Evaluating the Effect of No Child Left Behind on U.S. Music Course Enrollments

Abstract

The purpose of this study was to investigate nationwide enrollment in high school music courses from 1982 until 2009 to determine what trends in music enrollment existed and whether these trends were affected by the passage and implementation of the No Child Left Behind Act of 2001 (NCLB). With data from 10 separate nationally representative high school transcript studies conducted by the National Center for Education Statistics, a unique data set was constructed that tracked the transcript-indicated 9th- through 12th-grade music course enrollment patterns for the U.S. graduating classes of 1982, 1987, 1990, 1992, 1994, 1998, 2000, 2004, 2005, and 2009. Descriptive results showed that overall music enrollment patterns were relatively stable in the public schools, with roughly 34% of all students consistently enrolling in at least one music course during high school across all cohorts. Abbreviated interrupted time series analyses suggest that NCLB had no effect on overall music enrollment rates but exacerbated the preexisting underrepresentation in music courses of Hispanic students, English language learners, and students with Individualized Education Plans.

Keywords

No Child Left Behind education policy music enrollment demographics

The No Child Left Behind Act of 2001 (NCLB) generally is regarded as the strongest U.S. federal regulation of the nation’s education policy, an area that historically has been the provenance of the 50 states. In the ensuing decade between its enactment on January 7, 2002, and the effective end of the NCLB era when the U.S. secretary of education began granting waivers from the law to states that adopted other school reform policies on February 9, 2012 (Hu, 2012; Rich, 2012), the law’s intents and effects have been both championed and criticized in the popular press (e.g., Dillon, 2006), in think-tank reports (e.g., McMurrer, 2007; Peterson & West, 2003), and in academic position papers (e.g., Beveridge, 2009; Chapman, 2004, 2007; Colwell, 2005). The scientific record also includes several evaluations of NCLB’s effect on student achievement (e.g., Dee & Jacob, 2011; Lauen & Gaddis, 2012; Lee, 2006; Lee & Reeves, 2012; Wong, Cook, & Steiner, 2009b), although the results from these are generally mixed. Music educators generally have reported that NCLB has had deleterious implications for music education (Beveridge, 2009; West, 2012); however, few researchers within the field have investigated empirically the impact of NCLB on music education (Abril & Gault, 2006, 2008; Gerrity, 2009; West, 2012), and there have been no large-scale studies of the effect of NCLB on music education outcomes. The present study was designed to assess what effect, if any, NCLB had on the elective study of music in U.S. high schools.

NCLB was rationalized by a twofold theory of action: (a) that schools held accountable for student achievement based on the risk of consequential sanctions are incented to raise student achievement (Dee & Jacob, 2011) and (b) that the disaggregation of student achievement data by subgroup would ameliorate long-standing achievement gaps between students of varying race and socioeconomic status (SES) by incenting schools to raise achievement for all students rather than just by raising the school’s average achievement (Lauen & Gaddis, 2012). There are five mandates in NCLB through which this theory of action was implemented. The states were mandated to (a) develop rigorous standardized tests aligned with the state’s own content area standards and regularly administer them to students; (b) develop a sequential plan in which each year a greater percentage of students, overall and within targeted subgroups, would achieve a state-determined proficiency benchmark on the standardized tests until 100% of students achieve proficiency in reading and math in 2014; (c) penalize schools that fail to meet the yearly required percentages with consequences of increasing severity; (d) implement teaching strategies based on scientific studies of their efficacy or effectiveness; and (e) require all teachers to meet a state-defined standard of “highly qualified” (Wong et al., 2009b). Ultimately, these mandates were intended to improve scores on the state-developed and state-scored standardized tests, thus bestowing the results of state tests with “high stakes.” That states were permitted to design their own tests, and to set their own passing scores for achieving proficiency, did, in practice, produce state-level variation in the difficulty of achieving NCLB goals. The rigor of the various state assessments can be indexed against the high-quality-but-low-stakes National Assessment of Educational Progress (NAEP, commonly referred to as “The Nation’s Report Card”), and several researchers have exploited this variation to evaluate NCLB by comparing high-rigor to low-rigor states (Dee & Jacob, 2011; Wong et al., 2009b).

Studies of the effects of NCLB on standardized test scores have yielded mixed results. Investigating the effect of NCLB on fourth- and eighth-grade test scores with NAEP data from 1990 to 2005, Lee (2006) concluded that reading scores were unaffected by NCLB and that observed gains in NCLB-era math scores were the continuation of a pre-NCLB secular trend. Wong and colleagues (2009b), using NAEP data extending to 2009, found small positive effects for both fourth- and eighth-grade math scores but limited positive effects for fourth-grade reading scores. With the same years of NAEP data as Wong et al. (2009b), but using slightly different treatment and comparison group contrasts, Dee and Jacob (2011) also found small but significant positive effects of NCLB on NAEP scores in fourth- and eighth-grade math. However, and contradictory to Wong et al. (2009b), Dee and Jacob found no effect of NCLB on fourth-grade reading scores. Exploiting the variation in state-to-state implementation of the federal law, Lee and Reeves (2012) concluded that observed achievement gains nationwide in math during the NCLB era were a function of long-term state instructional capacity and teacher resources unrelated to the law, as the varying fidelity of NCLB implementation among the states was not associated with the achievement score gains.

Ostensibly, NCLB was designed to raise overall student educational achievement and to eliminate achievement gaps among varying subgroups of students. These intended outcomes have never been particularly controversial; however, the unintended consequences of the law, engendered by its enforcement mechanisms and the incentives it created to boost reading and math test scores by any means possible, have stirred debate among in-service teachers, school administrators, policymakers, and the educational research community (Peterson & West, 2003). The severity of the corrective actions imposed on schools under the law resulted in school efforts to raise test scores in ways never intended by Congress. For example, researchers have found that under high-stakes testing accountability policies, schools were more likely to suspend students with poor prior test scores on testing days than on other days of the year (Figlio, 2006), to classify greater numbers of students as in need of special education services (Jacob, 2005), and to increase markedly the caloric content of school-provided lunches on testing days (Figlio & Winicki, 2005). The incentives to game the system extend beyond school administrators and into classrooms as well: Jacob and Levitt (2003) found that high-stakes accountability pressure on teachers also can result in incidents of wholesale teacher cheating on the exams. The extremely narrow focus on math and reading test scores as arbiters of accountability gave rise to vehement criticism of NCLB from arts and music educators, as well as scholars from other subfields within education, that the nation’s school curricula were narrowing (Au, 2007; Beveridge, 2009; Chapman, 2004; Heilig, Cole, & Aguilar, 2010; Ravitch, 2011). Music educators, in particular, feared that the narrowing of the curriculum toward tested subjects would marginalize whatever eminence may have been gained in NCLB’s reaffirmation of “core” status for the arts.

