You are seeing this message because your web browser does not support basic web standards. Find out more about why this message is appearing and what you can do to make your experience on this site better.

Program Description

Overview Citizen Schools (CS) operates a national network of apprenticeship programs for middle school students that connects adult volunteers to youth in hands-on after school learning projects. CS is designed to help youth develop academic and leadership skills needed to succeed in school, get into college, and become leaders in their careers and in their communities.
Start Date 1995
Scope national
Type after school
Location urban
Setting public school
Participants middle school students (grades 6–8)
Number of Sites/Grantees 30 sites in 2006–2007 nationwide
Number Served approximately 3,000 in 2006–2007 nationwide
Components Twice a week after school, CS youth participate in apprenticeships, which consist of hands-on learning projects led by volunteer Citizen Teachers. Apprentices work collaboratively in small groups to build solar cars, litigate mock trials, publish children's books, and so on. On other weekday afternoons, CS staff lead structured after school activities to enhance youth’s school success, working on homework and building organizational and study skills to help youth navigate their education through high school, graduation, and beyond. Each semester culminates in a “WOW!”—a public presentation of the CS participants’ projects.

In 2001–2002, CS launched its 8th Grade Academy program in Boston for eighth graders who began CS in a prior year. In addition to offering activities similar to those at other CS campuses, 8th Grade Academy is intended to help youth apply to and succeed in competitive high schools and to introduce them to the college application process through experiential learning activities that build academic and life skills and give youth access to coaches, technology, internships, and other educational programs. Each participant is assigned a writing coach (typically a local lawyer). In 2004, CS launched an alumni program to support Academy graduates during the high school transition period.
Funding Level $5,250,000 in 2001–2002; $6,370,000 in 2002–2003
Funding Sources Major funders include Edna McConnell Clark Foundation, Atlantic Philanthropies, New Profit Inc., Skoll Foundation, Jack Kent Cooke Foundation, Goldman Sachs Foundation, Fidelity Investments, Dorot Foundation, Bank of America, Daniels Fund, Mott Foundation, Comcast, CVS Pharmacy and individual donors.

Evaluation

Overview A 7-year evaluation (2001–2008) is being conducted and focuses on the program experiences and outcomes of 4 cohorts of participants at the Boston site.
Evaluators Policy Studies Associates, Inc. (PSA)
Evaluations Profiled Citizen Schools: Evidence from Two Student Cohorts on the Use of Community Resources to Promote Youth Development. Phase II Report of the Citizen Schools Evaluation

Putting Students on a Pathway to Academic and Social Success: Phase III Findings of the Citizen Schools Evaluation

Preparing Students in the Middle Grades to Succeed in High School: Findings from Phase IV of the Citizen Schools Evaluation
Evaluations Planned The evaluation’s final phase will include data through the 2007–2008 school year. In addition, CS plans to conduct another study of non-Boston sites.
Report Availability Espino, J., Fabiano, L., & Pearson, L. M. (with Kirkwood K. P., Afolabi, K., & Pasatta, K.). (2004). Citizen Schools: Evidence from two student cohorts on the use of community resources to promote youth development. Phase II report of the Citizen Schools evaluation. Washington, DC: Policy Studies Associates. Available at: www.policystudies.com/studies/youth/CS%20Phase%20II.html

Fabiano, L., Pearson, L. M., & Williams, I. J. (2005). Putting students on a pathway to academic and social success: Phase III findings of the Citizen Schools evaluation. Washington, DC: Policy Studies Associates. Available at: www.policystudies.com/studies/youth/CS%20Phase%20III%202005.pdf

Fabiano, L., Pearson, L. M., Reisner, E. R., & Williams, I. J. (2006). Preparing students in the middle grades to succeed in high school: Findings from Phase IV of the Citizen Schools evaluation. Washington, D.C.: Policy Studies Associates. Available at: www.policystudies.com/studies/youth/Citizen%20Schools%20Phase%20IV%20Final%20Report_12-26-06.pdf

Contacts

Evaluation Lee Pearson
Research Associate
Policy Studies Associates, Inc.
1718 Connecticut Avenue, NW, Suite 400
Washington, DC 20009
Tel: 202-939-9782
Fax: 202-939-5732
Email: lpearson@policystudies.com
Program Meridith Polin
Director of Research and Evaluation
Citizen Schools
308 Congress Street, 5th Floor
Boston, MA 02210
Tel: 617-695-2300 ext. 132
Fax: 617-695-2367
Email: meridithpolin@citizenschools.org
Profile Updated June 7, 2007

