoolteachers’ assessment and grading practices. They found that teachers considered a variety of factors other than academic performance when grading students, such as effort, participation, attitude, and ability level of the students. In her review of the literature on teacher grading practices, Brookhart (1994) concluded that teacher practices are often inconsistent with recommended practice, particularly in confounding achievement with effort.
McMillan et al. (2002) studied the differences in which teachers weigh assignments; the variation in how teachers emphasize different factors in grading; and the inconsistency across teachers within schools and across schools in determining grades. In their study using elementary teachers, they found considerable variation in the sample on the extent they used particular factors. For instance, 23% of the teachers reported not using student ability level at all, while 47% reported using this factor quite a bit, extensively, or completely. Brookhart (1994) determined from her literature review that teachers varied considerably in how they regard achievement and nonachievement factors in determining grades.
McMillan et al. (2002) also examined the variation in the number of assessments and types of assessments used by teachers when they are determining student grades. They found considerable variation in the types of assessments teachers used for grading purposes from teacher-made assessments to published assessments. They also found that the factors teachers use for grading vary greatly between schools and even more so between teachers within schools. These researchers posited that teachers’ grading practices are highly individualized and based more on teachers’ beliefs than any other factors. Brookhart (1994) found that teachers in different locations weighed dissimilar items for grading purposes. In one location, teachers considered nonachievement factors such as attendance, behavior, effort, and work habits, whereas in another location, teachers used class work, homework, oral presentations, and projects to determine grades. These inconsistencies make comparisons in academic performance using grades over time or with other students difficult at best.
In addition, researchers have found that teachers’ grading practices differ considerably between elementary and secondary school. Randall and Engelhard (2009) surveyed a sample of 234 teachers and found that elementary schoolteachers assign higher grades than do secondary schoolteachers. Given the scenarios and results, the researchers concluded that middle school teachers were more likely to give lower grades to students exhibiting problem behaviors than were elementary schoolteachers. Thus, elementary and middle school teachers differed in the factors they considered when assigning students’ grade.
In this study, we considered the academic pathways of high school students who were recommended for expulsion. We conducted an archival study that examined the academic pathways in E/LA of high school students recommended for expulsion. We examined the available classroom letter grades and statewide assessment (SWA) scores across participants’ academic careers. We were interested in exploring the following research questions:
Research Question 1: Is there a consistent and predictable pattern in the E/LA performance measures over time for students recommended for expulsion?
Research Question 2: Are there differences between students’ scores on SWAs and the actual course performance grades in the E/LA?
Method
Participants
District demographics
Eighty-one students who attended schools within a large school district located in the southeastern United States were identified as participants in this study. The district comprised students from mixed socioeconomic backgrounds, with approximately 53% receiving free or reduced lunches. Caucasian students comprised 57% of the student body, and African American students were 39% of the student body. Students from Hispanic, Native American, and Asian/Pacific Islander accounted for the remaining 4% of the study body in the district. The district included schools in urban, suburban, and rural areas. A total of 37 schools served the district’s 27,378 students. Approximately 1.2% (n = 328) of the district’s students received either out-of-school suspensions and/or expulsions for violent or criminal offenses. Of the 328 students, a total of 126 students were recommended for expulsion from the school; however, due to missing data, 45 of the participants were not included in the data analysis. Based on district data, the percentage of students received either out-of-school suspensions and/or expulsions for violent or criminal offenses remained consistent over several years.
Participant demographics
Of the 81 students used for data analysis, 48% were 16 years of age, whereas 31% were 15, and the remaining students were either 14 (7%) or 17 (14%). Participants’ grade levels were 47% in 9th grade, 47% in 10th grade, and 6% in 11th grade. Gender demographics included 79% male and 21% female, while ethnically, 51% of participants were Caucasian and 49% were African American. Participants included those without disabilities (70%) and those with disabilities (30%).
