USFDC Home   RSS 
Material Information
Subjects
Record Information

Full Text 
PAGE 1 ASSESSMENT OF METHODOLOGIES FOR FUNDING OF FLORIDA'S STUDENT TRANSPORTATION Final Report Prepared for F lo rida Department of Educa tion and Commissioner's Tr an sportation Funding Study Group by Center for Urban Transportation Research College of Engineering University of South Florida PAGE 2 Student Transportation Funding II PAGE 3 ASSESSMENT OF METHODOLOGIES FOR FUNDING OF FLORIDA'S STUDENT TRANSPORTATION Executive Summary The current formula used to calculate how funds are to be distributed to individual F lorida school d istricts was originally established into law in 1973. The purpose of thi s formula was to ensure the equitable distribution of funds to all school districts for the statutorily PAGE 4 Executive Summary (continued) administrative eff ort. and improved auditability; and supportive of Blue prin t 2000 Figure 1 Actua l StMewide Expenditu..s to Current Formula Allocation and Legislative Funding Level .... .... <> ............ ....... / / mo / ,,. l'9IIHO tttl)fl t Mt.tl tttMl lft.Ht lff4.15 Under contract with the DOE, the Center for Urban Transportation Research (CUTR) at the University of South Florida, in conjunctio n with the FSG, conducted an indepth review and assessment !Jf the present formul a used to allocate student transportation funds in F lorida. The purpose of this review and assessment wa s twofold: to determine i f the present funding formula equitably allocates funds and promotes efficiency of stu den t transportation services; and, if it does not meet these c riteria, to modily it or to construct an entirely n ew formula that will realize the above goa ls. The detailed res ult s of this effort are contained in this f inal report and condensed in this executive summary. Numerous analyses were conducted to attain a better funding formula than the one that is currently being employed. Myriad alternative formulas were developed and tested by experimenting with variations of the current formula This experimentation induded, fO< example, adding variable(s) to the current formula, altering the natu r e of the density i ndex by computing bus miles traveled in several different conve ntions, changing the structure of the constant and variable, and add ing a squared term to the equation. After extensive development. review. and testing, i t was found that a noticeable improvement over the current formula could not be achieved by simply altering 0< changing its form. In ligh t of this finding. ot her funding formulas were developed and tested that included variables other than the density index. I n the FSG work sessions, several decisions were made in order to clearly define a basic level of service that any proposed funding formula would have t o support. These decisions inclu ded: fund ing for summer school transportation. when provided; compensating districts for the h i gher cost of tra nsporting students with disabilities; clarilying wh ich students who live less than 2 miles from school will be eligible for transportati on f unding (all those over 2 miles will remain eligible); lv PAGE 5 Exe PAGE 6 Executive Summary (continued) The following simple mathematical ill ustration of Model 2, shown i n Figure 2, was determ i ned to provide the most de sirable results. In t he equat io n, A i s the school district base a llocation, E is the base eligible t ran s p orted s t ude nts with (ESE) supplement, a n d 8 C and D are the indices wher e : 8 the florida price level ind ex for t&ch school district; C the ad;usttd average sc:hoo4 bus Oupancy factor fot each school district and 0 the fdjusted percent of a district' s population who live in rural areas. Figure 2 Model 2: Tranq>c>rted Student Allocation Model Equation T he base ESE supplement i s critical to t he success of Mode l 2 Assigning a wei ghting factor to ESE st ude nts, due to the higher costs associated with the transportation of t hese types of students will pre vent s chool d i stricts with a small n umber of ESE st u dents (usua lly rural) from having to supp lement other school districts. Appendix D conta ins the data calculat io ns and t h e r esultin g final scho ol district f unding allocations derived f rom Model 2 In addit ion t he body of t he f i n al report conta ins a full explanat i on o f t h e data sources and me thodo logy used to construct and cakula te Model 2. Concomitant with proposing a new funding f ormula, additi onal rec ommendations resulte d from t h i s effort. They are as follows: T ransportatio n funding sh o uld remain a categorical ap p ropriat ion within the to tal educat i onal program T he e l iminat ion of th e transpo rtat ion catego rical would rad ic a l ly c han ge the distrib ution of tran.sporta t io n funding a llocated to school d ist ricts The e l i mi nation of the transportation catego rica l would s h ift f u nds away f rom schoo l districts that c urren tly t ransport a h ig h percentag e of their students toward schoo l dist ricts tha t transport a low p ercentage of their students For example, Dade s sch oo l board transports approximately 1 9 percen t of its 308, 465 students in members h i p I f fund i n g were based on membership and not eligible t r ansported students, a s i gnificant portio n of the State's finite student transportation f u nds wou l d be redi rected from t he re mai n i ng 66 scho o l districts Considerati on s h oul d be given to improved data collection a n d to the dev elo pment and i mplementation of a sta tewi de a u t omated t r ans p orta t ion routing system. T he real izatio n of t his woul d re sult i n opti m i zed routing a n d schedu l ing and would allow for the collecti o n of precise data to s upport a l in ear regressio n model which may more accurate ly predict re q uired f u ndin g levels. Student T r ansportatio n Funding vi PAGE 7 Executive Summary (continued) This final report intentionally does not contain any r ecommendation r e lated to th e level of fund ing of the public school transportation program. It was felt t hat this decision i s within t h e purview of the Legislature based on recommendations of the Governor and the Commissioner of Education. The development of a funding formula that equitably distributes availabte funds i s the prindpal achievement of the Commissioner's Transportation Funding Study Group as supported by this Center for Urban Transportation Research study. Studtnt TranJportation Funding vii PAGE 8 Student TransporUtion Funding viii PAGE 9 TABLE OF C ONTENTS Executive Summary ............ ......... ... ... . . . . . . . m Table of Contents . . . . . . . . . . . . . ix List of T abies . . . . . . . . . . . . . . x List of F igures xi Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 I. Introd uction ..... ....................... ........................... 3 statement of problem . . . . . . . . . . . . . . . . . . . . . . . . . . 3 II. The !'rese nt Fun d in g F o rmula . . . .. .... . ..... 4 overvi ew of the presen t f undi n g formula . . . . . . . . . . . . 4 w hy the current fun ding formula is falling short . . . . . . . . . . 9 Ill. Proposed F und ing Formula Model Alternatives .................... J .... ..... ... 10 the importance of recogni zing the unique characteristics of each school district . . . . . . . 