Evaluation 1: Citizen Schools: Evidence From Two Student Cohorts on the Use of Community Resources to Promote Youth Development. Phase II Report of the Citizen Schools Evaluation


Evaluation Description

Evaluation Purpose To determine whether program services were implemented as intended and, if so, whether the program had a positive impact on participants’ academic achievement, school engagement, and psychosocial skills and well-being.
Evaluation Design Quasi-Experimental and Non-Experimental: Data were collected on 2 cohorts of youth in 8 Boston CS campuses. Cohort I consisted of a sample of 253 sixth and eighth graders attending CS in fall 2001 (Year 1). Cohort II consisted of a sample of 323 sixth and eighth graders attending CS in fall 2002 (Year 2). To assess whether CS serves youth who were among the most disadvantaged of all BPS students, the demographic and educational characteristics of participants were compared to the general Boston Public Schools (BPS) student population. Baseline and follow-up surveys were collected from all eligible CS participants attending CS at the time of each survey. Survey data from program staff (campus directors, deputy directors, CS teaching associates), volunteers (Citizen Teachers), and parents were collected for Years 1 and 2.

To estimate program impacts on school-related performance, data were collected on a comparison group of 225 matched nonparticipants in Cohort I, and 260 matched nonparticipants in Cohort II. The core matching criteria included gender, race, grade in school, eligibility for free or reduced-price lunch (FRPL), and grade 4 Massachusetts Comprehensive Assessment System (MCAS) test scores in math and English language arts (ELA). When possible, youth were also matched on school attended, bilingual education status, and special education status. With the exception of constants like grade 4 MCAS scores, these data were based on youth’s 1st year in CS. MCAS scores were collected for both participants and control group members. Matched nonparticipants were identical to participants on the core matching criteria. Eighty-three percent of Cohort I sixth graders and 81% of Cohort I eighth graders as well as 90% of Cohort II sixth graders and 78% of Cohort II eighth graders matched within 8 points on MCAS scores. Twenty-six percent of matched nonparticipants attended the same school as their participants. The majority of matched nonparticipants were similar to their participants with respect to their bilingual education and special education status. Evaluators also compared the following subgroups of participants to their nonparticipants counterparts: those with high CS exposure (attending over 80% of CS days), grade level, cohort, and those in CS programs offering special activities (e.g., special writing courses).
Data Collection Methods Observation: Evaluators visited each of the 8 campuses for 2 days in spring 2002.

Secondary Source/Data Review: Data were collected from BPS records for Year 1 CS participants (60 Cohort I and 260 Cohort II youth) and BPS students overall on the following: demographics, school performance (e.g., attendance, high school choices), and academic achievement (i.e., grade 4 MCAS test scores and first marking period English and math grades). Evaluators continued to access BPS data on youth who stopped attending or graduated from CS as long as they remained in the BPS system. For the measure of high school choices, BPS high schools were classified as high, medium, or low quality. CS encouraged 8th Grade Academy graduates to attend high or medium quality schools. In classifying schools, CS considered each school’s size, MCAS performance, percent of graduates attending college or postsecondary training, attendance and dropout rates, grade 9 retention, reputation, and relationship with CS (since an existing partnership would help CS track participants and support them through high school).

CS provided program enrollment and attendance data.

Surveys/Questionnaires: Participant surveys examined youth’s CS experiences and their psychosocial well-being and development. Cohort I baseline surveys were completed by 234 youth in fall 2001, of which 145 completed the first follow-up in spring 2002, and 31 completed a second follow-up in spring 2003. Cohort II baseline surveys were completed by 323 youth in fall 2002, of which 187 completed a follow-up in spring 2003 .

Staff surveys asked about roles and responsibilities in CS, training activities, parent involvement, staff meetings, program planning and activities, use of student teams, interactions with participants’ schools, and teaching methods. Parent surveys asked about demographics and family background, the child’s participation in CS, perceptions of and interactions with CS, and perceptions of program impacts on their child. Volunteer surveys asked about apprenticeships offered, training and support from CS, planning and preparation, teaching methods, perceptions of their overall experience, and background questions. Staff and parent surveys were administered once a year in the spring. The volunteer survey was administered in fall 2001, spring 2002, and spring 2003. This report mainly uses Year 2 survey data, but provides some comparisons with Year 1 survey data. In Year 2, 228 parents, 57 staff members (8 campus directors and deputy directors and 49 teaching associates), and 43 volunteers completed surveys.

