CHAPTER 4 Conducting a Trip Generation Study Pater A local jurisdiction may wish 60 conduct its own trip generation seudy to validate use of ITE Trip Generation rates or equations in its ‘community, establish its own rates reflecting unique conditions found in that community, or establish rates for land use types not included in Trip Generation. A state or province may wish to investigate trip generation rates in detail for land use types of partic- ‘lar concer in its jurisdiction. Consultants, ITE districts, sec- tions, or individual ITE members may want to supplement the ITE national database on trip genera~ tion. ‘To maintain consistency with TES nationally recognized database and procedures, local studies should fol- low procedures consistent with those described below. However, it is recognized that local jurisdictions ‘may need to tailor the process to meet the specific needs of the com- ‘munity and the characteristics of the sites being studied ‘To enhance the national database, ITE encourages the submittal ofall new trip generation data. Sample data collection forms for reporting the information are included at the end ofthis chapter. These forms should be used whenever possible. Reasons to Conduct a Trip Generation Study 42 ‘The general purpose ofa trip gen- eration study is to collect and ana- lyze data on the relationships between trip ends and site charac- teristics fora particular land use. Before initiating the study, is spe- cific purpose should be identified. ‘The specific purpose will help the analyst target the characteristics of the sites, the data to be collected, the number of sites to survey and the analysis t be conducted. @ Ifthe description ofa site is not covered by the land use classifications presented in Trip Generation, the analyst should collect local data and establish a local rate. © If the site is located in a down- town setting, served by significant public transportation, or isthe site ‘of an extensive transportation demand management program, the site isnot consistent with the ITE data and the analyst should collect local data and establish a local rate. (@ If the size of a site is not within the range of data points presented in Trip Generation for the land use, the analyst should collect Jocal data and establish a local rate. # LF the Trip Generation database has an insufficient number of data points, the analyst should collect local data and establish a local rate. $ Ifthe Trip Generation database produces curves with unsatisfac~ tory standard deviation or regression coefficients, the ana- lyst should collect local data and establish a local rate. IF local circumstances (eg. age of residents, worker shifts, other differences in independent vari- ables) make a site noticeably dif- ferent from the sites for which data were collected and reported in Trip Generation, the analyst should collet local data and estab- lish a local rate. If te site is a multi-use development, the analyst should refer to Chapter 7 in this handbook for guidance on special data collec- ‘ion and analysis efforts required for multi-use developments. Tip Generation Handbook, 2nd Edition Chapter 4m WTE 18.$ IF the applicability or validity of TTE Trip Generation data for local use is questioned by traffic pro- fessionals or local officials, the analyst may need to collect local data and either validate the national data or establish a local rate IF ivis desirable to establish tip {generation characteristics for a land use not included in the current edition of Trip Generation, the analyst should collect and analyze the data for focal use and submit the data to ITE. Trip Generation Study Design “Trip generation study design should include the land use to be surveyed, number of survey sites, selection of appropriate sites, sur- vey period, independent variable data to be compiled and traffic counting methodology. 43 Information is often available from analyses undertaken in either the same jurisdiction or other jurisdic- tions. In planning the local study, reviewing existing data is helpful in determining issues that may be ‘encountered and identifying ‘expected results. Also, because existing data may be integrated into the local study-co reduce the amount of new data that need to be collected, it is important to have prior knowledge of the availabilty and procedures used to collect the data. Selection of Land Use to Study ‘Tip generation studies should be considered to supplement the Trip Generation database when the fol- lowing conditions apply: @ Land uses of local interest for which ITE Trip Generation presents liede or no data; # Local land uses that do not fit into existing TTE land use classifications; ‘@ Land uses that are more specific than the generalized land use cate- gories in Trip Generation; ‘@ Land uses for which the range in development size in the Trip Generation data plots does not cover the local range in development sizes; or Land uses for which local trip generation rates are theorized to be substantially different from those in ‘Trip Generation. Sample Size Determination Sufficient sample