Empirical investigations of NCLB within the field of music education generally have been limited to surveys of principals (Abril & Gault, 2006, 2008; Gerrity, 2009) and qualitative interviews of in-service music teachers (Spohn, 2008; West, 2012). Abril and Gault found that elementary (Abril & Gault, 2006) and secondary (Abril & Gault, 2008) principals who responded to their national surveys believed that NCLB had had a negative effect on their music programs. In McMurrer’s (2007) national survey of school districts, 16% of administrators responding reported that they had decreased instructional time for music in response to NCLB mandates. In a survey of 179 principals from within the state of Ohio, Gerrity (2009) found 43% reported conditions arising from NCLB that Gerrity classified as having “weakened” the music program. Spohn (2008) interviewed arts educators in a rural Ohio district that had failed to meet the adequate yearly progress benchmark. Participants reported that the district had reduced music course requirements by restricting general music course enrollment only to those students who were not enrolled in band or choir; the NCLB-inspired policy change resulted in a de facto reduction in the number of music course sections offered in the district. In a qualitative study of music teachers in Michigan schools that had failed to meet the NCLB adequate yearly progress benchmark (West, 2012), many participants reported scheduling changes that made it more difficult for students to elect music courses and that expectations for music teachers to integrate math and reading concepts into their classes had increased.

West’s (2012) finding that electing to pursue music course work may have become more difficult is one key, empirically verifiable way in which negative effects of NCLB might manifest themselves in the nation’s schools. If schools under accountability pressure go to great lengths, such as altering school lunch menus, to boost test scores (Figlio & Winicki, 2005), policy changes that subtly increase the difficulty of students to pursue electives seem relatively minor undertakings in comparison. Elpus and Abril (2011) constructed a national demographic profile of 12th-grade high school music ensemble students in 2004 using high-quality data from the National Center for Education Statistics (NCES). They found that music ensemble participation was associated with race/ethnicity (white students were overrepresented among music students, while Hispanic students were underrepresented), SES (students from higher socioeconomic backgrounds were overrepresented among music students), prior academic achievement (high achievers were overrepresented among music students), and parental educational attainment (students whose parents had postgraduate degrees were overrepresented among music students, while children of parents who had not achieved a high school diploma were underrepresented). Elpus and Abril examined only one cohort at only one point in time during the NCLB era. Their finding that Hispanic students and lower-SES students were less likely to participate in music bears further investigation, particularly since Hispanic students and low-SES students are two of the target subgroups identified in the law’s disaggregation provisions. Comparing their self-reported enrollment estimate (21% of students were in band, choir, or orchestra) to an earlier estimate of 30.6% (Stewart, 1991) using transcript data, Elpus and Abril noted the decline but cautioned that two data points cannot reveal a trend and that measurement differences in the self-report and transcript data used to categorize music students may invalidate the comparison.

Purpose of the Study and Research Questions

The purpose of the present study was to investigate nationwide enrollment in high school music courses from 1982 until 2009 to determine what trends in music enrollment existed and whether these trends were affected by the passage and implementation of NCLB. The following research questions guided the inquiry: (1) From 1982 until 2009, what percentage of high school students were enrolled in music courses, in aggregate and in the public school subgroups targeted by the NCLB disaggregated accountability provisions? (2) What was the overall effect of NCLB on enrollment rates and trends in public high school music courses? and (3) What was the effect of NCLB on public high school music enrollment rates and trends among the targeted subgroups?

Data, Research Design, and Methods

Data

Sources

For the present study, I created a unique data set that tracks nationwide enrollment rates in high school music courses for the nearly three-decade period between 1982 and 2009. I constructed this data set by analyzing the music course enrollment data from 10 separate large-scale high school transcript studies conducted by the U.S. Department of Education’s NCES. Periodically, NCES collects complete high school transcripts, reflecting course work from all years of high school attendance, from large, nationally representative samples of high school students as part of the longitudinal education studies and as ancillary data for the NAEP. Cohorts included in the present study graduated high school as part of the classes of 1982, 1987, 1990, 1992, 1994, 1998, 2000, 2004, 2005, and 2009. Data from the cohorts of 1982, 1992, and 2004 were drawn from the Interuniversity Consortium for Political and Social Research data set of the longitudinal High School and Beyond (HS&B) study and from the NCES restricted-use data sets of the National Education Longitudinal Study of 1988 (NELS) and the Education Longitudinal Study of 2002 (ELS), respectively. Data from the remaining cohorts (1987, 1990, 1994, 1998, 2000, 2005, and 2009) were drawn from the restricted-use data sets of the High School Transcript Studies (HSTS), which are ancillary data collections associated with the NAEP. Each of these transcript studies, whether from NAEP HSTS or the longitudinal education studies, followed the same stringent collection procedures (Alt & Bradby, 1999), ensuring nationally representative data were collected at each time point and that the transcript data are comparable across time.

Preparation

I conducted extensive data preparation for each NCES study included in the construction of the analytic data set. For all sources except for HS&B, the data preparation included the manual verification of each course coded as a music class or those courses coded with the omnibus visual and performing arts codes (HSTS87,n [of courses] = 27,900; HSTS90, n = 19,370; NELS, n = 23,880; HSTS94, n = 37,540; HSTS98, n = 35,380; HSTS00, n = 33,240; ELS, n = 19,920; HSTS05, n = 45,810; HSTS09, n = 63,250). Verbatim course titles were not available in the HS&B (n [of courses] = 13,180) data set, preventing the manual verification step for that data source.

Empirical Approach

Research Question 1

To answer the first research question, I estimated the percentage of each cohort who had enrolled in at least one music course at some point in high school, with a 95% confidence interval, both for the entire population and for the various subgroups targeted by NCLB. I also estimated the percentages of each cohort who had enrolled in music at each grade level and who had earned credit for 2, 3, or 4+ years of music throughout high school. I examined potential associations between music enrollment and students’ race/ethnicity subgroup statuses at each time point using Rao-Scott adjusted χ².