10 other considerations i n develop i n g a new funding formula model . . . . . . . . . . 12 model 1: l inear regression model . . . . . . . . . . . . . . . . . . . . . . 13 model 2 : transported s t u dent allocation model . . . . . . . . . 1 4 base allocatio n . . . . . . . . . . . . . . . . . . . . . . . 1 5 base ESE suppl ement . . . . . . . . . . . . . . . . . . . . . . . . 1 5 f lor ida price level i ndex . . . . . . . . . . . . . . . . . . . . . . . 15 average school bus occupancy . . . . . . . . . . . . . . 1 5 percent of a distri ct's popul ation that is rur a l . . . . . . . . . . . . . 1 6 I V. Recommendations & Conclusions .... . . ..... . .. .. ...... ..... 16 justificat ion fo r selection of model 2 . . . . . . . . . . . . . . . . . . . . . . 16 Appendix A F S 236.083. F u nds fo r Student T r a n sportat io n (existi n g s tat ut e ) . . . . . A1 A ppendix B Current Funding Formula Calcula tions . . . . . . . . . . . .. . . B1 Appendix C Supplemental ESE Weight ing Factor . . . . . . . . . . . . . . C 1 Appendix D Mode l 2 Calculations . . . . . . . . . . . . . . . . . . . . . D1 Appendix E F.S. 236 083. Funds for Student Transporta1ion (proposed statute) . . . . . E1 Student Transportation Funding ix PAGE 10 Tabl e 1 Tabl e 2 Tabl e 3 Tabl e 4 Table 5 LIST OF TABLES P r oject Goals Realized by Mo del 1 and Mode l 2 ............... ..... .. Various Statistical Measures Associated with t h e Current F unding Formula . .. Project Goals Realized by Mode l 1 and Model 2 . . . . . . . . . . Sdlool Transportation Funding Formula cal culation Data, 199394 ............ Mode l 2 calcu lat ions . . ..................... ........ .. Studen1 Tr1nsportatJon Funding X v 8 16 83 03 PAGE 11 F i gu r e 1 F i gure 2 F igure 3 f i gure 4 Figure 5 F i gu r e 6 F i gu r e 7 F igure 8 F igure 9 f igure 1Q F i g ure I 1 F i gure 12 F i gu r e 13 LIST OF FIGURES Actual Statewide Expen d i tures Compared to Current Formula Allocation and legislative Funding leve l ....... ..... ...................... . . . Mo d e l 2: Transported S t u dent Allocatio n Mo del Equation .... ........... . . Actual Statewide Expend i tures Compared to Current Formula Allocat i on and Legislative F u nding Level ........ . . . ........................ Current Student Transportation Fund ing Mo del Equat i on .............. ... Curvilinear Relationshi p Between Actual PerStudent Cost and the Density Index .. ... ............. . ....... . . ...... L inear Relationship Between Actual PerStudent Cost and the Inverse of the Density Index (1/X) . ...... . . . ........ . . . . ....... Explanatory Power of the Densi ty Index .... ..... ... .. ..... Transported Stu dent Density ......... ............ . ... ... Per cent of Unpaved Highways ..... ... ...... . .. Average Bus Occupancy .... .... . . . . .... ....... Square Mi les of L and Area ......... ..... ........... . . ... Linear Regressio n Student T ransportation Funding Model Equation .... ..... Transported S tudent Allocatio n Model Equation .......... . .... ... Student Transportation Funding xi i v VI 5 6 7 7 8 1 1 1 1 1 1 I I 13 1 4 PAGE 12 Student Transportation Funding xi i PAGE 13 FOREWORD The Center fo r Urban Transportation Research (CUTR) at the University of South F lo rida, u nder contract with the F lorida Department of Educatio n and workin g i n conjunction with the Transportation Fund ing Study Group (FSG) established by the Florida Commissioner of Education i n September 1993, reviewed and assessed the current formula used to f und student transpo(tation i n Florida. After e PAGE 14 Student Transportation Funding 2 PAGE 15 ASSESSMENT OF METHODOLOGIES FOR FUNDING OF FLORIDA'S STUDENT TRANSPORTATION I Introduction statament of problem The current form ula used to calculate how f u nds are to be distributed to individual Florida school districts was originally established i nto law in t973. The p urpose of this formula was to ensure the equitable distribution of funds to all school districts for the statutorilyrequired t r ansportation of eligible students. Since 1973 substantial changes have taken place i n the variety and numbers of students who are taught i n Florida's public schools, the location and number of educational facilities, the demograp hics of Florida's school districts, the urbanization of Florida and its i mpact on transportation, and the variety of educational services t hat school districts provide. Thes e and a multitude of other factors have combi n ed to ren der the current student transportation funding formula obsolete a problem t hat has been consistently communicated to the Florida Department of E ducatio n (DOE) by F lorida school districts. In response, the Commissioner of Education, in September t993, established a T ransportation Formula Study Group (FSG) to determine the extent of the p roblem and to devise, if possible, a new f unding formula. The FSG consisted of representatives from tO Florida school districts. T hese representatives included s choo l d i strict assistant superintendents, finance officers, and student transportation directors, and Department of Education (DOE) personnel. The DOE personnel acted in the capacity of nonvoting technica l advisors. The FSG team members were as follows : Mr. Stan Mccall, D i rector of Operations, Broward School District; Mr. David Livingston, Director of Transportation, Clay School District; Mr. Dudley Brewton, Assistant Superintendent fo r Business, Duval School District; Mr. George latimer, Director, B udget and Finance, Manatee School District; Mr. George Baker, Director of Transportation, Palm Beach School D i strict; M r Mike Park, Director of Transportat ion, Pasco School District; Mr. John Powell, Assistant Superinten dent for T ransportation, Polk School D i s t rict ; Mr. Doug Dillon, Assistant Superintendent for Adminisuative Setvices. Santa Rosa School D i str ict; Mr. Hugh Mills, Director of Transportat ion, Suwannee School D i strict; and Mr. Wally Cox, Director, Financial and Human Resources, Highlands School District Du ring five extended work sessi ons, the FSG surmised that the inadequacies i n t h e current formula stem from its i nability to allocate f unds i n a manner that is equ i table and promotes efficient system operation. In addition, the FSG determined that the overwhelming administrative burde n, cost of collect io n a n d unreliability of the data require d to calculate t h e density i ndex (a variable i n the current funding formula), represent a serious prob lem. The FSG determined that these deficiencies were substantive eno ugh i n nature to warrant an re view and assessment of t h e present student transportation funding formula established i n Subsection (4) of Section 236.083, Florid a Statutes (see Appendix A). Studtflt Tramporu;tion Funding 3 PAGE 16 In defining its own mission, the FSG established the following fou r goals which they agreed any proposed student transportation funding formula would have to support: equitable distribution of funds; efficient delivery of student transportation services; simpl if ied and improved r e l iability of data collection, reduced administrative effort, and improved auditability; and supportive of Blueprint 2000 Under contract with the DOE, the Center for Urban Transportation Research (CUTR) at the University of South F lorida, in conjunctio n with the FSG, conducted an indepth review and assessment of the present formula used to allocate student transporta tion funds in Florida. The purpose of this review and assessment was twofold: to determine the present funding formula equitably allocates funds and promotes efficiency of student transportation services and, if it does not meet these criteria, to modify i t or to construct an entirely new formula t hat will realize the above goals. The detailed results of this effort are contained on the following pages. I The Present Funding Formula ovefVIew of the present funding fonnula According to Subsection (4) of Section 236. 