size is necessary to enable the analyst to draw valid conclusions from the trip genera- tion study. Hlowever, no simple statistical methodology has been established for determining the smumber of sites that should be studied to obtain statistically sig- nificant trip generation results. In reality, tip generation is influ- enced by far more than one or two 16 FTE i Tip Generation Handbook, 2nd Edition Chapter 4 independent variables. Asa result, significant variation of individual sites from the weighted average rate or regression curve is fre~ «quent. Common practice in the traffic planning industry has been to collect trip generation data at three to five sites that truly meet the recommended site selection criteria withthe assumption that these data will yield a relatively stable sample. local trip generation rate, itis rec- ‘ommended that at least three sites (end preferably at least five) be sur- veyed. The higher number is sug- gested because it will enable the analyst to more readily identify — and potentially discard—outler val- ues and to produce a more reliable estimate of local trip generation characteristics. Iti recognized, however, that budgetary constraints and perhaps even the lack of suit- able survey sites may limit the trip generation study to three sites. rn n nner nnn An Ar AANA NRA MA AA TAA AAAAAAAgTIF the analyst intends to validate ‘Trip Generation data for local wse, it is recommended that no fewer than three sites be surveyed. Ifthe ana- lst intends to supplement the Trip Generation data with local data and produce a consolidated trip genera- tion rate, itis recommended that at least two sites be counted. ITE will accept data from one site Site Selection Site selection is critical in achiev- ing representative and consistent trip generation rates. Failure to select sites appropriately may lead to inaccurate trip generation rates and equations. Use of unrepresen- tative sites asa basis for trip gener~ ation estimates can result in over- ‘or underestimating trips to be gen- erated by a proposed development. Suggested criteria for identifying sites for collection of trip genera- tion data are as follows: ‘# Data should be transferrable; therefore, i is critical that both trip data and development characteris- tics be representative of the land ‘uses to be analyzed. This includes developntent size, mix of develop- ‘ment components and geographic location with respect to the area roadway network and area develop- ‘ment patterns. ‘ The development should have reasonably fll occupancy (ie. at least 85 percent) and appear to be economically healthy (note: percent ‘occupancy at the time of the survey, if applicable, should be recorded). ‘© The development should be mature (ie, a least two years old) and located in a mature area so it represents the ultimate characteris- ties of a “successful” development. © The data needed to deseribe the independent variables should be available. The site should be selected on the ability to obtain accurate trip generation and development char- acteristics # Ieshould be possible to isolate the site for counting purposes: + No shared parking (unless the parking areas for the site are easly distinguishable); + No shared driveways (unless the driveways for the site are «easily distinguishable); * Limited ability for pedestrians to walk into the ste from nearby parcels; imited transit availability or use (unless transit usage can be counted—e.g., elementary stu- dents who ride a school bus), and No through-traffic. The site should have a limited number of driveways (asa data col- lection cost consideration). © The driveways (or the method of counting trafic) should ensure against double-counting vehicles. # Te should be possible for counts to be made safely. ‘The need for any special security measures should be identified. © The site should consist ofa sin- le land use activity (unless a multi- ‘use study is being conducted as described in Chapter 7) ‘There should be minimal to no on-site construction or adjacent roadway construction. Permission should be obtained from the owner or the building manager (note: owners/managers are sometimes more willing to be surveyed ifthe confidentiality of their site is guaranteed or ifthe results are provided to them). ‘Tip Generation Handbook, 2nd Edition Chapter 4 ml ITE 17Independent Variable Selection For a new land use being surveyed, ‘one or more appropriate indepen dent variables need to be identified, ‘measured and analyzed. When identifying a potential independent variable, the following points should be considered. The data for the independent variable should be readily availble, for the survey site and any poten- Gil proposed development ofthis land use type for which trip gener- ation estimates may be desired. When in doubt about which inde- pendent variables may be most © The number of trips generated appropriate refer to Trip Generation at the site should be influenced in a under the same or similar and use logical way by the independent to see which ones have produced variable. Correlation does not the most stable relationships and equal causation. reliable rates or equations, Typical independent variables include mum- Available site data should be her ofemployees, gross floor area, accurate, for sites being counted ——_gross leasable area and number of and proposed future development oceupied rooms or dwelling units. It (ie, if it eannot be projected for is critical that the definitions of new development, itis not an independent variables be the same appropriate independent variable). 