All participants were recommended for and received expulsion hearings during the same academic year. Expulsion hearings were held for participants because of “consistent problem behavior” (53%), fighting (17%), and zero tolerance policy violations (29%; that is, drugs, weapon possession, and sexual acts). Expulsion hearing outcomes resulted in 66% of students being expelled from the district for the remainder of the school term, while 30% were placed at an alternative school, and 4% received homebound instruction. Prior to the current expulsion, 65% of the students in the study sample had zero expulsions, 25% had been expelled at least once, 6% had two prior expulsions, and 4% had three expulsions.
Measures
An archival search of students’ permanent records was used to collect the data for this study. Two primary sources of data were used: letter grades received and SWA results.
Letter grades
Most results of classroom assessments were based on the letter-grade system. This system used letter grades A to F to determine a student’s level of mastery of the assessed skill. For the purpose of the current study, a 10-point scale (i.e., 90–100 equals “A,” 80–89 equals “B,” and so on) was used to determine the associated letter grade. For example, if a student received a numerical grade of 85, this was coded as receiving a letter grade of “B.”
SWA results
The Palmetto Achievement Challenge Test was the SWA used in South Carolina public schools in third through eighth grades during the years that data were collected for the current study. SWA was used to measure student progress toward meeting state standards in the areas of E/LA, math, science and social studies. SWA testing occurred during late spring of each academic year.
Students received one of four possible performance-based scores. These scores are listed below with their explanations:
Below Basic—The student has not met minimum expectations [italics added] for student performance based on the curriculum standards.
Basic—The student has met minimum expectations [italics added] for student performance based on the curriculum standards.
Proficient—The student has met expectations [italics added] for student performance based on the curriculum standards.
Advanced—The student exceeded expectations [italics added] for student performance based on the curriculum standards.
For students to “meet state standards,” they were required to score basic or higher. These students were considered academically prepared for promotion to the next grade level because they exhibited adequate knowledge of the state standards.
Procedures
Upon receiving Internal Review Board and school district administrator approval to collect these extant data, we reviewed expulsion packets and academic records of the students who received expulsion hearings. Expulsion packets contained demographic information and reasons for the expulsion hearing. Student data were selected from the expulsion packet based on meeting at least one of two criteria. Students had to be either (a) in 9th or 10th grade or (b) age 15 or 16. These criteria were selected because the state allowed students 16 years and older to drop out of school. Furthermore, most students are in the 9th or 10th grade when they reach 16 years of age. A total of 126 students met either of these criteria; however, 45 students were not included in data analysis because of the absence of at least 70% of their academic records due to transfer between districts or states. Due to missing data, it was determined that including these students in data analyses may not present an accurate depiction of student performance.
Following the selection of the 81 eligible participants, researchers gathered academic performance data from permanent academic records. Data were entered directly into a statistical analysis program and included the following academic indicators: (a) E/LA letter grades from 1st grade through the year of expulsion and (b) E/LA SWA scores from third through eighth grade.
Statistical Analyses
Statistical analyses were conducted using the Statistical Analysis Package for the Social Scientist (SPSS) v. 16.0. To answer the first research question, several analyses were conducted. First, mean and standard deviations of scores were determined to show average scores over time. Second, a one-sample t test was used to determine whether received letter grades and SWA scores varied significantly from the average score. For this analysis, a letter grade of “C” was chosen because it generally correlates to “average” performance. SWA score of “Basic” was chosen, because it is considered to show that students have met minimum standards and should be promoted to the next grade level.
Determining whether significant differences exist between E/LA achievements due to expulsion reason, expulsions were dichotomized into two categories, and an independent samples t test was conducted. The first category included participants expelled for consistent problem behavior, while the second category included participants expelled for “zero tolerance” policy violations (e.g., sexual acts, weapon violations, drugs, and fighting). Finally, to determine whether the number of expulsions was affected by E/LA performance, the number of expulsions participants had received was dichotomized into two groups: (a) no previous expulsions and (b) previous expulsions. An independent samples t test was conducted to determine whether significant differences existed.
