K12 Enrollment:The purpose of this section is to explore historical, current, and future trends in schoolenrollment. General regional population trends have been outlined in the About Berkshire Countysection of this report.School Enrollment:Various studies of enrollment have been conducted over the last decade. This section will draw fromseveral key resources: Historical analysis of K12 school enrollment patterns/trendsThe Berkshire Regional Planning Commission Berkshire County Public School EnrollmentProjections (Versions 2015, 2018, 2020)The MARS recent Regional School District Planning Board (RSDPB) analysis of Southern Berkshireand Berkshire Hills enrollmentCurrent estimates from the most available ESE data setsSome early indicators of enrollment as influenced by the COVID-19 pandemicIt is important to distinguish how enrollments are computed in that some models include PreKindergarten (PK) and Special education Post-graduate (SP) students, others do not. 15,348.14,748.In 2020, students enrolled in Berkshire County public schools with PK & SP.In 2020, students enrolled in Berkshire County public schools without PK & SP.28%Current21%Decline sinceEnrollmentFurther Decline by20002020203020,47714,74811,651Methodologies for tracking enrollment:Enrollment data is collected through several required ESE required reporting processes. These reportsinclude Student information management data accessed in October, January, March, and in June, yearend. Additionally, a school-attending report is submitted based on January 1 enrollment numbers.These figures are used to calculate enrollment and are critical in establishing foundation budgets andthe transfer of funds across communities via tuition, charter, and choice. District and school specificdata can be accessed at: process of projecting into the future is always somewhat speculative and imprecise. Influencingfactors can include births, migration patterns, choice, and both charter and vocational enrollments.Most models (such as BRPC’s) use available birth data, average five years of birth data to determine rateof change, which is then applied to project future enrollments. Additionally, patterns of choice andmigration are tracked and applied forward. For example, if a particular district has a pattern of studentschoicing out-of-district as they advance from 5th-6th grade (elementary to middle), this historical patternis assumed and applied to future cohort projection. These models, while accounting for historical1

patterns of student migration, do not assume all future possibilities such as the construction of a newschool (that could draw additional out-of-district students in) or a particular perception issue (such asviolence/safety concerns or below-average test scores) that might increase out-migration in a particularschool/district.Finally, this model does not account for external community factors such as in or out-migration due tobusiness development/recruitment, construction of housing complexes, or unforeseeable factors likethe recent pandemic that some believe will result in more remote workers who choose communities likethe Berkshires to work and raise families.Historical Studies & Projections:Several enrollment studies have been conducted in recent years. These include: UMass Donahue Institute (2016). Conducted in 2016, applied three methodologies and BRPCestimates with an average projected loss calculated. Berkshire Regional Planning Commission (2015, 2018, 2020). First comprehensive studycompleted in 2015 and has been updated twice with the most recent 2020 version. Massachusetts Association of Regional Schools (2020). Completed for Southern Berkshire andBerkshire Hills Regional School District Planning Board. New England School School Development Council (NESDEC). Has conducted enrollment studiesacross the Berkshires for district planning purposes.These reports used varying methodologies and focused on different time periods, see Table 1.However, there is some ability to compare the closest similar projected timeframe from each study todetermine to what degree they vary (or not). All projections suggest a consistent decline in populationacross the 5-19 age group and/or the K-12 total enrollment. What is striking, however, is that the morerecent models project a higher rate of enrollment decline. While the most conservative estimate of tenyear change was the UMDI V2015 model at -7.1% loss, the most recent BRPC study suggests this changewill be more than double at -17.0% loss.Table 1. Selected enrollment studies, 2015 - 2020Study20152025% Change 20152025Used several methodologies to evaluate population changeamong 5-19 year olds (extrapolated to fit 2025 time frame)UMDI V201519,78618,374-7.1%UMDI CCR21,82219,214-12.0%UMDI CSM21,28819,507-8.4%BRPC 201321,83818,408-15.7%UMDI Average21,18418,876-10.8%Evaluated K-12 EnrollmentsBRPC 201515,90414,115-11.2%BRPC 201815,911 (actual)13,478-15.3%BRPC 202015,911 (actual)13,099-17.0%Examined BHRSD & SBRSD K-12 projected enrollment through 203020202030% Change 2020 2030MARS 20201,7871,404-21.4%BRPC 20201,7801,280-28.1%2