Research Questions 2 and 3

I evaluated the effect of NCLB on high school music enrollment using an abbreviated (or “short”) interrupted time series design (Cook & Wong, 2008; Shadish, Cook, & Campbell, 2002; Wong, Cook, & Steiner, 2009a). This is the primary approach that has been used in the prior literature evaluating the effect of NCLB on standardized test scores (Dee & Jacob, 2011; Lee, 2006; Lee & Reeves, 2012; Wong, 2008; Wong et al., 2009b), and is a quasiexperimental approach designed specifically for estimating causal effects using repeated measures of the same variables over time. In the abbreviated interrupted time series design, causal effects are analyzed as either immediate changes in the level of the outcome following the initiation of the policy (a change in regression intercept following the policy change) or a difference in the variable’s rate of change over time (a change in the regression slope following the policy change). These causal effects are compared against an implied counterfactual—namely, the uninterrupted continuation of the preintervention trend or regression line. The primary threat to causal interpretation (internal validity) in abbreviated interrupted time series designs is the history threat—the possibility that an event external to the intervention occurred simultaneously with the intervention and affected the outcome of interest in ways that cast doubt on the causal interpretation of the study’s results (Shadish et al., 2002). Other possible threats to validity include the instrumentation threat (that the measures change over time) and statistical regression (that recent, sudden changes in the outcome would regress to the mean over time regardless of an intervention).

The history threat to internal validity can be ameliorated considerably if additional time series data are available to serve either as a counterfactual or as a nonequivalent comparison time series (Cook & Wong, 2008; Wong et al., 2009a). Possible choices include a nonequivalent comparison group that was not subject to the policy change but would be subject to similar history threats to validity or a nonequivalent dependent variable observable for the same individuals that would not be affected by the policy change (Shadish et al., 2002). Research methodologists generally contend that a suitable nonequivalent comparison group considerably strengthens the warrant for a causal claim from abbreviated interrupted time series designs but is not necessary provided certain conditions hold (Cook & Wong, 2008; Wong et al., 2009a). When a comparison time series is possible, causal effects are estimated as the difference-in-differences of regression intercepts or slopes between the groups, which is the approach followed in this study for Research Question 2.

The existing literature evaluating NCLB has compared public schools to private schools (Wong et al., 2009b), specifically, the Catholic schools, or has compared “high-dose” NCLB states to “low-dose” NCLB states categorized either by the rigor of the state tests relative to the NAEP framework (Wong et al., 2009b) or by the state’s adoption of consequential test-based accountability for schools prior to the passage of NCLB (Dee & Jacob, 2011; Lee, 2006; Lee & Reeves, 2012). In these comparisons, high-rigor states and those states that previously had not enacted consequential school accountability laws were considered to have received a “high dose” of NCLB. For the present investigation, my choice of comparison groups was necessarily limited by the availability and structure of the transcript study data—the data are strictly nationally representative and are not representative of the constituent states, making state-level contrasts using these data impossible. I was, however, able to compare national music enrollment trends in the public schools, which were subject to NCLB, to music enrollment trends in the private schools, which were not. This is one approach used by Wong et al. (2009a, 2009b).

Importantly, the public school/private school comparison potentially does introduce a specific history threat to the internal validity of time series studies of NCLB: The first news media reports of the Catholic child abuse scandal broke in January 2002, at virtually the same time as NCLB was signed into law. It is possible that students moved away from the Catholic schools simultaneously with the implementation of NCLB, a history threat to the validity of the causal comparison if withdrawal from Catholic schools was differential among groups of students with varying likelihoods of enrolling in music. For example, if students likely to enroll in music defected from the Catholic schools to the public schools in great numbers, a hypothetical NCLB-caused decline in public school music enrollment concomitant with the enrollment shift might be masked in the enrollment data. Wong et al. (2009b) and Dee and Jacob (2011) analyzed enrollment and attrition from the Catholic schools around the implementation of NCLB. Their analyses suggest that a preexisting declining trend of Catholic school enrollment accelerated beginning in 2002 but that this preexisting secular trend was not differentially accelerated among racial and ethnic groups. Differential acceleration of attrition from the Catholic schools among the various racial and ethnic groups would be a concern for the present study, given that the various racial and ethnic groups have been found to exhibit varying likelihoods of enrolling in music (Elpus & Abril, 2011).

Dee and Jacob (2011) did, however, report on one key demographic difference possibly spurred by the scandal that needs to be considered: parental educational attainment. After 2002, Dee and Jacob documented a relative increase in the parental educational attainment for students remaining in the Catholic schools. Although Elpus and Abril (2011) found that parental educational attainment was associated with school music participation, it was parents of higher educational attainment whose children were more likely to participate in music. The nature of the demographic shift in the Catholic schools ameliorates the history threat to the validity of the present study. Because NCLB is hypothesized in the existing literature to have affected public school music enrollment negatively (e.g., Beveridge, 2009; Chapman, 2004), an increase in the relative parental educational attainment of Catholic school students would not mask a decline in public school music enrollment, because the students who would be more likely to take music would have remained in the Catholic schools.

Unfortunately, the various racial and ethnic subgroups in the private schools were not represented sufficiently in the transcript study samples to draw meaningful public school/private school comparisons of the NCLB-targeted subgroups. Consequently, the estimates for Research Question 3 contain no comparison time series, and the estimated counterfactual is the continuation of the pre-NCLB trend in the public schools. While the lack of a comparison group reduces the statistical power to detect significant effects in the analyses for Research Question 3, potential changes in intercept or slope due to NCLB among subgroup music enrollment remain estimable, and those results are reported here. The loss of statistical power in these analyses makes causal estimates for Research Question 3 more conservative: To be detected as significant in this analysis, effects would need to be both larger and more proximate to the enactment of NCLB.

The other threats to the validity of abbreviated interrupted time series designs likely are not an issue for this study. Instrumentation for all the transcript studies followed a similar protocol and underwent no major change across the 10 underlying data sources, reducing or eliminating that threat to validity. NCLB was not introduced in response to enrollment rates in music, and so the statistical regression threat to validity is likely not of concern in the present investigation.

Estimation of the Analytic Models

Results for Research Questions 2 and 3 were obtained using a difference-in-differences approach. The analytic models, and their estimation through weighted least squares regression, are described fully in the online Appendix to this article, available at http://jrme.sagepub.com/supplemental.

Results

Research Question 1

Table 1 presents the estimates of nationwide enrollment percentages in music in both the public and the private schools. The top row lists the percentages of students in each cohort who enrolled in at least one of any kind of music course—performance or nonperformance—at any point during high school. Because each estimate is derived from a separate cross-sectional complex sample, the percentages are reported with appropriate 95% confidence intervals. Data presented in the top row of the table provide evidence that students’ election of taking at least one music course at some point during high school remained remarkably stable across the time period examined in this study. Although enrollment in at least one music course peaked at 37.54% of all high school students for those in the graduating class of 1994, across all 10 cohorts, just over one third of all U.S. high school students enrolled in a music course at some point during high school.