35 chose defined in Trip Generation (Chapter 3 in the Users Guide, ‘© Variables for similar sites should Seventh Edition) or Appendix D of be provided direetly and not merely {his handbook, estimated from a different variable. For example, the number of For a trip generation study involv- employees ata site may appear to ing a land use for which trip end bea valid independent variable, but and independent variable informa- it should not be used if the vale ition ig already provided in Trip ‘typically derived by factoring in Generatier, the choice of indepen- another independent variable, such dent yaiables should (atthe min a gross square footage of the ‘mum include those presented in development site. Trip Generation, I other indepen- dent variables appear to be logical and satisfy the criteria cited above, data for them should be collected and analyzed as well. 18 ITE Tip Generation Handbook, 2nd Euition Chspter 4 ‘In general itis recommended thee cata be collected and compiled for 4s many potential, appropriate independent variables as practical. As the Trip Generation database grows, itis quite likely that future analyses ofthe available data will identify additional relationships involving more than the currently used independent variables. ‘The sample data collection forms presented at the end of this chapter contains list of suggested data to obtain, Development Data Requirements “Trip generation estimates are based on development characteris- tics that are used as independent variables. This normally requires a check with the owners or man- agers ofthe development to ensure the availability of accurate data on physical characteristics. For example, itis insufficient to merely count dwelling units or square feet. A count of ecupied square feet or units is needed. The ‘occupied space represents the par- tion of the development that is actually generating trips. ‘Occupied square footage should also be carefully evaluated to make sure that it is actually being oceu- pied and wed rather than merely leased or purchased. For example, in some land use classifications (partic- ‘ularly warehousing, industrial and office), tis common practice for tenants to lease or porchase future expansion space, but not to occupy it for some time. Use of “leased or FRAP RAAF AA ARPA A ARM AMP AMP PAR AAASpurchased” square footage instead of “occupied and used” square footage can be misleading and may be one reason forthe scatter in the histori ‘cal data points within certain classif- cations in Trip Generation ‘At this stage inthe study design proces, its necessary to decide whether t include consideration of transi use, parking accumula- tion and automobile occupancy If these issues are to be considered as factors in the analysis ofthe local trip generation data, then appro- priate data should be colleeted. Survey Periods Site-generated trafic should be counted, if feasible, fora full 7-day period to determine when total site-generated traffie volumes peak: during weekdays and the weekend, At the minimum, automatic traffic recorder counts should be taken shrough a fall 24-hour period, although a preferred length of time would consist of 48 consecutive hours. Some sites require manual count- ing techniques because automatic traffic recorder devices will not capture all trips (or may not be accurate due to the configuration of the site driveways). Manual traffic counts should last for a minimum cof 2 hours for each peak period, depending on whether the adjacent street trafic peak or the generator’ peak is being surveyed. Ifthe desired traffic analysis requires other periods, counts for those periods should also be obtained. ‘The day of the week and time of day are also important consider~ tions in obtaining meaningful results, The purpose of the study will dictate the critical time period for analysis. Tn many cases, the season of the ‘year i also important. In general traffic generation for land uses with title oF a0 seasonal variation should be counted on average days. For land uses with significant sea- sonal variation, time periods repre senting the 30th to SOth highest hhours of the year may be used. Retail centers and recreational uses are typical examples of land uses ‘with significant seasonal variation. Care should be taken to avoid making counts during special events, holidays, construction periods, bad weather, or other times when conditions at the study site or in its vicinity may affect site trip generation. The time period being surveyed should represent typical activity for the site (e.g, no data collec- ‘Tip Generation Handbook, 2nd Edition Chapter 4 m ITE 