083, Florida Statutes, for a student to be eligible for full or partial State funding of transporta tion services, one of the following criteria must be satisfied (see Appendi x A). The student : Jives 2 miles or more walking distance from school: is physically handicapped or enrolled i n a teenage parent program, regardless of d istance to school ; is in a state prekin dergarten program, regardless of distance from school; is vocational, dual enrollment. or an exceptional student uansported from one school cente r to anot her to participate i n an instructional program or service; or exceptional students, except gifted, transported from one designation which is a school center and provided the student's indiv idual education plan ( IEP) i dentifies the need for the i nstruct ional program or service and transportat io n to be provided by t h e school district; is an elementary school student whose grade level does not exceed grade 6, who is subj ect to statutorily defined hazardous walking conditions enroute to or from school; and is a pre gnant student or student parent enrolled i n a teenage parent program, and the c hild of a student parent, regardless of distance from school. The t h eoretical intent of the current formula is to predict as accurately as possible the allowable cost per5tudent for each district. This allowable per student cost i s then used as a basis f or the request to the Florida Legi slature for student transportat ion funds. The funding amount u l timately approp r iated by the Legisl ature i s then allocated proportionally to the districts' predicted allowabl e cost per eligible student determined by the formula mul tip lied by t h e number of elig ible transported students. The allocated funds are used by each school d ist rict to pay for Stvdent Tronsponation FuMinv 4 PAGE 17 dr ivers' S PAGE 18 The current allocation for each school district for a 18oday school term is cakulated using the algebraic equation shown i n Figure 4. The variable A (718.683599) and the constant 8 (I 27.857257) are determined by the method of leastsquares regression,' and X is the density index of the school district raised to some power, n. Figure 4 Current Student Transportation Funding Model Equation The densi ty index (X), as used i n the present formula, is the ratio of eligible transported student membership to computed bus miles traveled (0.5 times miles with students plus 0.25 times miles without students the use of adjusted bus miles traveled in calculating the density index is presumed to optimize routing and scheduling by discouraging school d istricts f rom traveling unloaded miles). The specific equation for the computation of total bus miles traveled, used as the denominator in the derivation of the density index, is mandated by Statute. Densities of 1.7 a n d 4.7 are the minimum and maximum permitted by Statute, respectively. The "transporte d students per bus miles traveled" or density approach was established to ensure that rural school districts would be compensated for the inherently higher costs associated with fewer students per square mile. In theory, and, independent of other variables such as the cost of living, number of students with disabilities. and local school board decisions regardin g the location of educational programs, a highly urban school district, such as Palm Beach, Broward, or Dade, should be able to take advantage of their high student density to achieve a much more efftdent delivery of service, as measured on a per st udent cost basis. A density of 1.7 warrants the maximum allowable e nt itlement per student and 4.7 the minimum entitlement. Statute permits the formula to be recalibrated annually by raising the density index to some power, n, to keep pace with the changing costs associated with the provision of student transportation service delivery. I n addition, Florida Statutes call for the formula to be recalibrated annually on the basis of secondpreceding year data, the most current data available. This statutory provision was put into place to ensure that the calculations would reflect the most recent operating expenditures available for each school district. As illustrated in the scattergram plot in Figure 5, the statistical relationship between perstudent cost and d ensity is curvilinear and negative, i.e., as density i ncreases (x, the independent vadable), perstudent cost decreases (y, the dependent variable). The value of the variable A and the constant Bin the present formula were derived by regressing actual school district perstudent cost and the i nverse of the density index (1/density index). T his transformation of the den.si ty index causes the relationship between actual perstudent cost and density to become l inear and positive i n nature. as illustrated in Figure 6. This procedure yields the full a lgebraic equation used to compute the allowable perstudent cost for each district. The resulti ng allocation equation ls shown above in Figure 4. The perstudent cost amounts predicted by the equation are then multiplied by the eli gible transporte d 11he method of &eastsq u ares regression inYOives fitting a sttaight lint". the leMHq vares line. through the plotted points of pe r student cost (y) and 1/dtnsity i ndtx (x} so tha t the vertical differences between atl the points and the straight line are minimized. For each obstNed point the error {difference) between the point and the is computed. This er rOt' is then squared to give p oints above and below the line the same importance; the etrots are then summed. T he "best..fiuing" straig ht line is the that mini.mizes this sum of the squared errors (SSE). The ttstscwares tine is f ined usin PAGE 19 Fig u r e 5 Curvilin e a r Relationship Between Actual PerStudent Cost and the Density I ndex "Actu:.:=::I.;:Cot;:.:;.t ,...., St, OOOT ' ' ' Density Index s 1,000 $600 F igure 6 Unear Relationship Between Actual PerStudent Cost and the Inverse of the Density Index (1/X) Actual Cost ' ' ' ' ' .. ... i" J ; ... .... .. / .. ' S200 0 .75 1j0ensity Index 1.5 st u dent membership of each school dist rict to arrive at that particular school district's predicted tota l fundin g l evel. The data used to comp u te the va ri able A and t h e const ant 8 as used in t h e pr esen t formula, and the resu lting predicted perstuden t cost for each individua l schoo l d i strict, are conta i n e d i n Tabl e 4, in Appe n dix B. T h e statistical correla tion coeff icien t, r, or the measure of linear association betwee n actual perstudent cos t (y) a