Table 1.

Nationwide Music Enrollment Percentages 1982–2009, with 95% Confidence Intervals.

	Cohort graduation year
Variable	1982	1987	1990	1992	1994	1998	2000	2004	2005	2009
Any music	34.08 [32.5, 35.7]	32.30 [29.8, 34.8]	35.28 [32.9, 37.6]	33.21 [31.3, 35.1]	37.53 [35.5, 39.6]	35.36 [32.0, 38.7]	34.92 [32.9, 37.0]	35.48 [33.8, 37.2]	33.87 [32.6, 35.1]	34.43 [33.1, 35.8]
2+ years	13.83 [12.8, 14.8]	14.21 [13.1, 15.3]	16.15 [14.6, 17.7]	16.52 [15.2, 17.8]	18.72 [17.4, 20.1]	18.90 [14.8, 23.0]	17.64 [16.4, 18.9]	16.83 [15.7, 17.9]	17.52 [16.8, 18.2]	17.38 [16.5, 18.2]
3+ years	9.12 [8.3, 9.9]	9.50 [8.5, 10.5]	10.79 [9.4, 12.2]	11.22 [10.3, 12.2]	12.90 [11.7, 14.1]	13.48 [10.7, 16.3]	12.46 [11.3, 13.6]	11.80 [10.9, 12.7]	12.72 [12.1, 13.4]	12.52 [11.8, 13.1]
4+ years	5.42 [4.8, 6.0]	5.86 [5.1, 6.7]	7.05 [6.0, 8.0]	7.59 [6.8, 8.3]	8.42 [7.5, 9.3]	9.57 [7.3, 11.8]	8.74 [7.8, 9.7]	7.80 [7.0, 8.6]	9.11 [8.5, 9.8]	9.43 [8.8, 10.0]
9th grade	23.34 [22.0, 24.7]	22.98 [20.9, 25.1]	24.70 [22.5, 27.0]	23.37 [21.7, 25.0]	26.71 [24.9, 29.0]	25.74 [20.8, 30.7]	24.29 [22.5, 26.1]	23.77 [22.4, 25.1]	23.44 [22.4, 24.5]	23.02 [21.8, 24.3]
10th grade	18.76 [17.6, 19.9]	17.67 [16.1, 19.2]	19.58 [17.7, 21.4]	17.79 [16.5, 19.1]	21.73 [20.2, 23.3]	22.22 [17.5, 27.0]	20.46 [19.0, 21.9]	20.34 [19.2, 21.5]	19.88 [18.7, 20.9]	19.44 [18.6, 20.3]
11th grade	15.58 [14.6, 16.6]	15.27 [14.0,16.5]	17.21 [15.6, 18.9]	15.51 [14.3, 16.7]	18.91 [17.6, 20.2]	19.64 [15.4, 23.9]	17.39 [16.1, 18.3]	17.37 [16.3, 17.9]	17.17 [16.4, 17.9]	17.68 [16.8, 18.5]
12th grade	13.06 [12.1, 14.0]	12.52 [11.6, 13.4]	16.15 [14.8, 17.5]	14.09 [12.9, 15.2]	18.31 [17.2, 19.4]	18.17 [15.0, 21.3]	16.56 [15.5, 17.6]	15.49 [14.4, 16.5]	17.01 [16.2, 17.8]	16.84 [16.2, 17.5]
Choir	15.43 [14.3, 16.6]	15.45 [13.5, 17.4]	16.53 [14.8, 18.3]	16.32 [15.0, 17.6]	17.95 [16.2, 19.7]	17.84 [14.2, 21.4]	16.49 [14.9, 18.1]	16.78 [15.6, 17.9]	15.67 [14.9, 16.4]	14.61 [13.7, 15.5]
Band	13.74 [12.8, 14.7]	13.63 [12.5, 14.8]	14.21 [13.0, 15.4]	13.68 [12.2, 15.1]	15.69 [14.7, 16.7]	15.46 [12.7, 18.3]	13.63 [12.7, 14.6]	12.40 [11.5, 13.3]	12.93 [12.1, 13.7]	12.33 [11.6, 13.0]
Orchestra	1.57 [1.2, 1.9]	1.26 [0.8, 1.7]	1.52 [1.1, 1.9]	1.17 [0.8, 1.5]	1.64 [1.3, 2.0]	2.41 [1.86, 3.0]	1.85 [1.4, 2.3]	1.96 [1.6, 2.3]	1.83 [1.5, 2.2]	2.39 [2.0, 2.8]
Secondary general	4.88 [3.9, 5.9]	5.23 [3.7, 6.7]	4.78 [3.4, 6.1]	5.00 [4.0, 6.0]	6.28 [4.7, 7.9]	3.59 [1.6, 5.6]	6.20 [4.5, 7.9]	5.93 [4.7, 7.1]	5.38 [4.3, 6.4]	5.99 [5.2, 6.8]
Piano	0.92 [0.5, 1.3]	1.04 [0.6, 1.5]	1.13 [0.6, 1.6]	0.93 [0.6, 1.2]	0.85 [0.6, 1.1]	0.65 [0.4, 0.9]	0.99 [0.6, 1.4]	1.53 [1.1, 2.0]	1.70 [1.4, 2.0]	1.52 [1.2, 1.9]
Guitar	1.61 [1.2, 2.0]	0.91 [0.6, 1.2]	0.69 [0.4, 1.0]	0.52 [0.3, 0.7]	0.79 [0.4, 1.2]	1.20 [0.6, 1.8]	1.18 [0.7, 1.6]	2.02 [1.6, 2.5]	1.50 [1.2, 1.8]	1.93 [1.5, 2.4]
Music theory	1.59 [1.2, 2.0]	0.90 [0.6, 1.2]	2.44 [1.4, 3.5]	1.48 [0.9, 2.1]	1.39 [1.0, 1.8]	1.64 [1.0, 2.3]	1.44 [1.0, 1.9]	2.15 [1.7, 2.6]	2.00 [1.5, 2.4]	2.26 [1.9, 2.7]
Music technology	0.06 [0.0, 0.1]	0.20 [0.0, 0.4]	0.13 [0.0, 0.2]	0.08 [0.0, 0.1]	0.31 [0.0, 0.7]	0.17 [0.0, 0.4]	0.46 [0.2, 0.7]	0.36 [0.1, 0.6]	0.31 [0.14, 0.5]	0.77 [0.45, 1.1]
IB music	0.00	0.00	0.00	0.00	0.00	0.00	0.02 [0.0, 0.1]	0.01 [0.0, 0.03]	0.03 [0.0, 0.1]	0.06 [0.0, 0.1]

Note. Percentages for “any music” and the various music subareas are for students who enrolled in at least one course of the type indicated. Percentages for the 2+, 3+, and 4+ years are for students who earned at least two, three, or four Carnegie units of credit in music. Estimates include students at both public and private schools. “Secondary general” includes any nonperformance class that was not music theory or composition, such as Music Appreciation or Music History. Students who enrolled in more than one kind of music course are counted in each appropriate subarea percentage, so those percentages do not total to the “any music” estimate. The 95% confidence intervals for each percentage estimate are reported in brackets. IB = International Baccalaureate.