The most recent numbers for comparison are pulled from the RSDPB MARS study. This study projectedenrollments through 2020 through 2030 for the Berkshire Hills and Southern Berkshire regional districts.As a comparison, projections from the most recent BRPC (2020) study were used. Matched, it appearsthat the studies both predict significant enrollment declines (above 20%) between 2020 and 2030, withthe BRPC study expecting a higher rate (28%) than the MARS study (21%).Berkshire Regional Planning Commission, see Table 2, has generated three K12 enrollment studies since2015, with reports generated in 2015, 2018, and 2020. The accuracy of these studies can be evaluatedbased on predicted versus actual outcomes. From the 2015 study, the predicted 2018 versus actualonly varied by .07%, about 10 students, and the predicted 2020 enrollment was 2.1% off, with about 320fewer actual students than predicted. In the 2018 study, the actual enrollment was .37% higher thanthe predicted.Table 2. BRPC Projections, Predicted versus ActualEnrollment for Grades K - 122018Study YearPredictedActual% 2020Actual14,74814,748% Diff.2.1%.37%Based on this analysis of historical studies, the BRPC 2020 study, combined with ESE historical data, willbe used to inform the balance of this enrollment section. While no methodology is perfect, the BRPCestimates have demonstrated a degree of reasonable (one could argue high) reliability.Enrollment Trends, Overall:Overall, K12 enrollment across Berkshire County has experienced a consistent decline. Figure 1, and thecorresponding data table below, illustrates K12 enrollment change since 2000. The Berkshire decline isfour times greater than the loss of students across the Commonwealth, and is in contrast with generallyincreasing enrollments across the nation. The rate of change since 2015 has been around -1.5%/year.Figure 1. Enrollment change, 2000 through 43250,654,000% change 2000-2020-28%-7% 7%Source: 3 203.20.asp 9 203.10.asp3

Enrollment Trends, by District:Districts closely follow Berkshire population patterns. Below, Table 3 is organized by Berkshiresubregion. Three time periods were chosen (2000, 2010, and 2020) to examine patterns/trends relatedto population change in each district. Also, totals by sub-region were calculated.By subregion, population loss is relatively consistent, about 27% loss between 2000 and 2020, and 15%between 2010 and 2020. This matches quite closely to overall Berkshire losses of 28% between 2000and 2020, and 15% between 2010 and 2020.While this pattern looks somewhat balanced across subregions, there are wide variations in enrollmentdrop across districts. In the southern region, Southern Berkshire (-42%) and Farmington River (-47%)have experienced the greatest, while Lenox (-11%) has experienced the lowest decline since 2000.Central Berkshire has lost more (-37%) than Pittsfield (-24%). Finally, in northern county, the greatestlosses of students since 2000 have been felt in North Adams (-43%) and Hoosac Valley(-46%), whileMcCann and Savoy have seen slight increases.Table 3. Enrollment K-12, by district, 2000 through 2020DistrictBerkshire Hills RSDFarmington RiverLeeLenoxRichmondSouthern Berkshire RSDSouth Sub-region TotalsEnrollment K-122000 2010 20201,612 1,356 1,156154121818768196938438057501841621541,072 8656244,741 4,128 3,458% Change %-42%-28%-27%-16%Central Berkshire RSDPittsfieldCentral Sub-region Totals2,407 1,987 1,5226,721 5,930 5,1299,128 7,917 6,651-37%-24%-27%-23%-14%-16%Berkshire Arts & TechnologyClarksburgFloridaHancockHoosac Valley Regional RSD**LanesboroughMt. GreylockNorth Adams**Northern Berkshire RVTSavoyWilliamstownNorthern Sub-region CountyBerkshire 7NA-10%-30%-22%-46%NANA-43% 19% 9%NA-30%-28%72%-13%-29% 30%-31%NA-12%*-19% 1.4% ,748*Mt. Greylock (2020) reflects combination of Mt. Greylock, Lanesborough, and Williamstown**It should be noted that the opening of BART influenced enrollment of several districts, with specific impact to North Adams and HoosacValley.4