n d 1/X, i s 0.56366 An r equa l to 0.56366 indicates that a r e l atively st r ong positive relat i onsh i p exists between t he two var i ables. The values of r range from 1.0 t o +1.0. The va lues of 1.0 a n d + 1.0 i ndicate t hat the plot of points falls exactly o n a downward or upward s loping l i ne, r espect i ve l y Caution s h ou l d be exercised, howe ver, when interpreting the val u e of r. W hen a relativel y s t r o n g rel atio n s hip exi s t s betwe e n two variables, as exhibited between actua l perstud e n t cost and 1/X, i t does not necessarily mean that one variab l e causes t h e o t h er. In some cases, there may be an underty;n g causal relationship t hat is impossible to determine from the value of r. For examp l e the variab les may appear to be correlat ed if t hey are bot h r e l ated to a t h ird vari ab le, such as perso n s or students per square m i l e i n this case, or bot h variables may be associ a t ed pur e l y by h appe n stance In other words, correlatio n does not necessarily e q uate t o causali ty. Howeve r a method ex ists for t e s t ing t h e n u ll hypothesis (H.: 0) that x contr i butes no information toward the prediction of y against the alternative hypothesis (H,: p, "' 0) t ha t x a n d y are l i nearly associated A close inspect i on of t h e s h a ded box i n Tabl e 2, unde r the col um n headed P rob > I T I. gives t he obsetVed s i gnificance l eve l or p.va lue of t he test statistic used in determin i n g i f a l i near re l ationship ex i s t s between x andy. Since the pvalue of 0.000 1 for 1/X is smaller t han t he of 0.05, the null hy po thesi s (H.: 1), 0) is rejected in f avor of the alternative hypo thesis (H.: 1), ;o 0). T h e r efo re, i t can be stated t h at x andy a r e l inearly re l ated or tllat x is contributing some i n formation towa r d t h e p r ediction of y One method for measuring how much informat io n x contributes toward the p red ictio n o f y is the square of t h e corre lati o n coeffident, r . T his mathematical computation y ield s the statistic known as t h e coefftcient o f determination r'. T he coeffic i ent r' measures how muc h t h e errors or variation i n perstudent cost (y) can be reduced by 1/X (x). T h e square o f 0.56366 is 0.3 1 77. T h e value s of r' r ange f rom 0 0 to +1.0; t h e cl ose r t o 1 the value i s i n theory. t he better t he m ode l s or f o rmula' s explanatory p o wer. I n t h i s case, an rl of 0.3 177 r e lates Student Transportation Funding 7 PAGE 20 that approxi mately 32 percent o f t h e variance i n y is explained by the variat ion i n x. Stated a noth e r way, 68 percent of the variance i n perstuden t cost is l eft unexplained by the current model. Thi s s i tuatio n is clarified i n F i gure 7 Let the c ircle stand for t h e total variance of y, t his variance being 1.00. By grap hical ly illustr ating t h e r e l ationship betwee n x and y i n this manner, the expl anatory power of x o n y can be visually depicted and hence. understood with greater ease. In t h e c ircl e below, let t h e crosshatc h ing represent t h e variance i n y accounted for by x. The crosshatched portion of t h e ci rcle visually relates that onl y 32 percent of the variati on in y is accounted f or by t h e d e n sity index. While this is an artificial and contrived representation, i t has the virtue of demonstrating t h e associ atio n between the variables, and thus v isually illustrates the statistical concepts articulated above. In add ition, Tabl e 2 p rovides a variety of other statistical measures pertaining to t h e s impl e l inear association between actual perstu dent cost and t h e density index. Rgure7 Explan at...y Power of tho D ensity I nde x Unexplained Variance 68% Tota l Variance 1 0 0 Tabl e 2 Vario u s S tatistical Meas ures Associated wi th the CUrrent Fundin g Formula Dflcriptive SUtirtla Variable Meon Unit COSt 446.5 1/X 0.44 Regressio n Analysis Statistics Soutee Model trot C Total Constant Varib!e OF 1 65 66 Root MSE c.v O F I S t d Oev. Minimum 100 0 9 1 2 29 .22 0.08 0.3 Sum of Squares Mean Square 210050.2 3 2 1 0050 .23 451073. 1 1 6939 59 661123.33 83 .31 1 8.66 rsquar e Patam ete r E s t. Stan datd E rrot 127 857257 58 8 1 118 663599 130 .63 Maximum 1862.36 0 59 F Value Frob> f .30. 29 0.3012 0 56366 0 .3 17? T for H, : 2.114 5 502 Student Tra nsportation Funding 8 Ptob > ITI 0.03 3 3 0.0001 PAGE 21 why the cummt funding fotmul is falling short Under any fundin g formula system, especially one b."!sed on simple r.near (one independent variable) or mult iple linear regression (more than one independent variable) analysis, in order for the funds to be allocated equitably it is crucial that t h e independent vafiab l e(s) inclu ded i n t h e model (formula) parallel the inh erent characteristics that are drivi n g sch oo l distri ct operati n g costs t o the greatest extent possi b le, i.e.. in this case. possible i ndependent variables fo r inclusion i nto the mo del could be the magnitude of a dist r ict' s rur a lization and the number of students being transported. It has been shown in prior studies that have investigated this issue that r.nea r density, as used in the present formula is the single best predictor of school district per student =t. All irHl e pth review of the pr5ent formula (refer to preceding however indicates that th e density index is no long e r lhe single best pred'JCtor of perstudent cost. Given lhe everexpanding and complex nat ure of student transportation senrice delivery, finear density Is insuffic i ent by itself to account for the multitude o f factors that a r e driving school district operat ing costs. Stated dir e ctly, i t has outlived its usefu lness as a predictor of S<:hool district perstudent cost. I n a d d ition, and perhaps more i mportantly, the FSG determined, given the lad< o f r e l iabl e data, that a cost r e imbursement f undi n g formul a base d on least5quares regression analysis cannot p r o duce an equitable system for allocating f unds It was the i r co n s ensu s that the c urrent met hod of S tate fund ing does not provid e incentives for an efficient and comparable level of Statefun ded service among simi lar school districts. Moreover the fSG determ ined the need to remove the density index a ltogether, regardless of the m ethodology devised, from the determina tion of fund allocation due to the administrative burden and the unreliability and cost of data coftection associated with its calculation. What, then, are the i nternal and exwnal factors associated with the unexplained variance of 68 percent pertaining to school district per"5tudent cost? This unexplained variance i s due. i n part, to the fact that a single variable is being used t o expl ain the operating cost diffe r e nces between Florida's school districts that are being cau sed by a mul ti t ude o f f actor s Among these factors a r e th e increased complexity o f providin g t ransportation services for magnet schools, s tudents wit h disabilities and oth er special programs, changes resulting from m od ifi cat ions in the school year (i.e .. year ... ound educatio nal programs), summer school, mid PAGE 22 I Proposed Funding Formul a Model Alternatives Numerous analyses were conducted to attain a better f u nding