Although I have operationalized “music enrollment” as the election of a single music course at some point during high school, it also is important to consider music credit-earning and course-taking persistence. Whereas “enrollment” is the simple election of the course, course credit is earned only for the completion of a course and the earning of a passing grade. Rows 2 through 4 of Table 1 present estimates of the percentages of students who earned at least two, three, or four Carnegie units’ worth of music on their high school transcripts. It is clear from the table that within each cohort, there was a process of attrition from music courses—though slightly more than one third of students enrolled in at least one music course, in 2009, slightly fewer than 1 in 10 persisted in receiving at least 4 years’ worth of music Carnegie units. Some attrition across the high school career is to be expected as increased opportunities for electives and advanced study in other areas, or requirements for remedial work, compete with elective enrollment in music. However, the level of attrition from initial enrollment to earning at least 4 years’ worth of music credit decreased substantially since 1982, when only 5.42% of students persisted in earning at least four Carnegie units in music. While the overall enrollment rates between 1982 and 2009 do not significantly differ, t(1710) = 0.31, p = .759, the increase between the estimates of the proportion of students earning at least four music credits in 1982 and 2009 is significant, t(1710) = 8.89, p < .001.

Disaggregating data by student subgroup was an important requirement of NCLB. Public school population and music student percentages for the subgroups enumerated in the law—student race/ethnicity, free or reduced price lunch (FRL) status, and Individualized Education Plan (IEP) status—are presented in Table A1, available in the online Appendix to this article (http://jrme.sagepub.com/supplemental). The “Overall” columns in Table A1 present the percentage of the total cohort that each subgroup composed, so, for example, 12.20% of all students in the graduating class of 1982 were identified as African American. The “Music” columns present the percentage of students earning at least one credit of music who were members of the subgroup, so, for example, 13.49% of all music students in the 1982 graduating cohort were identified as African American. Any missing data in Table A1 were unavailable in the source data set.

Rao-Scott adjusted χ² (Rao & Scott, 1984) analyses, reported in Table A2 in the online Appendix (available at http://jrme.sagepub.com/supplemental), indicate that public school music enrollment was associated significantly with race for each of the graduating cohorts under investigation, with the exception of the graduating cohorts of 1992 and 1998. Across the cohorts with a significant association, white students tend to be overrepresented, while Hispanic students tend to be underrepresented. Of the 3 years for which data are observable, Rao-Scott adjusted χ² indicated that FRL status was associated with public school music enrollment only in 2009, with students receiving FRL significantly underrepresented among music students, F(<10, 60) = 12.36, p < .001. Public school students receiving special education services under an IEP were underrepresented among music students in all cohorts for which the data were observable: 1987, F(< 10, 40) = 119.18, p < .001; 1990, F(<10, 60) = 31.46, p < .001; 1994, F(<10, 60) = 15.80, p < .001; 1998, F(<10, 50) = 22.37, p < .001; 2000, F(<10, 60) = 17.37, p < .001; 2004, F(<10, 200) = 26.18, p < .001; 2005, F(<10, 60) = 96.32, p < .002; and 2009, F(<10, 60) = 118, p < .001.

Research Question 2

Figure A1, available in the online Appendix to this article (http://jrme.sagepub.com/supplemental), displays the national music enrollment percentages and the regression estimated trends for all students in the public and private schools. Visual inspection of the graph suggests that public high school music enrollment remained relatively stable from 1982 until 2009, although a slightly increasing trend commensurate with increases in the private schools stagnated after the passage of NCLB. Regression analyses, the results of which are presented in Table 2, confirm the visual interpretation of the graph. Although the mean level of music enrollment in the public schools lagged behind that in the private schools prior to NCLB, this difference (–3.01) is not statistically significant (p = .109). Similarly, although the pre-NCLB growth trend for public school music enrollment is smaller than that of the private schools (–0.14), this difference is also not statistically significant (p = .416). However, after the passage of NCLB, private school enrollment in music grew at a significantly higher rate than did the public school growth rate, which remained nearly flat. The difference between the public and private schools in the change in slope from pre- to post-NCLB was small (–1.11) but statistically significant (p = .003). The slight decline in enrollment level immediately after the passage of the law was not significant for either the public or the private schools. By 2009, NCLB was estimated to have caused a 2.88 percentage point drop in the number of students electing to take music courses in public high schools; however, this decline was not statistically significant (p = .070) and may be due to the expected cohort-to-cohort variation in music enrollment rates. The gain in music enrollment rates in the private schools after the passage of NCLB, roughly 2.28 percentage points, also was not statistically significant (p = .533). Thus, when examining overall enrollment rates in high school music, it appears as though NCLB had no discernable effect on the percentage of students electing to pursue music nationwide.

Table 2.

Estimates of the Effect of No Child Left Behind (NCLB) on Nationwide Music Enrollment, Public Versus Private Schools Contrast.

Variable	Coefficient	SE	t	p
Public schools
Initial 1982 enrollment percentage	33.02*	1.18	27.89	.000
Pre-NCLB slope (trend)	0.11	0.08	1.38	.193
Immediate change at NCLB enactment	–0.42	1.40	0.30	.772
Post-NCLB slope (trend)	–0.24	0.15	1.57	.143
Pre- to post-NCLB slope change	–0.35	0.17	2.04	.064
Total effect of NCLB by 2009	–2.88	1.44	1.99	.070
Private schools
Initial 1982 enrollment percentage	36.03*	1.28	28.13	.000
Pre-NCLB slope (trend)	0.26	0.15	1.73	.109
Immediate change at NCLB enactment	–2.99	2.87	1.04	.317
Post-NCLB slope (trend)	1.01*	0.18	5.55	.000
Pre- to post-NCLB slope change	0.75*	0.23	3.21	.007
Total effect of NCLB by 2009	2.28	3.56	0.64	.533
Pre-NCLB differences between public and private schools
Public-private mean level of enrollment difference	–3.01	1.74	1.73	.109
Public-private slope (trend) difference	–0.14	0.17	0.84	.416
Difference in differences pre- to post- NCLB
Difference in public-private intercept change	2.58	3.19	0.81	.435
Difference in public-private slope (trend) change	–1.11*	0.29	3.79	.003

Note. Coefficients reported in percentages; mean levels interpretable as the percentage of the cohort enrolled in music and the slopes interpretable as the growth rate in enrollment percentages across time.

p < .05.