Enrollment, high school:One point of keen interest throughout the BCETF effort has been enrollment in our high schools. Thequestions of how full our schools are, whether we have enrollment and cohorts necessary to supportrobust academic programming, what programs are (or are not) being offered across our high schools,and how many high schools we need have been critical areas of focus.Below, is historical data on our regional high school, alphabetized by high school name. It should benoted that enrollment is based on a traditional 9-12 grade span, recognizing that some schools containadditional grades (6, 7, or 8). Information regarding school configuration and building capacity/usagecan be found in additional report sections.Most schools across the region have experienced enrollment loss. The exceptions include BART, whichdid not exist in 2000, and McCann, which has grown and/or maintained enrollment over the last twodecades. While the reasons for variations, including school choice and tuition, are not detailed below,enrollment losses have occurred across all Berkshire high schools. An example is Lenox, that hasremained stable since 2000 (-0.4% loss), but largely due to an influx of choice students. The greatestlosses, since 2000, have been at Hoosac Valley (-56%), Drury (-53%), and Mt. Greylock (-40%).The reductions since 2010 demonstrate similar patterns and offer additional insights on how rapidlystudent loss is occurring. For many schools, much of the population loss has been realized since 2010including Drury, Hoosac, Lee, Monument, & Pittsfield – for example.Table 4. Grade 9-12 enrollment by high school, 2000 through 2020High SchoolBARTDruryHoosacLeeLenoxMcCannMMRHSMt EverettMt GreylockPHS*Taconic*Pittsfield VocationalWahconahBerkshire 456,407265,174Grades 9-12 NA5004,836290,201% Change since2000NA-53%-56%-30%-0.4% 12%-22%-28%-40%-22.5%*2010 102%-40%-44%-29%-3% 1.4%-16%-19%-19%-21.9%NA-33%-25% 9%NA-20%-19%0*In 2000, Pittsfield had a third high school (on paper), the Pittsfield vocational school. This is added to the 2000 Pittsfield totals to calculatetotal change. Percent change (on the PHS line) represents a combination of PHS, Taconic, and Pittsfield Vocational.This data is displayed below in graphic form on Figure 2, sorted by highest enrollment declines since2000.5

Figure 2. High school enrollment change, 2000 through 2020Enrollment, by grade:More broadly, enrollment by grade, see Figure 3, sheds light on historical and future patterns asstudents move into and through the schools. Below is a snapshot of changes between 2010 and 2020,the ten-year time frame used frequently throughout this report. Only PreK and Special education/Postgrad have seen increases since 2010. All other grades report declining enrollment with the highest %decrease in Grades 11 (-25.7%) and Grade 12 (-24.3%). Lesser change has occurred in Grades 2 (-7.1%)and7 (-12.8%).Figure 3. Enrollment change by grade level, 2010 through 20206

Enrollment, schools:To carry this analysis further, we’ve included school-level change. Below, Table 4, is a list of Berkshireschools with three enrollment checkpoints (2000, 2010, 2020). The enrollment below reflects theadditions of Pre-Kindergarten and Special education Post-graduate students. School reconfiguration isnoted in places where schools were closed, opened, consolidated and, in some cases renamed. Giventhe number of changes since 2000, % changes between 2010 and 2020 were added where applicable.With rare exceptions, enrollment has declined across Berkshire schools to varying degrees, from about 2.5% (Lee Elementary) to -38% (Undermountain Elementary).Table 4. Enrollment by school, 2000 through 20207

Some schools that have lower rates of change from 2010-2020 (Allendale, Crosby) saw a dramatic dropwhen examined over a twenty-year (2000-2020) timeframe, see Table 5 on next page. StearnsElementary is one of the outlier schools that saw an increase in enrollment. It should be cautioned that% change in some of the smaller schools is highly volatile in that a few students (more or less) in a givenyear can amplify the reported change, such as in Hancock, Florida, and Savoy.Analysis of school by grade level data can become complicated in a report this size. For context, belowis a listing of school, by grade level, across Berkshire County in 2020. Districts have been color coded byunique grade span in order to compare “like” schools. While particular staffing allocations for eachgrade cohort and/or mixed grade level cohorts are not noted, this data offers some insights into relativecohort and class sizes that could prove useful in future consideration of likely school/grade levelconsolidation possibilities. For example, Farmington River has 12 students in Grade 1 (with likely 1classroom teacher assigned), while Lee has 48 (with likely three teachers assigned). If the districts wereto consider consolidation, four sections (with average size of 13) could be reduced to three sections(with average size of 20). This would, in effect, free up a full-time teacher (or the resources associatedwith that teacher) for added remedial, enrichment, or student support functions.This sort of analysis would require careful attention that would need to account for travel distances andspace, for example. However, it does illustrate how class size balancing could generate additionalefficiencies and educational resources.8