fo rmula model t han t h e one t hat is currently being employed. Myriad alternative models wer e developed and tested by experimenting with variations of the current formula. This experimentation included, for example, adding variable{s) to t h e current formula, altering the nature o f the density index by computing bus m iles traveled i n several different conventions, changing the structure of the constant a n d variable and adding a squared term to the equation. After extensive devel opment, review, and testing, it was found that a noticeable improvement over the current formula coul d not be achieved by simpl y "ering or changing its form. In l i ght of this f inding, other formula model s were developed and tested that included variables other than the density index. I n order to develop a f unding formu l a mode l that does not rely solely upon or include the density index, an extensive database that contained 103 diff erent variables was created. This database consisted of a host of relevant economic, geographic, demographic, sode>onomic, and education related variable s pertaining to each individual school district For example, t h i s database included i nf ormation fo r each school district such as the fo llowing: popul ation density; transported student density; ratio of the number of eligible transported students to county popula tion; percent of unpaved highways; lane miles of highway (paved and unpaved); h i ghway density; p ercent of populatio n that is rural ; n umber of schools per square mile; b uses used in daily service; number of special needs students transported; Florida pric:e Jevel index; number of teen parent sites; number of bus drivers; percent of populat i on that is school age; average bus occupancy; square miles of service land area; and the Florida transportation price level index the importance of recognizing the unique charactflristics of each school district To construct a funding formul a model t hat will simultaneously allocate student transportation funds i n a manner that is equitab l e, promotes efficient system operation, and i s r elatively simpl e to calcul ate is no small task, g iven the fact that stu dent tra n sportation systems in Florida operate under vastly d i fferent condi t ions. For examp le, some school districts are 1 00 percent rural and others are highly urbanized; some school districts have a higher percentage of unpaved roadways than others; some school districts t ransport more specialneeds students than others; some school districts have to o r choose to pay more for labor than others; and some school d istricts have to pay a h igher cost for fuel and other vehicle maintenance expenses. These and other unique school district Student Ti'lnsportation Funding 10 PAGE 23 characteristics have an unavoidabl e effect on the costs of p roviding student transportation servi<:es. In order to build a fundi n g formula model that will equitably allocate funds and promote effi d ent system operati on it ls paramount t hat certain u n i que characteristics of each i ndividual school di strict be taken i nt o account by any proposed fund ing formula. To illust r ate the above poi nts, for example Figures 8 th r ough 11 show statewide extremes f or a variety of school district operating condit io ns, represented by t h e g ray shaded bars, in comparison to each district's perst u dent cost. represented by the black bars. In addition, fo r purposes of comparison as well, the statewide average perstudent cost is provided. The graphs show that Glades has th e lowest number of e l i gible s t u d ents tra nsported per square m i l e of lan d area (0 .52) and Pinellas has the largest (159), a rat io of 305. 7 to 1. Wi t h regard to the percentage of unpav e d h ighways, Wakulla has the h ighest and Dade has the lowest, 81.3 per cent to 0 .66 percent, respectively. Further examination of th e g raphs i llustrate th at Collie r has the most square m i les of s ervice land area to cover ( 1,994) and Pinellas has the fewest (280). Theoretically, t h e densely populated districts s houl d be able t o achieve a much more efficient oper ation, as measured o n a p e rst u d e nt cost basis, due to the f act t h at they have to travel fewer miles to transport students than the sparsely populated districts, a geographic advantage that s hould greatly reduce their per1tudent operati n g cost. This i s illustrated by F i gure 10, wh ich compares the average bus occupancies (ABO) of the Hamilt o n and Volusia school districts with their perstudent cost of opera tio n usin g the most recent data (199 1 92). The graphic clearly illustrates Vol usia's l ower perstudent operating cost. Given all of these operating constraints it is clearly u nfai r to compare districts i n Aorida u nless certain districts are compensated for their relative i n h e r ent geographic, social a n d economic characteristics. Figure 8 Transported Student Density Hamilton Pinellas Figure 10 Bus S PAGE 24 other consitkratlons In developing a new funding fonnula model In the FSG work sessions, several decisions were made in order to clearly define the basic level of service that any proposed funding formu l a model would have to support. These decisions included the following: accommodating the fun ding of summer school transportation, when provided; compensating d istricts for the higher cost of transporting students wit h disabilities; clarifying .which students who live less than 2 miles from school will be eligible for transportation funding; eliminating the doublecounting of students; accommodating varying term lengths including year round schools; providing incentives for efficient utilization of equipment; compensating for the varying cost of living in different school districts; a n d compensating for geographic constraints beyond school d i strict control which affect the costs of providing student transportation services. The FSG decisions. as listed above, are detailed in the following paragraphs. Summer school and varying term lengths, induding yearround schools, should be accommodated by annuali z ing the number of students transported to a 180 PAGE 25 population that is rural. as defined by the 1992 Florida Statistical Abstract. s hould be used as t h e appropriate measure of school district rural d e nsity. Based on t h e directio n p rovided by the FSG, two different f u n d ing formu l a model s were developed. The first model i s based on multip l e l inear regression that follows the logic o f the present f u ndin g formula. l ike the current formula, this model l s a costreimbursement model based on secondprecedingyear expenditures. The second model i s a straight matht>matical regression based allocation model that provides a basic statewide level of f unding per s tudent (depending on legis lative app r opriat ion), adjusted by indices wh ich reflect certai n uniqu e characteristics o f each school d istrict. 