Test/Assessments: MCAS is a standardized test required by Massachusetts public schools. Results are reported as scaled scores and performance levels, defined as follows: Advanced, a comprehensive and in-depth understanding of rigorous subject matter and provides sophisticated solutions to complex problems; Proficient, a solid understanding of challenging subject matter and solves a wide variety of problems; Needs Improvement, a partial understanding of subject matter and solves some simple problems; and Warning, a minimal understanding of subject matter and does not solve simple problems.

The following scales are embedded in youth surveys: (a) Connection to CS, 5 items measuring agreement with statements such as “I feel like I belong here;” (b) Positive Relationships With Staff, 9 items measuring agreement with statements such as “CS staff always keep their promises;” (c) Peer Support, 3 items measuring agreement with such statements as “Students go out of their way to help each other;” (d) Leadership Opportunities, 4 items measuring frequency of engaging in activities such as “helped plan a CS event or activity like an Exploration or a Potluck Dinner;” (e) Active Learning Opportunities, 18 items measuring the frequency that CS activities, staff, and volunteers accomplish things like “make learning fun for you;” (f) Self-Efficacy, 8 items measuring agreement with statements such as “If something looks too hard, I will not even bother to try it” and 9 items measuring how sure youth are that things will “work out well” when faced with occurrences such as having “to learn something new at school;” (g) Locus of Control, 4 items measuring agreement with statements such as “In my life, good luck is more important than hard work for success;” (h) Social Competence, 5 items measuring agreement with statements such as “I can make suggestions to other students without being bossy;” (i) Positive Interpersonal Behavior, 6 items measuring frequency of engaging in behaviors such as trying to understand why others feel a certain way when they have problems with other youth; (j) Relationship With Parents or Guardians and Relationship With Adults, two 4-item scales measuring frequency of talking with parents/guardians or other adults about topics such as what is going on in their lives; and (k) Future Optimism, 10 items measuring chances of accomplishing things such as graduating from high school. For all scales, a rating of 1 indicates the absence or lowest level of that characteristic, with higher numbers indicating greater presence or strength of that characteristic.
Data Collection Timeframe Data were collected during the 2001–2002 (Year 1) and 2002–2003 (Year 2) years; data refer to Year 2 unless otherwise indicated.


Findings:
Formative/Process Findings

Activity Implementation Depending on the campus, program activities were offered from 2.25 hours to 4 hours per day, with an average of 3.5 to 4 hours, over an 8- to 10-month period. On average, programs offered activities on 81 days, with 6 campuses offering 3 days of activities each week, and 2 offering 4 days. Activities were mainly offered weekdays after school, but 2 campuses, also offered activities on Saturdays.

All campuses held apprenticeships once a week for 1.5 to 2.75 hours, in 3 to 8 areas each semester, mostly in the fields of arts, law, and business.

Writing and data analysis/math projects were each offered once a week for 1.25 hours to just under 3.5 hours; 6 campuses offered these projects at least 1.5 hours a week.

All campuses offered explorations of local institutions, neighborhoods, and universities once a week for anywhere from just under an hour to approximately 3 hours, with half of the campuses offering fewer than 1.5 hours and the other half offering more than 2 hours.

Five campuses offered homework time every day of programming, while the remaining 3 offered this activity on all but one program day. The weekly amount of homework time ranged from just under 2 hours to nearly 5 hours.

Compared to other CS activities, campus directors rated their implementation of explorations and apprenticeships as more successful and reported more challenges with writing and data analysis/math projects and homework. Directors cited youth absenteeism and lack of time as the most pervasive challenges in implementing program elements.
Parent/Community Involvement The majority of surveyed parents reported that staff contacted them at least twice a month to discuss their child’s progress in the program (73%), and academic performance (62%). Parents also reported that staff requested their attendance or assistance at events such as parent meetings or potluck dinners.

Analyses of parent survey data and CS characteristics revealed that staff’s prior experience working with youth was positively and significantly related to the frequency with which staff talked to parents about their child’s academic progress (p = 0.03). As the number of hours of exploration activities increased, parents’ reports of the frequency with which staff talked to them about their child’s progress in CS also significantly increased (p = 0.02).

The activities in which the greatest numbers of parents reported being involved were campus registration/orientation meetings (54%) and potluck dinners (53%). The activity that parents reported participating in least was helping to organize a family event (3%).