Table 5. Enrollment by school and grade, 20209

Class size, overview:Class size is a very difficult standard to evaluate in that districts/schools make site-based decisions onhow to allocate staff in providing optimal academic and support services to students. Variables thatchallenge any reported class size number include: Multi-section classrooms. Examples include several sections of a foreign language beingdelivered in a single class period. Small remedial, special education, or independent study sections. These can deflate class sizewhen averaged against larger class sections, but are needed for specialized instructionaldelivery. Enrichment, electives, and high rigor courses. These can include courses in arts, AdvancedPlacement, vocational/technical, or technology (such as programming) that are offered withsmaller numbers of students in order to deliver a broad range of curricular offerings. Flexible scheduling. In some schools, grade-span teachers work in teams, thus class size may befluid on any given day/week/month. Students are often grouped based on a tiered instructionalmodel, with higher needs students organized into smaller group instructional settings. Multi-grade classrooms. In some cases, more than one grade level will be contained in a singleclassroom.There are a few ways to understand trends in class size and allocation of staff. This analysis, recognizinglimitations, uses several strategies: Ratio of adults who work directly in the classroom (teachers, substitutes, paraprofessionals,tutors) per 100 students. Student/Teacher ratio Curated inventory of class size by two common content areas, English Language Arts andmathematics. Advanced Placement courses & electivesClass size, ratio of classroom adults to students:A relative (albeit imperfect) way to measure class size is to examine the number of adults who aredirectly assigned to classroom instruction to the number of students. It is possible to measure thenumber of teachers and long-term substitutes and paraprofessionals and tutors (as categories) and thendetermine the ratio per 100 students in the school, see Table 6. Below, this data is displayed by districtand the entire county. We recognize the limitations of this data as impacted by school size, theutilization of these roles differently across schools, and without consideration for school or grade spaneffect. Rather, this data is offered as a broad index of class size/support. An analysis of staffing levelswill be provided in a later section of this report.Both the number of Teachers/subs (9.4) per 100 students and Paraprofessionals/tutors (4.3) rosebetween 2010 and 2020, despite the fact that total staff positions and student enrollment declined.Several of the smaller schools (Hancock, Florida, Savoy, Farmington River, and Richmond) report thehighest Teacher/sub ratios along with the Southern Berkshire RSD. In contrast, Clarksburg, CentralBerkshire RSD and Northern Berkshire RVT have the lowest ratios.Farmington River, Hoosac Valley RSD, North Adams and Savoy report the highest ratios ofParaprofessionals/tutors, while Northern Berkshire RVT, BART, Central Berkshire and Lenox have amongthe lowest ratios.Overall, Hancock (recognizing its small size) has the highest classroom staffing ratios along with Florida,Savoy, Richmond, and Farmington River. Central Berkshire and Northern Berkshire RVT have among thelowest staffing ratios.10

These data suggest a need for further analysis to determine how staff are used. In some cases, decisionsto hire more (or fewer) classroom teachers rather than more (or fewer) paraprofessionals may be apersonnel strategy. Also, the unique needs of students in early grades and special education programsmay necessitate additional support paraprofessional (assistant teacher) staffing and, in some cases, bedeployed in a 1:1 basis to avoid expensive outplacement costs.Table 6. Teachers/substitutes and Paraprofessionals/Tutor per 100 students, by districtDistrictBerkshire Hills RSDFarmington RiverLeeLenoxRichmondSouthern Berkshire RSDSouth County TotalsNumber per 100 studentsTeachers & SubsParas & Berkshire RSDPittsfieldCentral County e Arts & TechnologyClarksburgFloridaHancockHoosac Valley Regional RSDMt. GreylockNorth AdamsNorthern Berkshire RVTSavoyNorth County TotalsBerkshire size, per selected commonly delivered courses, ELA and mathematics:The ESE has deepened analysis of class size over the last decade. In that class size across variouscourses and programs may vary and skew aggregated class size data, the ESE now provides tools thatallow deconstruction of class size by selected populations (gender, ELL, students with disabilities,economically disadvantaged, and race/ethnicity) and by content area.A comprehensive report of class size could be extrapolated by district, school, and a variety of selectedcategories, and we will limit our analysis to all students in two content areas, ELA and mathematics, seeTable 7. These have been selected because all students across all grades participate in these contentareas. We suggest these data can serve as another reasonable measure of class size variability.Using the Massachusetts ELA average of 17.0 as a benchmark, six districts reported larger class size(Berkshire Hills, Richmond, Central Berkshire, BART, North Adams, and Northern Berkshire RVT). In11