Research Question 3

Figures A2 through A9, available in the online Appendix to this article (http://jrme.sagepub.com/supplemental), display the public school music enrollment rates for most of the subgroups targeted by NCLB’s disaggregated data accountability provisions, with the exception of FRL status, which had no usable pre-NCLB data points. The percentages reported in these figures represent the percentage of students within each subgroup who enrolled in a music course during high school; so, for example, in 1982, 27.90% of Hispanic students in the public schools enrolled in at least one music course. Without enough reliable data for private school students in each of the targeted subgroups, these graphs compare the observed values in the NCLB-era with the continuation of the pre-NCLB trend to illuminate potential effects of the law on music enrollment. In nearly all cases, with the notable exceptions of the graphs for Hispanic students, Asian students, and students with IEPs, the 2009 observed enrollment rates do not appear to vary widely from the levels predicted by the pre-NCLB trends.

Table 3 presents the regression estimates analyzing the time series data shown in the figures. Confirming the visual interpretation of the graphs, the negative slope change (–0.61) and the negative total effect of NCLB by 2009 on the percentage of Hispanic students enrolling in music classes (–5.67) are significant, p = .027 and p = .047, respectively. These results suggest that NCLB caused the enrollment trend among Hispanic students, which was relatively flat prior to NCLB, to turn negative after the passage of the law, with roughly 5.67% of Hispanic students who would have been predicted to enroll in music courses not pursuing music courses in high schools. The significant increase in the enrollment trend among Asian students after the passage of NCLB, with 13.95% more of the 2009 Asian student population enrolling in music than pre-NCLB trends would have predicted, suggests that NCLB did not prevent Asian students from enrolling in music and that Asian students may have found it easier to enroll in music courses during the NCLB era. Figure A9 shows that students with IEPs were taking music at increasing percentages across the pre-NCLB years but that this trend turned negative after the passage of the law. The slope change (–0.89) is statistically significant, p = .006; however, the total effect of the decreasing trend by 2009 (–7.92 percentage points) fails to be significantly different from the prediction based on the pre-NCLB trend (p = .059), although it is possible that the reduced statistical power of this analysis due to the absence of a true comparison group may be the reason this total effect estimate is nonsignificant.

Table 3.

Estimates of the Effect of No Child Left Behind (NCLB) on Music Enrollment Within Subgroups.

	Subgroup
	(1) White	(2) African American	(3) Hispanic or Latino	(4) Asian/Pacific Islander	(5) Native American	(6) Other race or ethnicity	(7) ELL students	(8) Students with IEPs
Variable	b/(SE)/p	b/(SE)/p	b/(SE)/p	b/(SE)/p	b/(SE)/p	b/(SE)/p	b/(SE)/p	b/(SE)/p
Initial 1982 enrollment percentage	33.40*(0.80).000	36.59*(2.12).000	27.90*(2.02).000	31.88*(2.87).000	28.04*(0.98).000	36.84*(3.92).000	32.98*(3.91).000	17.68*(1.87).001
Pre-NCLB trend	0.17*(0.07).042	–0.03(0.16).837	0.003(0.15).983	–0.20(0.25).448	0.03(0.14).825	–0.81*(0.26).026	–0.69*(0.25).032	0.49*(0.15).030
Immediate change at NCLB enactment	0.14(1.90).946	0.93(2.17).683	–1.40(1.70).441	1.57(5.09).768	3.54(7.65).660	13.52(5.99).073	10.35*(4.00).041	–1.67(2.22).494
NCLB trend change	–0.40(0.25).161	–0.29(0.17).143	–0.61*(0.21).027	1.77*(0.68).041	0.16(1.07).884	–0.61(0.76).462	–1.45(0.59).050	–0.89*(0.17).006
Total effect of NCLB by 2009	–2.67(1.46).108	–1.12(3.01).723	–5.67*(2.28).047	13.95*(4.37).019	4.68(3.92).278	9.28(4.95).120	.19(3.19).955	–7.92(3.02).059

Note. All quantities reported in terms of the percentage within each cohort who were music students. ELL = English language learner; IEP = Individualized Education Plan.

p < .05.

A complementary finding to the significant negative effect of NCLB on Hispanic student music enrollment is evident in Figure A8 and in Model (7) of Table 3. Since 1982, there has been a statistically significant declining trend in the percentage of English language learner (ELL) students who enroll in music. The pre-NCLB slope (–0.69) is significant, p = .032, and although the NCLB-induced slope change fails to achieve significance (–1.45, p = .050), the resulting post-NCLB slope itself (–2.14) is still significantly negative and nonzero (p = .007), suggesting that with each passing year, fewer and fewer ELL students are enrolled in music.

Discussion

The present study was designed to evaluate the effect of NCLB on elective enrollments in high school music courses. The results obtained in these analyses suggest that while NCLB may have had no discernable effect on overall enrollment rates in music, NCLB’s disaggregated data provisions may have systematically prevented Hispanic students and students with IEPs from enrolling in music courses and may have exacerbated the declining trend of fewer ELLs enrolling in music courses. These results all corroborate and extend the findings of Elpus and Abril (2011), who examined the demographic profile of one nationally representative cohort of music ensemble students in the NCLB era. Given that research has demonstrated that Hispanic students, ELLs, and students with IEPs underperform their White, Asian, and nonclassified peers on standardized tests (Stiefel, Schwartz, & Chellman, 2007) and that schools have responded to consequential accountability mandates in ways unintended by policymakers (Figlio, 2006; Figlio & Winicki, 2005; Jacob, 2005; Jacob & Levitt, 2003), it is perhaps not surprising that these subgroups of students enrolled in music at declining rates during the NCLB era. The precise mechanism through which these students were excluded from music courses remains an open question. It is possible that school administrators, responding to consequential accountability pressures, systematically denied access to music courses for these subgroups in favor of courses more closely aligned with the high-stakes test, as was reported by teachers interviewed by West (2012) and principals surveyed by Gerrity (2009) and Abril and Gault (2008). This may have occurred through specific targeting of students within schools for enrollment in remedial course work or in the elimination or reduction of music programs in schools with higher concentrations of students in these subgroups. Future research examining within-school variation in course-taking patterns would help to elucidate the mechanism through which NCLB negatively affected the music enrollment of Hispanic students, ELLs, and students with IEPs. Understanding this mechanism would help the music education profession understand the potential effects of future similar state or national government initiatives, such as the Race to the Top reforms or the implementation of the Common Core State Standards and its associated testing program. It is particularly important for the field to explore empirically whether these students were, or are, systematically denied the opportunity to study music—a potential injustice with far-reaching implications for the profession.