A paired samples t test was conducted for third- through eighth-grade E/LA letter grade and SWA scores to address the second research question. Comparison of letter grades, with five possible scores, and SWA scores with four possible scores, required that letter grades be recoded into four possible scores. Letter grades of D and F were combined into one, a common score, because these two letter grades usually correspond to students not meeting the standard set in the class, which highly correlates to the SWA score of “Below Basic.”
Participants were assigned into dichotomous groups based on disability. An independent samples t test was conducted to determine whether significant differences existed between participants with and without disabilities in SWA and letter grades. Finally, ethnic and gender differences in E/LA performance was determined by using an independent samples t test. It is important to note that all t tests have degrees of freedom lower than 81 due to individual students missing one or more pieces of data (i.e., specific grade or result).
Results
Is There a Consistent and Predictable Pattern in the E/LA Performance Measures Over Time for Students Recommended for Expulsion?
Figure 1 presents SWA results based on percentage of students receiving each score. Examination of Figure 1 shows that no participant received an “Advance” score on the SWA. The score of “Below Basic” made up one third to almost one half of scores received during each administration of the SWA, which suggests these participants were not academically prepared to move to the next grade level. Letter grades earned by school level (e.g., elementary, middle, and high) are presented in Figure 2. A trend line was added to show the general downward trend in the letter grade “A” and the upward trend in the combined letter grades “D/F” between primary grades (i.e., Grades 1–2) and high school (i.e., Grades 9–10). Higher letter grades were positively skewed during primary and elementary (i.e., Grades 3–5) and negatively skewed in high school. Letter grades received in middle school represented an even grade distribution.
SWA scores on E/LA assessment—third through eighth grades.
Letter grades earned by school level.
A group of one-sample t tests was conducted based on SWA and letter grade results, with a test value of “2” or “Basic” and “3” or “C,” respectively. These values were chosen because they are considered to be average scores. Results shows that the entire sample of participants scored significantly lower (p < .01) than average on the SWA in Grades 5 to 8; however, students with disabilities scored significantly lower (p < .05) than average in each administration of the SWA. An interesting finding is that students without disabilities scored significantly lower than average during the transition years of fifth (t = −2.51, p < .01) and eighth grade (t = −2.73, p < .01).
Results of the t test measuring letter-grade achievement returned the entire sample and participants without disabilities scoring significantly higher than average during primary (p < .001) and elementary (p < .01) and significantly below average during high school (p < .01). Middle school resulted in a slightly below average finding; however, statistical significance was not attained. Participants without disabilities experienced diminishing success as they progressed through school. For students with disabilities, the only significant t test results were found in ninth grade (t = −2.45, p < .05) suggesting that these participants earned near average (e.g., “C”) letter grades representing a more consistent classroom performance over time.
An independent samples t test was conducted between students expelled for consistent problem behaviors and students expelled for “zero tolerance” violations. Results indicated that there were no significant differences in the scores for SWA or letter grades based on expulsion reason. Independent samples t tests were conducted to determine the effects of prior expulsions on achievement. Results show no significant differences in letter grades or SWA testing.
Are There Differences Between Students’ Scores on SWAs and the Actual Course Performance Grades in the E/LA?
Paired samples t tests were conducted between recoded letter grades and SWA results in Grades 3 to 8. As previously stated, the letter grades of “D” and “F” were recoded into one score for comparison to the four scores received from SWA results. Results of the t test returned significant differences (p < .05) in all grades except Grade 7 (p = .07). In each instance of statistical significance (p < .05), students received a higher letter grade score than SWA score. This finding indicates that content mastery may not have generalized between classroom performance (i.e., letter grades) and performance on SWA tests.
Results of the independent samples t test measuring effects of disability on E/LA performance found significant differences (p < .05) between students with and without disabilities on E/LA performance. Students without disabilities received significantly higher scores (p < .05) for classroom performance (i.e., letter grades) from Grades 1 through 5. During Grades 6 through 10, there were no significant differences found. SWA results showed that participants without disabilities scored significantly higher (p < .05) in Grades 3 through 7.