mathematics, Richmond and Central Berkshire exceed the Massachusetts average of 17.9, withBerkshire Hills, BART, and North Adams just under this average.Smallest average class sizes go to Lee, Florida, Savoy, Farmington River, Pittsfield, and SouthernBerkshire in ELA. In math, smallest class sizes were reported in Lee, Florida, Savoy, Farmington River,and Southern Berkshire.The state also aggregates a student-to-teacher ratio overall for each school. This is an accounting ofhow many students attend school for every one teacher. This aggregated variable, as with most classsize variables, must be considered at face value recognizing that students and teachers are grouped andassigned in a variety of ways across the schools and, in some cases, coded on state reports differently.Still, this analysis offers some additional insights.Overall, the lowest student to teacher ratio belongs to several of the smallest districts includingHancock, Savoy, Richmond, Farmington River, Florida and Southern Berkshire RSD. This continues asomewhat consistent trend of more adults per child in smaller schools. The largest class size alsobelongs to a small school, Clarksburg, with Hoosac Valley RSD, Central Berkshire, and Northern BerkshireRVT also among larger class size districts. In total, the Berkshires generally have smaller aggregateteacher-to-student ratios than Massachusetts (12.6). While this does raise questions about whetherdistricts are maximizing student grouping (class size) across the county, it does also serve as a point-ofpride when we share benefits about Berkshire public schools.Table 7. Class size for ELA, Mathematics, Aggregate, by districtDistrictBerkshire Hills RSDFarmington RiverLeeLenoxRichmondSouthern Berkshire RSDAverage class size 17.918.513.913.5AggregatedAll classes/subjects10.48.610. Berkshire RSDPittsfield18.213.618.714.712.810.7Berkshire Arts & TechnologyClarksburgFloridaHancockHoosac Valley Regional RSDMt. GreylockNorth AdamsNorthern Berkshire 13.88.817.412.811.010.512.08.912.612

Class size, enrollment in Advanced Placement (AP) courses:While AP courses will be explored in greater detail in the educational indicators section, enrollments inAP courses are offered as a proxy for elective courses class size at the high school level. Below, Table 8,is a list of all Advanced Placement courses offered in 2020. The total number of students is followed by,in parentheses, the number of sections offered at each high school. In cases where small numbers ofstudents are followed by large numbers of sections (example, AP Music Theory in Lenox) it is expectedthat course was offered as an independent study.The aim of this grid is to determine to what degree high schools are making the decision to offer lowenrolled AP courses. Clearly, there is evidence where single sections (AP Calculus at Drury with sixstudents and AP Physics with four students in Lenox) are scheduled. In other cases, two sections of acourse (AP English with 21 students in 2 sections, and AP Statistics with 18 students in two sections atMt. Greylock) are scheduled with 10 or fewer students.This signals a desire for schools to maintain their high-level academic courses, even when enrollmentsdon’t result in a full classroom. It could be argued that additional consolidation across schools couldlead to filling sections, offering more sections (to avoid scheduling conflicts) and utilizing teachingperiods saved to offer other courses/programs. For example, Lee and Lenox could combine two APSpanish classes of 5 and 7 into a single section, or Mt. Greylock and Drury could manage three sectionsof AP Statistics into two. Both situations would free up a teaching period that could be utilized to offer aunique course that expands educational breadth and opens new opportunities for students at bothschools.Table 8. Advanced Placement Enrollment (# Class sections), by high school13