19 tion should be conducted during a super sale at a retail site) unless the study is specifically designed for collecting data dur- ing a peak time (eg, holiday shopping season for retail sites). PRY Conducting the Study ‘The following guidelines should be reviewed and followed to the extent possible when collecting traffic vol- ‘ume and site characteristics data. Count directional traffic volumes (entering and exiting) by 15-raimute period. Where directional counts cannot bbe made automatically, manual ‘counts should be made during the street peak periods, plus the peak period of the generator to record the peak-hour entering and exiting volumes. Two or more days of peak period traffic counts are desirable. If possible, collect hourly trafic ‘volume data (or obtain from the governing jurisdiction) on all streets adjacent to and with access +0 or from the site so chat adjacent street peak hours can be deter- mined. ‘The trafic counts of multi- ple driveway volumes must be done ‘concurrently. {© Surveys or traffic counts con- ‘ducted on public streets may require a courtesy cal to the proper governing authority. Providing a copy of the trafie volume data or the final study to either the public or private agency involved is, another good policy.‘# Data concerning the site should be obtained through interviews with the site owner or manager and, if necessary, by means of mea- surements. Verify automatic counts with ‘manual counts for short period(s). If pneumatic oad tube counters are used, exercise extra caution and ver~ ification because the accuracy of this equipment may degrade at low- speed traffic conditions. ‘ IF the site could be considered a ‘multi-use development, refer to (Chapter 7 (Multi-Use Develop- ments) for guidance on additional data collection needs: # If pass-by data are being col- lected, specific intercept surveys will be needed (as described in Chapter 5, Pass-By, Primary and Diverted Linked Trips). ‘# If needed, record hourly enter- ing and exiting traffic by vehicle classification and vehicle occupancy and compare with corresponding azutomatic counts to determine a factor for adjusting the raw auto- ‘matic counts. This may require classifying vehicles by number of ales if automatic counters have been used. Refer to the latest edi- tion of the ITE Mamual of ‘Transportation Engineering Studies for guidance, Pyy Establishment of a Local Trip Generation Rate or Equation ‘This section provides guidance if the purpose of the trip genera- tion study isto establish a new local trip generation rate or equation. Ifthe purpose ofthe study isto validate the use of Tip Generation for a local application, Section 46 provides appropriate leis recommended thatthe frst analysis step in the establishment of «local trip generation rate or equa tion be the development of hypothesis for why the Trip Generation data might not be appro- ate for local application. For cexample, the rationale could involve the age of residents, metropolitan area characteristics and/or the avail- ability of transit. Ie is important that the local community have a com- ‘mon-sense rationale for believing that the local rate is or should be significantly different from that presented in Trip Generation. (Note: the absence of any data cov- ering a particular land use is also a 20 ITE i Tip Generation Handbook, 2nd Edition Chapter 4 valid reason for conducting a local trip generation study.) ‘The second analysis step in the establishment of a local trip genera- tion rate or equation should be confirmation that a loca tip gen- eration rate/equation is indeed jus- tified. This confirmation should be predicated on satisfying the follow- ing three criteria 4 Acleast three local sites are counted (five sites are preferable); 4 The weighted average rate for the counted sites is at least 15 per- cent higher or lower than the com- parative Trip Generation cate or if ‘Trip Generation provides only two or fewer data points; and ‘© The local counts provide consis- tent results. IF establishment of a local trip gen- eration rate or equation is justified ‘based on these three criteria, the next step should involve the selec- tion of either the computed trip generation rates or equations (if applicable) as the local trip genera tion estimator. The development of the local rate or equation should likewise satisfy the standards assigned to Trip Generation data for its use. In other words, the local data should be used with the same caution and desire for statistical integrity a the ITE database, [As described! in Chapter 3 (Guidelines for Estimating Trip TOM ANA APH MAHAR AAA AAA AHRAAnAAAaneaaeGeneration), an acceptable use of the weighted average trip genera- tion rate requires at least three data points with a computed standard deviation that is no more than 110 percent of the weighted average rate, The acceptable use of a regression equation requires at least four data points with a com- puted R? of atleast 0.75. ‘The local trip generation documen- tation shold clearly state the local rates and/or equations, the situations in which they are applicable