The present study extends an important line of inquiry that examines selection into school music. Elpus and Abril (2011) examined the demographic profile of high school music ensemble students who participated in those ensembles as 12th graders in the 2003–2004 school year. They correctly noted the underrepresentation of Hispanic students among music students in that cohort but lacked the trend data analyzed in the present study to contextualize their results. The analyses reported here suggest that the underrepresentation of Hispanic students among music students long predates the 2004 graduating cohort and also that NCLB likely exacerbated the underrepresentation of Hispanic students in all types of music courses—not just ensembles. In the results presented in Table A1, it is evident that the proportion of music students who are Hispanic lies within 1 or 2 percentage points of the proportion of Hispanic students in the population in the pre-NCLB era and that all of the post-NCLB data points show disparities of 4 or more percentage points. Elpus and Abril speculated that geographic differences, issues of access, or cultural issues may have contributed to the relative scarcity of Hispanic students in music ensembles. The results of this study suggest that educational policy, too, may be an important factor to consider in the uneven representation of certain subgroups among music students.

Beyond evaluating the effects of No Child Left Behind on enrollment rates in music, I also sought to describe music enrollment rates across the 10 cohorts studied. With the operationalization of a “music student” as a student who enrolls in at least one music course during high school, it is clear from this study that the uptake rate of music courses in high school is remarkably stable at roughly 34%, despite oft-repeated criticism that our course offerings attract a far smaller proportion of students. This number is greater than the number reported by Elpus and Abril (2011), who looked only at ensemble courses taken during the senior year and who incorrectly speculated that there was a decline in music enrollment from a high point in 1982 (reported by Stewart, 1991). The results of the present study suggest that the decline Elpus and Abril observed was likely due to the instrumentation differences between the multiyear transcript data analyzed by Stewart (1991) and the cross-sectional self-report data used by Elpus and Abril. Transcript data analyzed here suggest that the overall music enrollment trend was flat, not declining, in the period between 1982 and 2009.

Although the subgroup disparities may be troubling, it should be heartening for most music teachers to learn that a core group of just over one third of all U.S. high school students, for nearly 30 years, has consistently chosen to enroll in a music class. That music teachers reach over one in three of all high school students with at least some courses should encourage music teachers to consider carefully the breadth as well as depth of their curricular offerings, bearing in mind that 10% or more of all students are likely to pursue a nonperformance class. Also encouraging is that for those students who do elect music, the average number of courses taken was increasing across this sample, with 9.43% of the class of 2009 persisting through at least 4 years’ worth of music courses, compared to just 5.42% of the class of 1982 achieving the same persistence benchmark. Given the relative stability of the overall uptake rates, it is plausible that efforts to recruit new students into music over the past three decades have resulted in less attrition from among the ranks of music students, also a result favorable to the field.

Caution in accepting the warrant for causal claims arising from the results of this study lies entirely with the potential that threats to the validity of short interrupted time series designs (Cook & Wong, 2008; Wong et al., 2009a) were not addressed adequately here. The most salient threat to the validity of this design in the present study is the history threat—that other events simultaneous with the implementation of NCLB caused the observed slope and intercept changes observed here and that the aforementioned Catholic sex abuse scandal is the most significant simultaneous historical event to consider (as did Dee & Jacob, 2011; Wong et al., 2009b). Historical events potentially associated with music enrollment occurring prior to NCLB, such as the introduction of the National Standards and economic conditions of the era (Elpus, 2013), are accounted for in the pre-NCLB trend; historical events potentially associated with music enrollment occurring in the post-NCLB era are accounted for similarly in the post-NCLB trend. Plausible alternative explanations for the observed results due to other simultaneous historical events not considered here, if they exist, may limit the warrant for causal claims in this study. Other limitations from this study arise from the structure of the existing data—most notably, the lack of state-level representative transcript data, which would permit the high- and low-dose comparisons also used by Dee and Jacob (2011) and Wong and colleagues (2009b).

Education policy generally is intended by policymakers at the federal, state, and local levels to improve educational quality and outcomes. Although education policies are rarely focused exclusively on music education, music educators and music education researchers should remain conversant in education policy trends and proposals to understand their potential effects on the status and quality of music education in the nation. Music educators often worry that all new educational policy is deleterious to music, yet a more nuanced approach to policy, undergirded by empirical evaluations of existing policy, is needed to inform the profession. That NCLB precipitated no overall decline in high school music enrollment rates is perhaps surprising to some, but the disparities in enrollment trends among the racial, ethnic, ELL, and special education subgroups that were exacerbated by NCLB is a legitimate cause for concern. At the time of this writing, NCLB waivers that have been granted to states by the Department of Education are tied to the adoption of other school reform efforts, including the linking of student standardized test scores with individual teacher evaluation. Authors of future research in music education ought to consider what effect these new federal education policies will have on access to and election of music education course work by secondary students.

Footnotes

Supplemental Material

The online Appendix is available at .

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

Author Biography

Kenneth Elpus is an assistant professor of music education at the University of Maryland, College Park. His research interests include music in education policy, music education as a context for adolescent development, and understanding the demographics of music students and the process of selection into music education.

References

Abril

C. R.

Gault

B. M.

(2006). The state of music in the elementary school: The principal’s perspective. Journal of Research in Music Education, 54, 6–20. doi:10.1177/002242940605400102

Abril

C. R.

Gault

B. M.

(2008). The state of music in secondary schools: The principal’s perspective. Journal of Research in Music Education, 56, 68–81. doi:10.1177/0022429408317516

Alt

M. N.

Bradby

(1999). Procedures guide for transcript studies (Working Paper No. 1999-05). Washington, DC: U.S. Department of Education, National Center for Education Statistics.