Campuses that provided the most hours of academic skill-building tended to have significantly more parents reporting attending a CS orientation/registration meeting (p = 0.03) or a potluck dinner (p = 0.01). Also, campuses that employed staff with more prior CS experience tended to have significantly more parents reporting attending a CS potluck dinner than campuses with less experienced staff (p = 0.00).
Program Context/Infrastructure During their 1st year of CS, Cohort I and II participants scored an average of 2.74 out of 4 on the Peer Support scale on 1st year follow-up surveys. Among Cohort I sixth graders who participated for a 2nd year, the average score on this scale was 2.68.

First year follow-up surveys indicated that during their 1st year of CS, Cohort I and II participants scored an average of 1.90 out of 3 on the Leadership Opportunities scale. The average score of Cohort I sixth graders who participated for a 2nd year was 2.04. A positive relationship was found between scores on the Leadership Opportunities index and the number of hours of exploration activities offered and staff with prior experience in CS. In contrast, a negative relationship was found between scores on this measure and the number of youth per CS staff member and hours of academic skill-building activities.

The average score for Cohort I and II participants during their 1st year of participation on the Active Learning Opportunities scale was 2.89 out of 4. The average score was 2.98 among Cohort I sixth graders who participated for a 2nd year.
Recruitment/Participation CS participants, like BPS students overall, were about evenly split between males and females. African Americans made up 69% of participants, a higher percentage than for BPS students overall (47%). Hispanics were the next largest ethnic group among participants (23%) compared to 29% of BPS students overall. Whites, Asians, and Native Americans together constituted 9% of all participants. In contrast, 14% of BPS students were White, 9% were Asian, and less than 1% were Native American.

A higher percentage of CS participants than BPS students overall were FRPL eligible (90% vs. 73%) and in special education programs (23% vs. 19%). However, a smaller percentage of CS participants were enrolled in a bilingual education program than BPS students overall (5% vs. 10%).

Of surveyed parents, 21% reported something other than a job as their primary source of income and 32% reported completing college (10% of which also had a graduate degree).

Both CS participants and BPS students overall attended school 93–95% of the time. Fourteen percent of CS participants were suspended at some point during the year.

Less than half (45%) of Cohort II CS sixth graders performed at the standard of Needs Improvement or better on the MCAS math test as fourth graders, with only 5% performing at a Proficient or Advanced level; the remaining 55% performed at a Warning level. In contrast, 42% of BPS sixth graders performed at a Warning level on the 2001 MCAS math test as fourth graders. Among Cohort II CS eighth graders, 61% performed at a Needs Improvement level or better, with 14% performing at a Proficient level. Thirty-nine percent of CS eighth graders performed at the Warning level, compared to 43% of BPS eighth graders who took the test as fourth graders in 1999.

Among Cohort II CS participants with available data, the average first period math grade was a C- for sixth graders (15% failed the period, and only 34% earned a B- or better) and a C for eighth graders (15% failed the period, and 38% passed with a B- or better).

About a quarter of Cohort II CS participants performed at a Warning level on the ELA MCAS test (28% of sixth graders and 25% of eighth graders), with similar proportions at a Warning level for BPS overall. However, a higher percentage of BPS sixth graders (24%) performed at Proficient or Advanced levels than CS sixth graders (13%).

Among Cohort II CS participants with available data, both sixth and eighth graders had an average first period English grade, of C ; 7% of sixth graders and 10% of eighth graders failed the period while 45% of sixth graders and 34% of eight graders had a B- or higher.

The number of days a youth attended CS in Year 2 ranged from 1 to 98 (median = 62). Two thirds of youth participated for at least 60% of the days the program was offered, while 50% participated for 80% or more.

Sixty-eight percent of youth enrolled in CS in Year 2 were enrolled for the duration of the fall and continued to be enrolled at the time of survey administration in the spring. Nearly equal percentages of youth withdrew from CS during the fall semester, between the fall 2002 and spring semesters, and during the spring semester (8–10%).

Youth who continued to participate in CS for a 2nd year had significantly higher baseline Locus of Control scale scores than those who withdrew (2.91 vs. 2.72 out of 4, p = .03).

When asked on the survey how often they missed a CS activity due to arriving late or leaving early, 48% of participants reported never, 22% reported usually or almost always, while the remaining youth reported sometimes.
Satisfaction The average score on a survey item measuring youth satisfaction with CS in Year 1 was 3.22 out of 5. The average score for Cohort I youth who returned for a 2nd year was 3.44.

