Open Access Open Badges Research article

Validity of instruments to assess students' travel and pedestrian safety

Jason A Mendoza123*, Kathy Watson1, Tom Baranowski13, Theresa A Nicklas13, Doris K Uscanga1 and Marcus J Hanfling24

Author Affiliations

1 USDA/ARS Children's Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine, Houston, TX, USA

2 Academic General Pediatrics, Department of Pediatrics, Baylor College of Medicine, Houston, TX, USA

3 Dan L Duncan Cancer Center, Baylor College of Medicine, Houston, TX, USA

4 Pediatric Injury Clinic, Ben Taub General Hospital, Harris County Hospital District, Houston TX, USA

For all author emails, please log on.

BMC Public Health 2010, 10:257  doi:10.1186/1471-2458-10-257

Published: 18 May 2010



Safe Routes to School (SRTS) programs are designed to make walking and bicycling to school safe and accessible for children. Despite their growing popularity, few validated measures exist for assessing important outcomes such as type of student transport or pedestrian safety behaviors. This research validated the SRTS school travel survey and a pedestrian safety behavior checklist.


Fourth grade students completed a brief written survey on how they got to school that day with set responses. Test-retest reliability was obtained 3-4 hours apart. Convergent validity of the SRTS travel survey was assessed by comparison to parents' report. For the measure of pedestrian safety behavior, 10 research assistants observed 29 students at a school intersection for completion of 8 selected pedestrian safety behaviors. Reliability was determined in two ways: correlations between the research assistants' ratings to that of the Principal Investigator (PI) and intraclass correlations (ICC) across research assistant ratings.


The SRTS travel survey had high test-retest reliability (κ = 0.97, n = 96, p < 0.001) and convergent validity (κ = 0.87, n = 81, p < 0.001). The pedestrian safety behavior checklist had moderate reliability across research assistants' ratings (ICC = 0.48) and moderate correlation with the PI (r = 0.55, p =< 0.01). When two raters simultaneously used the instrument, the ICC increased to 0.65. Overall percent agreement (91%), sensitivity (85%) and specificity (83%) were acceptable.


These validated instruments can be used to assess SRTS programs. The pedestrian safety behavior checklist may benefit from further formative work.