1: Unear Regressio n Model With the aid of a computer and the stat istical s o f twa r e package SAS ( Stat i s tical Application Software ) several diff e r e n t regression models were developed and tested using a stepwise regression analysi s procedure at the 0.05 s ignificance level (ex 0 .05). The small signif icance level o f 0.05 was specified to guard against the inclusion of any variables t h at do not contribute to the overall predictive power of the model. S tepwise regression performs regression i n "steps," i.e., i t substitutes each variable, forward and b a c kward, one by one, in an a tt empt to f i n d the "bt>st' model. It is a usef u l and vali d statistical procedure. particul a rly whe n trying to identify a f ew sign ificant variables t hat should b e included in t h e model from a m u ltitude of possible variables. Expressly, f r om the myriad variables included in t h e SAS dataset (1 03 variables). the stepwise p rocedur e iden tifie d the following variables for inclusion i n Mo del 1: eligibl e transported students as a percent o f t otal enr o llment: F lorida price level index; and an average school bus occupancy factor. After extensive review and testing, i t was found that the follow ing s impl e linear regression model. shown in Figure 12, produced the most desirable resu lts Figure 12 Unear Regression SWden1 Transportation Funding Model Equation c PAGE 26 x, el'ugib&e transported students as a peou:ent of tota l enrollment for each school district ":t t h e price lewl index for each school district; and x ao age school b u s occul)lnc:y facto r for each school district. The use of the regression ana lysis p r ocedu re developed a simple l i near model that real i zed t h e followi ng majo r project objectives: a model was generated that i s useful for prediction, one that has a h igh r' (0.67) a n d is high l y sig n ificant (pro b > F 0.0001); a mode l was p rovided that i s e<:onomical i.e., one that uses on l y a few i n dependent variables, is simp l e to calcu l ate, r ed uce s data collect ion, and utilizes data that is easily obtai nab l e and reliable; and a mode l was provided that considers to the extent practical given the avai l ability of data, certai n un ique character i stics of each school district Model 2: Transported Student Allocation Model At the reques t o f t h e FSG, after extensive revie w o f Model 1 and the rea l ization of i ts vul n erabilities g i ven t h e i nherent prob l ems associated wit h manua l data collection, a second funding mode l was deve l oped t ha t utilizes a uniform sta tewide base allo c ation per transported student as a foundation. The base allocation for each district was derived by mul tiplying t h e total statevvide a l locat i on b y the perce ntage of total statewide transported student s w i thin each resp ective sc hool district The school d i strict bas e allocation is then multiplied by variables that re f le<:t certa i n un ique characteristics o f each school d istrict Expressly, the FSG iden tifie d the following t hree factors that are to be appl ied to t h e base allocation T h e three factors are as f ollows: F lorida price l eve l index; ave rage schoo l bus occupancy; and percent of rural pop u la ti on The foiJo wing s imple math emat i ca l equation shown i n F i g ure 13, was determined to provide the most desi rabl e results I n the equatio n A i s the sc hool dist rict ba se allocation E is the base e ligibl e t r ansported students w i th di sabilit ies (ESE) supp l ement and 8, C, and 0 are the indices w h ere : B t h e Florida price index for eac h school district; C the adjust ed averag e sch oo l bus OCOJpancy factor for each school district; a n d D the adjuS1ed perCfflt of a district's population that is rural. Figu re 1 3 Transported Student ARocation Model Equation Ftmil iflg alloauitm (A X B X c ll: D) (E B c J( IIJ Student Transporta t ion hlndlng 14 PAGE 27 The variab les l isted on the previous page, as used i n the calculation of the transported student allocation model and displayed i n F igure 13, were derived as follows: Base allocation ( varia ble A) Step A1: Step A11: Step A111: S tep AIV: Determine the number of eli gible transported students by cal culating the average number of students transported during the October t 992 and February t 993 survey weeks. This t otal includes unweigh ted except ional (ESE) students. Annualize students (including E S E ) attending summer school and multiple programs as discussed earlier. Determine the sum of steps A1 and I above for each school district and fo r the state. Divide each school dist rict's transported membership by the statevvide total and multiply by the tota l statew ide dollars a llocated i n 199293 ($232,364,265). Base ESE supp lement ( va riable E) Step E1: Step E 11: Step E 111: S tep EI V : Step EV : S tep EVI: Deter m ine the total statewide number of annualized transported students (induding ESE) for 199293. Same as step A111 above. Determine statevvide total dollar expenditures for stud ent transportation fo r 199293, from Annual Financial Report, O b ject 7800 ($423, 7 88,355). D ivide step Ell by step E 1 to determ ine the average expendi ture per transported student for 199293. App l y the same methodology used in t h e base allocatio n (variabl e A) above to iden t ify the number of annualized transported ESE students f or each school dist ri ct and fo r the state. Determine the number of we i ghted ESE students by mul t iplying s tep EI V above by the supplemental ESE wei ghting factor of 1.8 (see Appendix C) Determine the base E S E supplement for each district by m u ltiplying step EV by step Florida price level index (variable B) The natural log of t h e most recent yeafs FPL I (1993) was util ized i n Mode l 2 The n atura l log was used to normalize the range in the data. This data transformation is a common statistical practice when using economic data. Average school bus occupancy (ABO) (va riabl e C) ABO is a measur e of the average number o f eligible students transported per day per bus u s ing data obtai ned from the October 1992 and February 1993 transportation surveys. The data were indexed at a range of 0 .90 to 1.10 i n orde r to moderat e the negative effects of the w ide variatio n in ABO amon g Florida school districts Student Tr1nspol"tttion Funding 15 PAGE 28 Percent of a district's population that is rural (variable D) T he percent of a school district's population that is ru r a l was obtained from the 1992 Florida Statistical Abstract These data were in dexed at a range of 0 .90 to 1 10 (same as t he range tor ABO) in order to moderate the negative effects of the wide variation in t he percent of a schoo l district's population that l ive in areas def i n ed as rural. I Recommendations & Conclusions The two propose d f un d ing formula models included in this report were developed with t he i ntention of achieving the followi ng four major pro j ect goals : equitable distribution of fun d i ng; efficient delivery of student transportation services; simp lified and i mproved rel i abi l ity of data collectio n reduced administrative effort, and improved auditab i l i ty; and supportive of B l u ep rint 2000 The overrid i n g conside r ation of the FSG i n gu i ding their recommendation of a f unding formula model was to promote flexibility at the local level by a llowin g school districts to provide a chose n level of service t ha t is not at t he expense of other school di stricts and with a mi n im um of fundi n g rules or constraints at the State leve l. Table 3 provides a sidebysi de comparison