The average youth score on the Connection to CS scale was 2.87 out of 4 in Year 1. The average for Cohort I sixth graders who continued to participate for a 2nd year was 3.04.

The average parent survey score of overall satisfaction with CS was 4.04 out of 5. Parents of children at campuses that offered more hours of academic skill-building were more satisfied than parents of children a campuses with fewer hours of such activities, though this difference was only marginally significant (p < .10). Parents of children at campuses that offered more hours of interpersonal skill-building were significantly less satisfied than parents of children at campuses with fewer hours of such activities (p = 0.03).
Staffing/Training The youth-to-staff ratio averaged 10:1 across campuses.

The majority of CS staff had a college degree or above—all leadership staff (campus directors and deputy directors) and 86% of Citizen Teachers had at least a college degree; 51% of Citizen Teachers also had a graduate degree—and female (especially leadership staff). Direct service staff tended to be under 25 years old, and White; leadership staff were more likely than other staff to be non-White and older than 25. In Year 2, almost two thirds of staff had experience working with CS, an increase of 24 percentage points from Year 1.

The average survey rating of staff/volunteer enjoyment of working at CS was 3.61 out of 4.

The average survey rating of staff training and support satisfaction was 2.75 out of 4.

The average score on the Relationships With Staff scale was 3.02 out of 4 for Cohort I and II youth during their 1st year of CS participation. Among Cohort I CS sixth graders who continued to participate for another year, the average score on this scale was 3.15.


Summative/Outcome Findings

Academic A significantly greater percentage of 8th Grade Academy graduates selected high quality high schools than matched nonparticipants (70% vs. 12% for Cohort I, and 70% vs. 46% for Cohort II, p = .001).

A higher percentage of Cohort I 8th Grade Academy participants were promoted to grade 10 on time than matched nonparticipants (87% vs. 74%); this difference was marginally significant (p = 0.07). A significantly higher percentage of high exposure participants were also promoted to grade 10 on time than matched nonparticipants (90% vs. 73%, p = 0.05).

In Year 1, CS sixth graders overall performed significantly better than matched nonparticipants on attendance, suspension rates, and grade promotion (p < .05).

Sixth graders with high CS exposure outperformed matched nonparticipants on attendance, suspension rates, math and English grades, and promotion (p < .05).

Seventh graders in Cohort I outperformed matched nonparticipants on suspension rates, MCAS ELA scores, and math course grades.

CS eighth graders performed better than matched nonparticipants on attendance.

In their ninth grade year, Cohort I 8th Grade Academy graduates did not significantly outperform matched nonparticipants on any of the evaluation’s indicators.
Youth Development Cohort I youth's scores on the Locus of Control scale and a confidence in public speaking subitem of the Self-Efficacy scale demonstrated a marginally significant increase from fall 2001 to spring 2003 (p < .10). They did not exhibit any changes on the two overall Self-Efficacy scales, or on a confidence in writing subitem. Cohort II youth did not increase any of their self-efficacy/confidence scores from baseline to first follow-up.

No significant changes were seen from fall 2002 to spring 2003 for either CS cohort on scales of Social Competency, Positive Interpersonal Behavior, Relationship with Parents or Guardians, Future Optimism, or importance of school for future success subitems.

Comparisons of the average difference between baseline and first follow-up scores for Cohort I and II revealed that in Year 1 and 2, participants experienced similar changes on measures of psychosocial skills and well-being.

Evaluation 2: Putting Students on a Pathway to Academic and Social Success: Phase III Findings of the Citizen Schools Evaluation


Evaluation Description

Evaluation Purpose To describe CS participation and the development of psychosocial skills and well-being, school engagement, and academic achievement of CS participants.
Evaluation Design Quasi-Experimental: Data were collected on all CS participants enrolled in Boston Public Schools (BPS) in 2003–2004 (Year 3, N = 855). To assess whether CS serves youth who are among the most disadvantaged of all BPS students, participants’ demographic and educational characteristics were compared to the general BPS population in grades 6–8 (N = 14,183). In addition, 8th Grade Academy participants’ enrollment in college-track high schools and on-time promotion to grade 10 were assessed. Surveys were collected from all eligible CS participants attending CS in the spring.

