Click for next page ( 43


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 42
42 The work of survey staff is often reviewed during the DATA CLEANING AND DATA PROCESSING fieldwork as well as at the end of the shift. Supervisors may randomly observe and monitor surveyors on board buses Once surveys have been returned from the field, the task of and trains or at transit centers. Some agencies also use checking, cleaning, and tabulating the data can be substan- "mystery shoppers" (anonymous observers) to monitor sur- tial. This is especially the case for large O&D surveys where veyors. One agency asks drivers to report any problems addresses, intersections, and landmarks must be geocoded. with survey administration. Particularly in the case of O&D surveys, transit staff Transit agencies (and their consultants) employ a vari- emphasizes the importance of beginning data editing and ety of approaches to dispatching survey workers to their coding as soon as surveys are completed. That way, correc- starting locations. The approach depends on the configura- tive action can be taken to mitigate aspects of the fieldwork tion of the transit service, time of day, and other factors. that may be affecting data quality. In addition, the overall In large bus surveys using temporary workers, for exam- workflow goes more smoothly if data collection commences ple, survey workers are often required to provide their own during the fieldwork. transportation to the starting location. Staff may be board- ing the first bus run of the morning at bus garages in Transit agencies report various requirements and proce- relatively remote parts of town and public transportation dures for completeness, data editing, and cleaning. Agencies may not be available. also use a variety of software for data processing and analysis. In other situations, most or all bus routes serve a central Completeness terminal such as a downtown transfer center. Dispatching survey workers from a terminal facilitates close supervision More than one-half of agencies surveyed required that either from a central location. a certain percentage of questions be answered or that certain key questions be answered for the survey to be considered An important consideration for on-board bus surveys complete and usable. The percentages used ranged from involves bus operators. Transit agencies noted that bus oper- 25% to 90%; most agencies that used this approach cited a ators must be aware that a survey will be conducted and must percentage of more than 50%. feel comfortable with having survey workers on the bus. The workers must not interfere with passenger boarding. Los Agencies that required certain key questions to be Angeles County Metropolitan Transportation Authority answered tended to focus on questions related to O&D, trip (Metro) has surveyors board buses at the terminal rather than purpose, and fare type. Greater Portland (Maine) Transit on the street. This ensures that the surveyor boards the District required that the O&D questions be completed. correct bus and that the bus operator trusts the surveyor. WMATA required that O&D and jurisdiction of residence The surveyor rides out to the bus in a support vehicle with be completed. TriMet's Annual Fare Survey required that the operator if it is a second or third shift. fare and transfer questions be answered. Sun Tran required that O&D and other trip-specific questions be answered, but not satisfaction and demographic questions. SAFETY Four in 10 agencies surveyed required that all questions One-third of transit agencies surveyed reported taking steps on the survey be answered for the survey to be considered to protect survey workers' physical safety, whether from complete and usable. Many of these were short surveys crime or accidents. Agencies issue high-visibility or fluores- or surveys conducted by means of personal interviews. cent safety vests to protect against accidents, and also to However, self-administered surveys for the Lane Transit convey to the public that these workers are part of the transit District, Fort Worth Transportation Authority, Milwaukee system. During training, survey workers may be instructed County Transit System, Metra (Chicago area), Southwest on when to stand on buses or trains and when to be seated to Ohio Regional Transit Authority, Potomac & Rappahan- avoid injury. Surveyors working on trains are sometimes nock Transportation Commission, Transit Authority of given rail safety instruction. Northern Kentucky, Tompkins Consolidated Area Transit, Southeastern Pennsylvania Transportation Authority To address concerns about crime, transit agencies have sur- (SEPTA), Sacramento Regional Transit District, and veyors work in teams of two, particularly at night. Agencies Merrimack Valley Regional Transit Authority required that sometimes limit the hours the survey is conducted (to the all questions be answered. 8 a.m. to 5 p.m. period, in one case) and avoid assigning female surveyors to night shifts. Agencies also issued workers Six in 10 O&D surveys reported by transit agencies identification badges, alerted police officials to survey sched- required that all O&D survey questions be answered for the ule dates, and notified bus operators that survey workers survey to be considered complete and usable. Agencies not would be on their buses. requiring that all O&D information be complete generally

OCR for page 42
43 required that at least one part of the start of the trip (origin or Consultants (or academic institutions) were responsible boarding) and one part of the end of the trip (alighting or des- for data cleaning and data processing for 64% of the surveys tination) be complete. reported by transit agencies. The transit agency was respon- sible for these tasks in 23% of the surveys. The consultant and agency shared data cleaning and data processing duties Data Editing and Cleaning in the remaining 13% of surveys. Two in three transit agencies surveyed implemented editing procedures or other steps in data cleaning and data processing Tabulating to ensure the accuracy of the data. Most commonly, agencies checked for logic and consistency in the answers--e.g., Several database software packages are used for tabulating whether transfers made sense based on the routes involved survey results. Surveys reported by transit agencies used both and eliminating self-reported "home to home" trips. Other traditional database software such as SPSS, Microsoft Access, procedures were to eliminate outliers in the data based on the and SAS, and spreadsheet software such as Microsoft Excel number of standard deviations from the mean and proofing (Table 18). the accuracy of data input. Some agencies have had positive experiences with For O&D surveys many agencies conducted extensive using scanners to input data. Scanning reduces data entry address editing procedures to geocode as many locations as needs and can produce data quickly. Other agencies how- possible. Data processing procedures included checking the ever have found that the intelligence applied during data consistency of trip direction, distance, and speed with O&D entry (especially O&D data) is essential to the process and locations and transit routing information; verifying addresses cannot be replicated using scanners. Some agencies have against geographic information system databases, and using also found that scanners introduce errors during the data landmark look-up tables. See Table 17 for survey results on capture process and as a result, take more time than man- data processing for address information. ual data entry. TABLE 17 STEPS TAKEN IN VALIDATING ADDRESS DATA TABLE 18 SOFTWARE USED IN TABULATING RESPONSES Action taken Percentage O&D direction logical with bus/train direction 62 Software Percentage Looked at route/line used for trip 62 SPSS 50 Used logic or consistency checks 57 Verified addresses/intersections using GIS software 48 Microsoft Excel 50 Used landmark listing 43 Microsoft Access 12 Speed of trip logical for O&D distance and time of trip 10 provided by respondent SAS 9 Speed and distance reasonable for originboarding 5 Wincross 3 and alightingdestination pairs based on mode of access/egress (e.g., don't walk 20 miles or at 60 mph) Other software, primarily 14 Other 38 GIS software Total number of respondents, 21. Total number of respondents, 19. O&D = origin and destination; GIS = geographic information system. GIS = geographic information system.