of t h e project goa l s realized by both of the funding formu l a mode l s i ncluded i n this report, Mod e l 1 being the multiple l i near regression model and Model 2 being t he t r ansported student allocation model. As the table makes dear, Model 2 satisfies all of the p r oject s goals. Therefore, it i s recommended tha t a new student transportation funding formula model be adopted that is based on Model 2. T able 5, i n Appendix D, contains th e r esulting funding allocation for each school district and a variety of othe r information pertaining to Model 2. In addition, Appendix E contains th e p r oposed revi sions to Subsection (4) of Section 236 083, Florida Statutes. Table 3 Project Goals Realized by Model 1 and Model 2 Project Go als Equi table distribution of funds among districts Pro m otion of efficiMt delivery of student tran sportat ion services a n d inJp rO\'ed q f data reduction of admn iSt rabOn, and 1mptoYed auditabllity Support of Btuepri n t 2000 jurtifia>tion for Sf!lection of model 2 Mod PAGE 29 up to 10 percent above the statewide perstudent reimbursement. Conversely, a district with a lo w ABO will be f unded at a level as low as 10 percent be low the state average. This process promotes cont inuing efforts to improve etfident t r ansportation service del ivery. An example of an efficient district is Volusia, whic h effe PAGE 30 transportation program. I t was felt that this d ecisio n is within the purv iew of the legislat ure based o n recommendations of the Governor and the Commissioner of Education. The development of a fund i ng formu la m odel that equ itably d ist ributes available funds is the prindpal achievement of th e Commissione(s T ransp o rta tio n Fund ing Study G roup as supported by t h is Cenler for Urban Transportatio n Research study. Student Trenspottation Funding 18 PAGE 31 APPENDIX A F.S. 236. 083, Funds for Student T r ansportation (existing statute) Student TranSf)Orutlon Fund.lng A1 PAGE 32 Stvdetrt Tta ns:potUtion Flolnding PAGE 33 Existing Statute The annual allocation to each district for transportation to public school programs of students i n membership i n kindergarten through grade 12, in migrant and exceptional student programs below kindergarten, and in any other statefunded prekindergarten program shall be determ i ned as follows: (1) Subject to the rules of the state board, each district shall determ i ne the membersh i p of students who are transported: (a) By reason of l iving 2 miles or more from school; (b) By reason of being physically handicapped or enrolled in a teen age parent program, regardless of distance to school; (c) By reason of being in a state prekindergarten p rogram, regardless of distance f rom school; (d) By reason of being vocational, dual e n rollment, or exceptional students t ransporte d f rom one schoo l center to another to participate in an instructional program or service; or exceptional st udents, except gifted, t ransported from one designation to another i n t he state, p rovided one des i gnation is a school center and provided the student's individual educational p lan (IEP) identifies the need for the i nstru ctiona l program or service and t ransporta t i on to be provided by the school district. A "school center" i s defi ned as a public school center, p ublic community college, p u blic or other faci l ity rented, leased, or owned and operated by the school district or anothe r public age ncy. A "dua l enrollment st u dent i s defined as a p u blic school student in in both a pub l ic secondary school p rogram and a pub l ic community college or a public program under a written agreemen t to part i ally ful fill ss. 229.814 and 240.115 and earning f u ll time equivalent under s 236.081(1 )U); (e) With respect to elementary schoo l students whose grade leve l does not exceed grade 6, by reason o f being subjected to hazardous walking conditions i n rou te to or f rom school as provided in s. 234.021. S u ch rules shall, when appropriate, provide for the de te rmination of membership under t his paragraph for less than 1 year to accommodate the needs of students who require transportation only until s uch hazardous conditions are corrected; and (f) By reason of bei ng a pregnant student or student parent, a n d the c h i l d of a student paren t as provided ins. 230 2316, regard less of distance f rom sch ool. (2) Subject to the r egu l ation of the state boa rd each district shall determine a n d report onehalf of the roundt ri p route mileage requ ired to transport students to and from schoo l and onehalf of t he roun d t rip mileage on routes between sch oo l centers required to transport exceptional students and vocational students to and from centers where appropriate programs are p rovided. Onehalf of t he ro un dt ri p ro u te mileage shall be comp uted by add i ng : (a) The loaded mi les of each school bus route as designated i n accordance with s. 234.06 1 and served by a bus as defined by regulations of the state board, except that miles traveled for a side route to pick u p students l ivi ng within 1 1/2 miles of the mai n trunk ro u te and mi les trave le d to transport students to evening schools and enrichment programs shall not be added; and (b) f ifty percent of the mi les of the bus route trave led without s t udents (3) A density indo. for each d i strict shall be computed by the depa rtment annually by dividing the membership of tra nsported students as de term ine d i n subsect i on (1) by the bus r oute mileage as de term i ned i n subsection (2). (4) T he allocat i on for each d is trict for a 180day school term shall be calcu l ated i n accordance vvith t he follovving fo r mula: Student Transportation Funding A3 PAGE 34 in which A, B. and n are constants the values of which shall be computed by the method of leastsquares based on the variables i n transportation expenses of salaries. employee benefits, purchased services, materials and supplies, and other expenses, and th e cost of replacing 10 percent of buses per e l i g i ble transported membership The density index s ha ll be computed as specif ied i n subsection (3), except that the districts with a density index of 1.70 students per route m i l e or less will be computed as hav i n g a density index of 1.70 and d i stricts with a density index of 4 70 or more students per route mile will be counted as h avi n g a dens ity index of 4.70. T his formu l a shall be r ecomputed annually by the Department of Education on the basis of the l atest available data for the dens i ty i ndex and from the second p receding year for constants A and B The allocation to each district fo r transportation shall be determ i ned by multiplying the allowable perstudent cost by the membersh i p of a ll students who are transported as dete r m i ned in subsection ( 1 ) (5) I f i district operates schools more or less than 180 days. the allocation per student for transportation to such schoo l s shall be cakulated by multiplying t h e quotient of the days the schools operate divided by 180 days t imes the allocation per st udent determ i ned i n subsection ( 4 ). T he allocation for each district for transp o rtation o f students in membership more or less t han 180 days shall be d ete r m in ed by multiplying t h