To estimate program impacts on school performance, data were collected on a comparison group of youth not participating in CS (n = 855). The core matching criteria included gender, race, grade in school, free or reduced-price lunch (FRPL) eligibility, and grade 4 Massachusetts Comprehensive Assessment System (MCAS) test scores in math and English language arts (ELA). When possible, youth were also matched on school attended, and bilingual/special education status. With the exception of like grade 4 MCAS scores, these data were based on youth’s 1st year in CS. MCAS scores were collected for participants, control group members, and BPS students overall. Youth with high and low CS exposure levels were compared to matched nonparticipants. High exposure was defined as attending 60% or more of the days CS was offered in both the fall and spring. For 8th Grade Academy participants, this requirement further stipulated that this participation rate be maintained for 2 or more years in each semester. Low exposure was defined as all participants who did not meet these criteria. Demographic characteristics did not significantly differ by exposure level. Exposure level differences are only presented when they are statistically significant.
Data Collection Methods Secondary Source/Data Review: BPS provided data on student achievement (MCAS scores and grades), demographics, and school performance (e.g., attendance, high school choices) for CS participants and BPS students overall. Evaluators continued to access BPS data on youth who stopped attending or graduated from CS as long as they remained in the BPS system. In addition, CS provided program enrollment and attendance data. For the measure of high school choices, BPS high schools were classified as high, medium, or low quality. CS encouraged 8th Grade Academy graduates to attend high or medium quality schools. In classifying schools, CS considered each school’s size, MCAS scores, percent of graduates attending college or postsecondary training, attendance and dropout rates, grade 9 retention, reputation, and relationship with CS (since an existing partnership would help CS track participants and support them through high school).

Surveys/Questionnaires: Surveys measured participants’ perceptions of their CS experiences as well as psychosocial well-being and development.

Test/Assessments: MCAS is a standardized test required by Massachusetts public schools. Results are reported as scaled scores and performance levels, defined as follows: Advanced, a comprehensive and in-depth understanding of rigorous subject matter and provides sophisticated solutions to complex problems; Proficient, a solid understanding of challenging subject matter and solves a wide variety of problems; Needs Improvement, a partial understanding of subject matter and solves some simple problems; and Warning, a minimal understanding of subject matter and does not solve simple problems.

The following scales are embedded in the youth survey: (a) Connection to CS, 5 items measuring agreement with statements such as “I feel like I belong here;” (b) Positive Relationships With Staff, 9 items measuring agreement with statements such as “CS staff always keep their promises;” (c) Peer Support, 3 items measuring agreement with statements such as “Students go out of their way to help each other;” (d) Leadership Opportunities, 4 items measuring frequency of engagement in activities such as “helped plan a CS event or activity;” and (e) Active Learning Opportunities, 18 items measuring the frequency that CS accomplishes things like “really get you thinking.” For all scales, higher numbers indicate greater presence or strength of that characteristic.
Data Collection Timeframe Data were collected during the 2003–2004 year (Year 3).


Findings:
Formative/Process Findings

Program Context/Infrastructure Average participant survey assessment scores (on a scale of 1–4, unless otherwise noted) were as follows: (a) Connection to CS: 2.94 for sixth graders, 3.01 for seventh graders, 2.84 for eighth graders; (b) Positive Relationships With Staff: 3.11 for sixth graders, 3.13 for seventh graders, 2.95 for eighth graders; (c) Peer Support: 2.80 for sixth graders, 2.77 for seventh graders, 2.64 for eighth graders; (d) Leadership Opportunities (scale of 1–3): 1.97 for sixth graders, 2.05 for seventh graders, 1.83 for eighth graders; and (e) Active Learning Opportunities: 3.00 for sixth graders, 2.99 for seventh graders, 2.81 for eighth graders. Eighth graders scored approximately 4% lower on each measure than other grades.
Recruitment/Participation The CS sample had a significantly higher proportion of African Americans than BPS overall (67% vs. 46%, p < .001). With the exception of Native Americans, who made up less than 1% of both, BPS overall had significantly higher proportions of other races than CS (p < .001), including Hispanic (30% vs. 24%), White (14% vs. 6%), and Asian (9% vs. 3%). Both BPS overall and CS participants were evenly split between girls and boys.

A significantly higher percentage of CS sixth graders were enrolled in the BPS program for low-achieving students than BPS sixth graders overall (34% vs. 28%, p < .01). Similarly, a significantly smaller percentage of CS sixth graders were enrolled in the BPS program for high achieving students than BPS sixth graders overall (3% vs. 9%, p < .001).

Of CS youth, a significantly higher percentage were FRPL eligible (90% vs. 81%) and a significantly lower percentage were enrolled in a special education program (24% vs. 30%) than BPS students overall ( p < .001). The percentage of youth in the Bilingual Education program (5%) did not differ between CS and BPS overall.