A series of independent samples t test was conducted to determine mean differences in E/LA performance between participants from the two primary ethnic backgrounds represented in the study. Classroom performance, as determined by letter grades, showed significant differences in second grade, t(45) = 3.16, p < .01, and third grade, t(49) = 2.52, p <. 05, with Caucasian participants receiving higher reported grades. E/LA achievement on SWA showed significant differences in scores in third grade, t(45) = 2.38, p < .05; fourth grade, t(48) = 2.48, p < .05; fifth grade, t(58) = 2.16, p < .05; and sixth grade, t(61) = 3.24, p < .01. In each instance, Caucasian participants performed significantly higher than African American participants. Analyses to determine gender differences returned no significant differences on either assessment of E/LA.
Discussion
The purpose of this study was to examine the academic pathways of students recommended for expulsion by answering two research questions. First we sought to determine whether there were consistent and predictable patterns in the E/LA performance measures over time for students recommended for expulsion. Results of this study affirmed that students recommended for expulsion did perform in a consistent and predictable pattern. As presented in Figure 2, as students moved into higher grades, their letter grade scores consistently decreased. The trend in SWA scores shows a similar pattern, with 38% of the participants scoring Proficient or Advanced in third grade to only 7% scoring Proficient or Advanced in eighth grade. As discussed previously, early academic failure has been identified as a risk factor for school failure and delinquency. These outcomes pose detrimental effects on the individual and place a tremendous burden on society.
The results of this study reinforce Brunner’s (1993) conclusion regarding reading failure, frustration, and the snowball effect that occurs from the end of elementary school, increasing frustration and vulnerability to later risks such as maladaptive and aggressive behaviors. Whereas most studies of high school students who have been expelled have focused on factors related to secondary education, the current study extends the literature by looking back to elementary school for patterns in the academic pathways of students who were eventually recommended for expulsion. Other researchers have discovered predictable and persistent academic pathways of students at risk for school failure. For example, Hickman, Bartholomew, Mathwig, and Heinrich (2008) demonstrated that dropouts’ academic achievement levels were lower than graduates’ levels and as early as kindergarten. The differences in course performance grades and standardized test scores persisted over time regardless of participant. Researching academic pathways of students at risk for school exclusion and school failure holds great promise for informing educators of policies and practices that may not benefit students, as well as those policies and practices that may greatly benefit students during times of heightened risks.
Another finding from this study was that students scored significantly lower than average on both E/LA measures during the transition years of fifth and eighth grade. Other researchers have found that academic performance declines as students transition to high school. For instance, McIntosh et al. (2008) found that 35% of the students in their sample needed additional support in academics, behavior, or both when transitioning into ninth grade. These researchers suggested that school personnel need to assess and support the academic and behavioral needs of students as they start high school. Indeed, school personnel need to assess and support students during critical transition periods such as into kindergarten, in later elementary school as coursework is increased to include science and social studies, into middle school, and into high school.
The second research question asked if significant differences existed between students’ scores on SWAs and the actual course performance grades in E/LA. When SWA and course performance grades were compared, it is evident that there are disparities between SWA performance and course grades given by teachers.
To be sure, assessing student learning is one of a teacher’s most important tasks. Why would this important task of assessing student learning be fraught with chronic problems such as including factors unrelated to academic achievement and weighing assignments inconsistently across classes and participants? Why would teacher-given grades deviate so greatly from SWA scores, particularly at the elementary level? Researchers and experts in assessment have offered several reasons for the disparities in teacher grading practices. For example, Guskey and Bailey (2001) describe chronic issues with grading systems for general education teachers. One issue is that preservice programs provide minimal preparation in how to develop grading systems. Another issue is that teachers see grading as burdensome, time-consuming, and conflicting with their roles as teachers. For example, Green and Johnson (2010) discussed the struggle teachers face with evaluating and judging students while trying to advocate and nurture them. In addition, the authors pointed out that teachers who are strong advocates may be more concerned about the powerful effect bad grades can have on students’ self-perception and thus give better grades than the performance warrants. Grades are typically left up to teacher discretion and are judgments based on subjective decisions such as which elements to count and how to weigh various elements. Most districts do not provide specific grading policies to address these issues. As a result, discrepancies between teacher-given grades and SWA scores will continue to confound researchers and confuse students and their parents.