and. ‘what to do in situations where they are not applicable. The docamenta- tion should also presen the ste- specific information, Consideration should be given to submitting the data to ITE for use in subsequent editions of Tip ‘The second step involves analysis of, Generation, Sources will be cited, the local data and comparison af t but the identity of specific sites will to the ITE Trip Generation data, A be kept confident be transmitted to: ial. Data should Trip Generation rate/equation should be considered valid for local use if it meet the following citer: Institute of Transportation Engineers 4 The trip generation rate foreach 1099 14th St., NW, Suite 300W- of the locally surveyed sites falls ‘Washington, D.C. 20005-3438 within one standard deviation of ‘Tak: +1 202-289-0222 the Trip Generation rate; Fax: +1 202-289-7722 Validati Trip Get Rates/Equa' Local Use This section provi “# Of the sites surveyed locally, at least one has a rate higher than the jon of ‘Trip Generation weighted average eration rate or equation and one has 2 HONS fOr ower rate; oral ofthe survey sites generated trips with totals within ides guidance if 15 percent of the Trip Generation the purpose of the trip genera- average rate or equation (calculated tion study is to validate the use of Trip Generation data for a Jocal application. the study isto establish a new local rate, Section 4.5 ps ate guidance. Validation of Trip or equations for use in a pasticular locale should be accomplished using a two-step process. The first step is to collect local trip genera- tion data at no fewer than three local sites (or supplemental data obtained from other sources to ce- Cgqsideration should be given to submitting the data to ITE for use in subsequent editions of Trip ate a database of three or more local data sites). as follows: the difference between the survey site rate and the Trip Ifthe purpose of Generation rate, divided by the Trip Generation rates rovides appropri- The locally collected data gener- ally fall within the scatter of points shown in the current Trip Generation data plot; and © Commion sense derived from the local trip generation study indicates that the Trip Generation data are valid for local application. Generation rates Tethe local data do not meet all of the above criteria, development of local rate or equation should be considered (refer to Section 4.5 for suidance). “Tip Generation Handbook, 2nd Edition Chapter 4 m ITE 21Generation. Sources will be cited, but the identity of specific sites will bbe kept confidential Combining Trip Generation and Local Data Ifthe Trip Generation database for a particular land use is relatively small (e.g, nine or fewer sites), the local community should consider merging the national and local databases to create a consolidat- cd trip generation rate. Itis rec- ‘ommended that this merging of the datasets take place ifthe local and a national average rates are rea~ sonably close (eg. within 15 percent of each other). The ‘merging ofthe two databases under those circumstances should Improve the statistical strength of the overall database forthe particu~ har land use, Ifthe local and national rates are ‘substantially different, refer to Section 4.5 for guidance on the establishment of a local rate. ‘The following procedure demon- strates the proper steps for merging, the loca! and national databases. 22 ITE mi Tip Generation Handbook, 2nd Exition Chapter 4 This procedure can be used for any land use, time period and indepen- dent variable for which weighted average trp rates, average size of the independent variables and number of studies are provided in ‘Trip Generation. ‘Note: This method of combining datasets does not allow precise calealation of the standard devia tion or ofa revised regression equation because of the unavailabil- ity ofthe exact data points in the ITE national database.‘Tip Generation Handbook, 2nd Edtion Chapter 4 ml ITE 23pig 4019 uo ue ae\duI09 aeeDIK Sa Sew meee s a z —— : ry patna : Wavaomns sa wa WS AS 5B — : meee — 5 ‘nowy — 2 & see — ey 2 (ero wth g 5 i a ; o a a 2 3 2 a a o a xp ere a Locos ers oust ws se aa or a a t ORR ee 5 son saaTes ner aca ST AEE me Ss fecha ([ ung) wsoy vIDG UONDAQUayH As] sioouidug uonepodsues, Jo oynnsu =9MBREUER ERUUU UEP UYU eee eee 883 pg i ct - szmmgieo snelepy rennet “tay 24 pions fours _pataene wonmuZopy MUONPRE 0pUR E HES HONDO) i 3 a] pao op 7 ea] ope pay za 7 Fra] mon nel Par Sea] mec ioam| ma | meal roe ny CI ORME TOT TTA sui Jo wane my20} 249 as08n dopuseimy da Greed par RAE Gn vo RAL ONAL BAPHNT Poo HONEA TV = IVD ugg Georars emmS (¢ Lg) 40, DIDG UoYDsIUaDH AA], sJoouiGua uonepodsues) Jo oymysu) 99 26 ITEM Trip Generation Handbook, 2nd Exltion Chapter 4HF Institute of Transportation Engineers Trip Generation Data Form (Part 3) Name/Organization: Telephone Number: City/State: Dated Disecy Poh heck sat Fer and? fpocae prot eb] wrmaon Dey of he we aietrat za ae [ae Te are reer ITER (tA ede Cont, nding race; Traha~ Hay Dy Dh aes) a me ar Tae [eee a ce TEIN TESTER zt TOOT 7aa5 Trip Generation Harcibook, 2nd Ecition Chapter 4m ITE 27