Most CS youth scored in the bottom 2 MCAS performance levels for grade 4 in ELA (89%) and math (92%); BPS fourth graders overall were somewhat less likely to score at the bottom levels (84% for ELA and 86% for math). BPS students overall had significantly higher scaled grade 4 MCAS ELA (p < .01) and math (p < .001) scores than CS youth.

Participants averaged about a C in English and math in the first period that they began CS (course grade data were unavailable for BPS students).

CS youth and BPS students did not differ in pretest school attendance or suspension rates.

Of all CS youth, 30% were eligible to participate in all 3 years that CS was offered, of which 28% participated for all 3 years, 40% in grade 6 only, 20% in grade 6 and 8 only, and 12% in grades 6 and 7 only.

The following attended 60% or more of CS days in Year 3: 59% of sixth and seventh graders in their 1st year of CS; 57% of seventh graders in their 2nd year of CS; and 75% of 8th Grade Academy participants (regardless of how many years they had participated).


Summative/Outcome Findings

Academic 8th Grade Academy youth attended school significantly more often than comparisons (p < .01) overall (92% vs. 88%), and for those with high exposure (95% vs. 90%), although the two groups did not differ on grade 8 school suspension rates, overall or by exposure.

Former 8th Grade Academy participants with high exposure had significantly higher (p < .05) school attendance (94% vs. 88%) and lower suspension (2% vs. 10%) rates in grade 9 than matched nonparticipants; these differences did not hold overall or by low exposure.

8th Grade Academy youth were promoted to grade 9 at a significantly higher rate (p <. 01) than comparisons overall (98% vs. 94%) and for high exposure (100% vs. 92%). Former 8th Grade Academy youth and comparisons did not differ on grade 10 promotion rates.

8th Grade Academy youth had slightly but significantly higher scaled grade 8 MCAS math scores than comparisons (p < .01 overall, p < .05 for high and low exposure groups). Like comparisons, however, they performed at low levels; over half scored at the Warning level.

8th Grade Academy youth did not differ from comparisons on grade 8 English and math grades, as measured in the third marking period. Former 8th Grade Academy youth had significantly higher grade 9 English and math grades than comparisons (p < .05); high exposure youth also had significantly higher English grades (p < .05), and low exposure youth had significantly higher math grades (p < .01) than comparisons.

Former 8th Grade Academy youth selected high-quality high schools at a significantly higher rate (p < .001) than comparisons, overall (72% vs. 32%) and for high (70% vs. 34%) and low (74% vs. 31%) exposure.

First year CS sixth and seventh graders attended school at a significantly higher rate (p < .001) than comparisons overall (93% vs. 90%) and for the high exposure group (95% vs. 90%), with an average of 9 more days of school for participants overall.

High exposure seventh graders in their 2nd year of CS had significantly higher school attendance rates than comparisons (95% vs. 93%, p < .05); this difference did not hold for the sample overall.

First year CS sixth and seventh graders were suspended significantly less often (p < .001) than comparisons overall (10% vs. 16%) and for those with high exposure (5% vs. 16%). Suspension rates of 2nd year CS seventh graders did not differ from comparisons.

First year CS sixth and seventh graders were promoted to the next grade at a significantly higher rate than comparisons overall (96% vs. 94%, p < .05) and for those with high exposure (98% vs. 94%, p < .01). Promotion rates of 2nd year CS seventh graders did not differ from comparisons.

First year CS sixth and seventh graders overall did not differ from comparisons in third period English and math grades, though high exposure youth significantly outperformed comparisons (p < .001 for English; p < .05 for math). Seventh graders in their 2nd year of CS significantly outperformed comparisons in math grades overall (p < .01) and for those with high exposure (p < .05 ), although they did not differ in English grades.

First year CS youth scored significantly higher on the grade 7 MCAS ELA test than comparisons overall (p < .05) and amongst those with high exposure (p < .01), although the two groups did not differ in the percent at the Warning level. First year CS youth performed similarly to comparisons on grade 6 MCAS math scaled scores and the percent scoring at Warning level. Youth in their 2nd year of CS were less likely than comparisons to score at the Warning level on the grade 7 MCAS ELA test (p < .01 overall, and p < .05 for high- and low-exposure), although no differences were found in scaled scores.