Limitations and Future Research
This study has three major limitations that readers should note when interpreting the results. First, participants were selected from one district, which limits the ability to generalize the results to other geographical locations and the broader U.S. school population. It is important for future research to include multiple districts’ data and perhaps even multiple states’ data. These data would improve the validity of results for the present study and provide a clearer measure of achievement for students who receive expulsions. Second, in collecting letter-grade data, the use of numerical grades, as opposed to categorical, would have provided more variance within the results. This variance may have better explained the results. Future research in this area should attempt to use categorical (i.e., letter grades) and numerical data for classroom-related academic performance. Third, due to missing data, some of the statistical analyses did not take into account the full sample size. Further investigations should attempt to account for 100% of each participant’s data. Fourth, the current study did not utilize a control group to compare letter grades and SWA results. Comparisons between students not receiving expulsion hearings and those who do would add heft to the results.
Implications and Recommendations
This study raises concern for the more than two thirds of 8th graders and 12th graders that score lower than proficiency for reading (Perie, Grigg, & Donahue, 2005). Efforts to address these concerns have led to a growing body of research regarding literacy teaching and learning in adolescence. Because lifelong consequences are experienced related to academic difficulties in E/LA, it is imperative that these issues be addressed for students on the path to expulsion from school. Because the literature is clear in identifying the problematic academic trajectory of students having behavioral challenges, and because the educational response has not adequately addressed the academic needs of these students, it is urgent that more information be known regarding the details of the journey toward the academic destruction, that is, expulsion from school. Based on this study, we make the following recommendations regarding the academic and behavioral policies and practices that may result in students being expelled from school.
We recommend that school personnel make unified efforts to address students’ deficits in academics and behavior early and throughout their school careers and provide remediation instead of just passing them on to the next grade. There are specific initiatives for struggling adolescent readers such as Adolescent Literacy, an educational initiative that includes a national multimedia project (www.AdLit.org) offering information and resources to the parents and educators of struggling adolescent readers and writers. There also are evidenced-based interventions for students who need increased behavioral support at any grade level, such as the Behavior Education Program (Crone, Hawken, & Horner, 2010), which incorporates daily behavioral feedback, positive adult attention, and increased home–school collaboration.
We also conclude that there is a dire need for teachers to follow established measurement principles and to evaluate the quality of their own assessments as well as those provided by others. The confounding of achievement elements with behavioral elements and the inconsistency of grading practices contribute to the confusion regarding students’ level of academic achievement, especially when teacher-given grades do not match SWA scores. We recommend that teachers consult resources, such as Green and Johnson (2010) in which the authors provide guidelines and examples for fair and accurate grading decisions. They provide a list of best practices for making grading decisions:
1. Learn and follow your school district’s grading policy. 2. Design a written policy in which you base grades on summative assessments that address your learning goals derived from the standards. 3. Ensure quality of assessments by having sufficient information (reliability) and by using different types of assessments that address the learning goals (validity). 4. Communicate with and involve students in developing standards of performance for graded assessments. 5. Allow recent and consistent performance to carry more weight than strict averages. 6. Avoid unduly weighing factors unrelated to a student’s mastery of the learning goals, such as the following: effort, participation, attitude or behavior, aptitude or ability, improvement, extra credit, late work, zeroes, group grades. 7. Avoid using grades as rewards and punishments. 8. Review borderline cases carefully. (Green & Johnson, 2010, p. 295)
We recommend that preservice teacher education programs provide enhanced preparation in how to develop grading systems. We also recommend that school districts provide specific grading policies to address these issues and monitor teachers to increase consistency in teacher grading practices within and across schools, and in comparison with SWA scores. In addition, we recommend that preservice programs collaborate with local school districts to ensure consistency in grading policies and practices.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.