(2007). High-stakes testing and curricular control: A qualitative metasynthesis. Educational Researcher, 36, 258–267. doi:10.3102/0013189X07306523

Beveridge

(2009). No Child Left Behind and fine arts classes. Arts Education Policy Review, 111(1), 4–7. doi:10.1080/10632910903228090

Chapman

L. H.

(2004). No Child Left Behind in art? Arts Education Policy Review, 106(2), 3–20. doi:10.3200/AEPR.106.2.3-20

Chapman

L. H.

(2007). An update on No Child Left Behind and national trends in education. Arts Education Policy Review, 109(1), 25–40. doi:10.3200/AEPR.109.1.25-40

Colwell

(2005). Whither programs and arts policy? Arts Education Policy Review, 106(6), 19–29. doi:10.3200/AEPR.106.6.19-30

Cook

T. D.

Wong

V. C.

(2008). Better quasi-experimental practice. In Alasuutari

Bickman

Brannen

(Eds.), Sage handbook of social research methods (pp. 134–165). Los Angeles, CA: Sage.

10.

Dee

T. S.

Jacob

B. A.

(2011). The impact of No Child Left Behind on student achievement. Journal of Policy Analysis and Management, 30, 418–446. doi:10.1002/pam.20586

11.

Dillon

(2006, March 26). Schools cut back subjects to push reading and math. The New York Times. Retrieved from http://www.nytimes.com

12.

Elpus

(2013). Music in U.S. federal education policy: Estimating the effect of “core status” for music. Arts Education Policy Review, 114(1), 13–24. doi:10.1080/10632913.2013.744242

13.

Elpus

Abril

C. R.

(2011). High school music ensemble students in the United States: A demographic profile. Journal of Research in Music Education, 59, 128–145. doi:10.1177/0022429411405207

14.

Figlio

D. N.

(2006). Testing, crime and punishment. Journal of Public Economics, 90, 837–851. doi:10.1016/j.jpubeco.2005.01.003

15.

Figlio

D. N.

Winicki

(2005). Food for thought: the effects of school accountability plans on school nutrition. Journal of Public Economics, 89, 381–394. doi:10.1016/j.jpubeco.2003.10.007

16.

Gerrity

K. W.

(2009). No Child Left Behind: Determining the impact of policy on music education in Ohio. Bulletin of the Council for Research in Music Education, 179, 79–93.

17.

Heilig

J. V.

Cole

Aguilar

(2010). From Dewey to No Child Left Behind: The evolution and devolution of public arts education. Arts Education Policy Review, 111(4), 136–145. doi:10.1080/10632913.2010.490776

18.

(2012, February 9). 10 states are given waivers from education law. The New York Times. Retrieved from http://www.nytimes.com

19.

Jacob

B. A.

(2005). Accountability, incentives and behavior: The impact of high-stakes testing in the Chicago Public Schools. Journal of Public Economics, 89, 761–796. doi:10.1016/j.jpubeco.2004.08.004

20.

Jacob

B. A.

Levitt

S. D.

(2003). Rotten apples: An investigation of the prevalence and predictors of teacher cheating. Quarterly Journal of Economics, 118, 843–877. doi:10.1162/00335530360698441

21.

Lauen

D. L.

Gaddis

S. M.

(2012). Shining a light or fumbling in the dark? The effects of NCLB’s subgroup-specific accountability on student achievement. Educational Evaluation and Policy Analysis. doi:10.3102/0162373711429989

22.

Lee

(2006). Tracking achievement gaps and assessing the impact of NCLB on the gaps: An in-depth look into national and state reading and math outcome trends. Cambridge, MA: Civil Rights Project at Harvard University.

23.

Lee

Reeves

(2012). Revisiting the impact of NCLB high-stakes school accountability, capacity, and resources: State NAEP 1990–2009 reading and math achievement gaps and trends. Educational Evaluation and Policy Analysis, 34, 209–231. doi:10.3102/0162373711431604

24.

McMurrer

(2007). Choices and challenges: Curriculum and instruction in the NCLB era. Washington, DC: Center on Education Policy.

25.

Peterson

P. E.

West

M. R.

(2003). No child left behind? The politics and practice of school accountability. Washington, DC: Brookings Institution Press.

26.

Rao

J. N. K.

Scott

A. J.

(1984). On chi-squared tests for multiway contingency tables with cell proportions estimated from survey data. Annals of Statistics, 12(1), 46–60. doi:10.1214/aos/1176346391

27.

Ravitch

(2011). The death and life of the great American school system: How testing and choice are undermining education (Rev. and expanded ed.). New York, NY: Basic Books.

28.

Rich

(2012, July 6). “No Child” law whittled down by White House. The New York Times. Retrieved from http://www.nytimes.com

29.

Shadish

W. R.

Cook

T. D.

Campbell

D. T.

(2002). Experimental and quasi-experimental designs for generalized causal inference. Boston, MA: Houghton Mifflin.

30.

Spohn

(2008). Teacher perspectives on No Child Left Behind and arts education: A case study. Arts Education Policy Review, 109(4), 3–12. doi:10.3200/AEPR.109.4.3-12

31.

Stewart

(1991). Who takes music? Investigating access to high school music as a function of school and social factors (Doctoral dissertation). Available from ProQuest Dissertations and Theses database. (UMI No. 9208660)

32.

Stiefel

Schwartz

A. E.

Chellman

C. C.

(2007). So many children left behind: Segregation and the impact of subgroup reporting in No Child Left Behind on the racial test score gap. Educational Policy, 21, 527–550. doi:10.1177/0895904806289207

33.

West

(2012). Teaching music in an era of high-stakes testing and budget reductions. Arts Education Policy Review, 113, 75–79. doi:10.1080/10632913.2012.656503

34.

Wong

(2008). Studies of educational change from three disciplinary perspectives: Sociology, policy, and human development (Doctoral dissertation). Available from ProQuest Dissertations and Theses database. (UMI No. 304543742)

35.

Wong

Cook

T. D.

Steiner

P. M.

(2009a, March). Comparison groups in short interrupted time series: An illustration evaluating No Child Left Behind. Paper presented at the Society for Research in Educational Effectiveness Annual Conference, Crystal City, VA. Available from ERIC. (ED514988)

36.

Wong

Cook

T. D.

Steiner

P. M.

(2009b). No Child Left Behind: An interim evaluation of its effects on learning using two interrupted time series each with its own non-equivalent comparison series. Evanston, IL: Institute for Policy Research Working Paper Series. Retrieved from http://www.ipr.northwestern.edu/publications/papers/2009/ipr-wp-09-11.html

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.63 MB