Public School Attending:There are various ways that students participate in their educational experience (K12) including public,private, home schooling, etc. Below, data from 2020 shows the relative % of students across the regionbased on their schooling option.Public schools, see Table 9, continue to service about 93% of students locally, as compared with around88% nationally. Eighty-five percent of students in the Berkshires attend their public school of residence.As it turns out, choice-out (10%) may be the most common option for students seeking to attendschools other than their school of residence. About 14% of Grades 9-12 students are in Chapter 74(vocational) programs, although this is not balanced across the county with 25% of north countystudents in Chapter 74 programs, 11% in central county, and less than 3% in south county. Note:Choice, tuition, Chapter 74 will be discussed in later sections.Overall, about 1.5% of students are home schooled, and about 4% (excluding special education settings)attend private/parochial schools.Table 9. Schooling options for resident students, Berkshire CountySchooling optionResident students, in public schools, in BCAttend their home/resident schoolAttend charter (BART)Classified as Chapter 74 studentsChoice-out to another BC schoolTuitionResident students, attending schools, outside BCPublic school, out of BCHome schooledPrivate school in-stateOf these, 100 are Special EducationPrivate school out-of stateOf these, 10 are Special EducationTotal number of eligible students13,0873726531,476413149246777Number15,348% Total93%1,2047%16,552100%32Students who attend public school options outside of Berkshire County include the following in Table 10.The majority attend one of the two Massachusetts virtual schools. This data pre-dates the pandemicand it is expected that these numbers may shift in the 2020-21 school year.Table 10. Locations where resident students attend public schools outside of Berkshire CountySchoolGreenfield Commonwealth Virtual DistrictTEC Connections Academy Commonwealth Virtual School DistrictNorthampton-Smith Vocational r Valley Performing Arts Charter Public (District)Southwick-Tolland-Granville Regional School DistrictTotal# students from BC25682912932114914

School Choice:Finally, unpacking choice (free movement between districts) offers a deeper understanding of patternsand trends. It should be noted that choice is allowable via state regulation as long as districts participateand establish enrollment targets by school committee vote. Choice students/families, while able toattend school outside of the town of residence, have to provide their own transportation to/fromschool. It has been argued this creates equity issues of who does and doesn’t have access to the choiceprogram. Several studies have been conducted about choice, and reasons for family/studentparticipation in choice can be based on school/district size, perception of quality and breadth ofcurriculum, geographic convenience, and perceptions of school safety, for example.Below, Table 11, is a listing of the % of choice-in students that comprise a district’s total enrollment.Additionally, % non-resident is noted and includes tuition students who arrive in a district from a townthat does not operate its own school (entirely or by grade span).Choice continues to serve as a primary enrollment driver for several districts such as Lenox (39%),Richmond (36%), Clarksburg (34%) and Florida (27%). Districts with low choice-in numbers includePittsfield (2%), North Adams (4%) and Hoosac Valley (3%). When tuition is added to choice, total nonresident students rises among several districts including Richmond (44%), Florida (42%), and Savoy(50%). The movement of students has significant fiscal implications that will be discussed later in thisreport.Table 11. Choice and Non-resident, by districtBerkshire Hills RSDFarmington RiverLeeLenoxRichmondSouthern Berkshire RSDChoice as percentage (%) of total enrollment 36%44%13%15%Central Berkshire RSDPittsfield14%2%15%2%Berkshire Arts & TechnologyClarksburgFloridaHoosac Valley Regional RSDMt. GreylockNorth AdamsNorthern Berkshire RVTSavoy34%27%3%7%4%0%45%100%35%42%4%13%8%15%50%15

Choice by district can also be described by how many students are received and/or leave each district.Below, see Table 12, is list of how many students were received or left via school choice in 2020 and2000. The largest choice receivers include Berkshire Hills and Lenox, with Lee, Richmond, CentralBerkshire, Mt. Greylock, Florida, Clarksburg, and Savoy all receiving more students than they send out.Pittsfield, North Adams, Hoosac and Farmington River all send more students than they receive.Southern Berkshire also lost students (net) to choice out in 2020, in contrast with 2000 when theygained students via choice.Overall, choice patterns have increased since 2000, s

Enrollment Trends, Overall: Overall, K12 enrollment across Berkshire County has experienced a consistent decline. Figure 1, and the corresponding data table below, illustrates K12 enrollment change since 2000. The Berkshire decline is four times greater than the loss of students across the Commonwealth, and is in contrast with generally