e allowable cost per student determined i n this subsectio n by the members h ip of such students who are transported. (6) When author ized by rules of t he state board an allocat ion of th e maximum amount specif i ed i ns. 112 061(7) per mi l e shall be allowed for m i les traveled by passen ge r car.; or boats providing for transportation of isolated students, o r physica lly hand icapped students as defined by ru l e when more efficient than a schoo l bus (7) The tota l alloca t ion to each district for transportat i on of stu d ents shall be the s u m of the amounts determined i n subsect i ons (4), (5), (6), and (9). I f the funds appropria t ed for the purpose of impfementing thi s section are not sufficient to pay the requirements i n full, the Department of Education shall pro r ate the available funds on a pe r centage basi s (8) No district s h all use funds to purchase transportation equ ip ment and sup p l ies at prices which exceed those determined by th e department to be the l owest which can be obta ined, as prescribed i ns. 229 79. (9) Fun ds allocated or apportioned fo r the payment of student transportat i on serv ices may be used to pay l oca l genera l pu rp ose transportatio n systems for lransportatio n of students to and from school. Subject to t he ru les of the State Board o f Educat i o n each sch o ol d i st rict shall determine and report the n umber of assigned stude nts u s in g gene r al purpose transportation to transport students to school for the f i rst t im e o n any sch oo l day an d t he oneway mi les on routes between schoo l centers r equ ired t o t r ansport exceptional stude n ts and vocatio n al students to centers w h e r e app r opriate programs are prov i ded The alloca t ion t o each district for such trans portation s h a ll be determ i ned by mu l tiplying t he allowab l e cost per stude n t b y t he membe r s h ip of students who are transported by g eneral purpose transportatio n The allowable cost per stu d e n t s h all be equal to t h e roundt ri p fare charged by the ge n eral p urp ose transportat i on system or the allowable cost pe r stude n t r i d i ng a schoo l bus. w h i chever is less. Hi s t ory. s 6, ch. 73345; s. 4 ch 74227; ss. 12, 14 ch. 75284; s. 3, ch. 7 8104 ; s 1 ch. 78 128; s 3 ch 812 54; s. 4, ch. 861 46; s 8, ch 89101; s. 9, ch. 89 278; s 22. ch. 89379; s 1 5, ch 901 72; s. 57, ch. 90288 No t e Substiluted by the editor.; for a reference to s 236.081(1)( h ) to conform to t h e r edesignatio n o f s u bun i ts by t he rev iser inc i dent t o the compilation of s 17, ch 89298, and s 60, c h 89381. Student Transportation Fundlng A4 PAGE 35 APPENDIX B Current Funding Formula Calculations Student Transportation Funding PAGE 36 Stvdont T...,.portatlon fUnding PAGE 37 Table 4 School Transportation Funding Formula calculation Data 199394 District Actual Cost Eligible Predicted Cost Density of per Student Students per Student o.as; ty ( 1 /XJ U12.48 380 l44U4 2.287 OAl1 ...... $425.28 1835 s.,., 1.932 0 518 OAV U91,,. \1160 S"Uf ..... ... OAADfORO muz .... SJ40 .09 ..... 0 .295 8MVAAO SS67. 81 ,..., 1$04._$9 .... 0.524 BROW.tJIO S4S4.13 = 1.501 .... (AVfQ,.N $369.40 1ll4 ...... LS81 (1,387 """""' 5384.93 !317 S40UZ U.l6 (1.38l (ITM ...... ,. .. ,. ..., .. U)l (1, 3)9 ClAY Uc1 04 11102 ....... 2 .SS1 0 .391 COUI(A 10530 $C81.3 8 2.033 COlUMI!A $4CU2 5)11 S41U2 2.462 0 .000 """' S86U6 59627 SS49.83 1.703 .. ,., OESOTO S468.14 2135 1416. 1 0 2.49) O . PAGE 38 District Ar.cnkll Cost Eligi ble Transport ed Predicted Cost Oensity f n\'trse of pe:r Studef"'t S t udtnts per Student Density ( 1 /X) """""' ....... 11400 So&IS,)8 2.500 ..,. MAAIOH S)H.74 181.)1 S4SII. 24 U<40 O .SIS ....,.. 1470, 7 5 .... U18.St 2.866 ..... ..,._ SS .:W ..,. s.to3. n .... .... """"" U7US ... U69.SS U74 0.))6 0""00SA USI.Ol ,,.,. SJ83.2S 2.1'4 O )SS OKlEOtOIIU i6M. n "" $ PAGE 39 APPENDIX C Supplemen ta l ESE Wei g h t in g Factor StiMMnt Transportation Funding Cl PAGE 41 Supplemental ESE Weighti n g Factor T h e supplementa l ESE weight in g factor of 1. 8 was derived by the FSG to reflect the in h e r ently higher cost associated wit h providing transport to students with special needs. The we i ght i n g facto r of 1.8 incr eases the pe r student transporta t ion allo PAGE 42 C4 PAGE 43 APP ENDIX D Mode l 2 Calculations Student Trans:porutlon Fund:fng D 1 PAGE 44 Student Transportation funding D PAGE 45 1 J II I! I I; li I; l i Ia f i!!. PAGE 46 'i ... f i a I h PAGE 47 1 ,., I ... l q :::! .. Jtritttr1r1ritttrriittr1r1ttttrritttr1r1riti f l j PAGE 48 g c: ]l "' {! PAGE 49 APPENDIX E F S 236.083, Funds for Student Transportat i on (pr o posed statute) Student Transportation funding E1 PAGE 50 Student Transportation Funding E2 PAGE 51 ProJ>Osed Statute 236.08 3 F un ds for student transportation.The annual allocation to each district for transportation to public school programs of studen t s in membership i n kindergarten through grade 12, i n migrant and exceptional student programs below kindergarten and in any other stat .. f unded prek i nde rg arte n program shall be determined as follows : (1) Subject to the ru le s of the state board, each d ist ri ct shall de termi ne the membership of students who are transported: (a) By reason of l iving 2 miles or more f rom school; (b) By reason of being students with djsabjlitjes or enrolled i n a teenage parent program, regardless of distance to school; (c) By reason of being in a state pre ki nde r garten program, regardless of d istance from s choo l; (d) By reason of being vocational, d ual e n rollment, or steeAts students wit h disabilities transported from one school center to anot her to part i cipate in an i nstructiona l program or servi ce; or eMEeJ!ItieAal W:tdeAts, e"'ee,t liked students wjth d isabilities. transported from one designation to another in the state, provided one designation is a school center and provided the student's in d ivid ual educat i onal p lan (IEP) i dentif ies the need for the instruc t ional program or service and t r ansportation to be provided by t he school dist rict. A 'school center" i s def i ned as a public sch ool center, p u blic community college, public un iversity, or other faci l ity r ented, leased, or owned and operated by the school district or anot h e r pub l ic agency. A dual e n rollment student' is defined as a public schoo l student in membership in both a public secondary school program and a pub l ic community college or a public u n i versity program under a written agreement to partially fulfill ss. 229.814 and 240.115 and earning f u lltim e eq uivalent membersh i p under s. 236.081(1)(j); (e) W ith respect to e l ementary school students whose grade level does not exceed grade 6, by r easo n of being sub j ected to h azardous wal king con dit ions enroute to or from scho o l as provided i ns. 234.02 1. Such rules shall, when appropriate, p rovi de for the de term inat ion of membership unde r this paragraph for les s than 1 year to accommodate the n eeds of students who require transportatio n only u n til such hazardous conditions are corrected; and (f) By reason of being a pregnant student or student parent, and the chi l d o f a student parent as provided ins. 230.2316, regard less of d i stance from school. (2) 