Evaluation 3: Preparing Students in the Middle Grades to Succeed in High School: Findings from Phase IV of the Citizen Schools Evaluation


Evaluation Description

Evaluation Purpose To analyze program impacts on former participants’ school engagement and academic achievement in ninth and tenth grades over time.
Evaluation Design Quasi-Experimental: Data were collected on 8th Grade Academy youth who attended Boston Public Schools (BPS) in one of four cohorts: 2001–2002 (Year 1) through 2004–2005 (Year 4). Each participant was matched to 3 nonparticipants who attended BPS, based on five core matching variables: gender, race, grade in school, free or reduced-price lunch (FRPL) eligibility, and grade 4 Massachusetts Comprehensive Assessment System (MCAS) test scores in math and English language arts (ELA). When possible, youth were also matched on school attended, and bilingual/special education status. With the exception of constants like grade 4 MCAS scores, these data were based on the year prior to each youth’s participation in the evaluation. Each participant’s 3 matches were ranked based on how similar they were to the participant on the matching criteria. Participants were compared to their most similar nonparticipant. Matched nonparticipants and participants were similar on all matching criteria. A participant’s matched nonparticipant sometimes changed from year to year; the second or third most similar match sometimes had to be used when a matched nonparticipant left the BPS system or was not promoted on schedule.

Between Years 1 and 4, 354 youth participated in 8th Grade Academy and had parental consent to be included in the evaluation: 48 in Year 1, 85 in Year 2, 118 in Year 3, and 103 in Year 4. To be included the analyses, youth had to be promoted to grade 9 at the end of the year in which they participated in 8th Grade Academy; 97% of the sample did so. Grade 9 promotion did not vary significantly by class year nor between participants and nonparticipants. Analyses focused on former 8th Grade Academy youth as they transitioned to high school and progressed through grades 9 and 10. Youth with high and low CS exposure levels were compared to their matched nonparticipants. High exposure was defined as 2 or more years of CS participation at an attendance rate of 60% or more each semester for all semesters. Low exposure was defined as all participants who did not meet these criteria. Of former 8th Grade Academy participants, 49% were classified as high exposure. Exposure level differences are only presented when they are statistically significant.
Data Collection Methods Secondary Source/Data Review: BPS and 3 Boston-area charter schools provided data on student achievement (MCAS scores and course grades), demographics, and school performance (e.g., attendance, high school choices). For the measure of high school choices, BPS high schools were classified as high, medium, or low quality. CS encouraged 8th Grade Academy graduates to attend high or medium quality schools. In classifying schools, CS considered school size, MCAS scores, percent of graduates attending college or postsecondary training, attendance, dropout rates, grade 9 retention, reputation, and relationship with CS (since an existing partnership would help CS track participants and support them through high school).

Test/Assessments: MCAS is a standardized test required by Massachusetts public schools. Results are reported as scaled scores and performance levels, defined as follows: Advanced, a comprehensive and in-depth understanding of rigorous subject matter and provides sophisticated solutions to complex problems; Proficient, a solid understanding of challenging subject matter and solves a wide variety of problems; Needs Improvement, a partial understanding of subject matter and solves some simple problems; and Warning, a minimal understanding of subject matter and does not solve simple problems.
Data Collection Timeframe Data covered 2001–2002 (Year 1) through 2004–2005 (Year 4).


Findings:
Summative/Outcome Findings

Academic Former 8th Grade Academy participants selected high quality high schools at a significantly higher rate than their matched participants, overall (65% vs. 26%, p < .001) and among both the high exposure (65% vs. 22%) and the low exposure (64% vs. 31%) groups.

Former 8th Grade Academy participants attended school at significantly higher rates in their ninth-grade year than comparisons (87% vs. 82%, p < .05), for a difference of approximately 9 days of school.

Former 8th Grade Academy youth with high exposure (but not overall) had significantly lower suspension rates in their ninth-grade year than matched nonparticipants (1% vs. 10%, p < .01).

Former 8th Grade Academy participants had significantly better (p < .05) grade 9 English course grades than comparisons overall and among the high exposure group. They did not significantly differ on average grade 9 math course grades.

Former 8th Grade Academy youth with high exposure (but not overall) were promoted on-time to grade 10 at a significantly higher rate than comparisons (92% vs. 81%, p < .05).

Former 8th Grade Academy participants had significantly higher grade 10 math course grades than comparisons (p < .05). The two groups did not differ on other grade 10 measures.

© 2010 Presidents and Fellows of Harvard College
Published by Harvard Family Research Project