An empirical reappraisal of the level of traffic stress framework for segments

Highlights

• An empirical update to Level of Traffic Stress using stated preference survey.

• New LCC framework finds bikeway characteristics suitable for three cyclist types.

• LCC framework yields more conservative connectivity estimates than LTS.

Abstract

The Level of Traffic Stress (LTS) framework is widely used to assess the suitability of roadway environments for cycling. Its main strength is the identification of infrastructure characteristics that meet the needs of different cyclist groups. The four levels of infrastructure ratings, LTS 1 to LTS 4, roughly map to the cyclist types defined by the Four Types of Cyclist typology. Despite its popularity, the LTS framework has several limitations, including reliance on a cyclist typology that was developed subjectively, and a lack of empirical evidence to define thresholds between levels. This work builds on our previous empirically-based findings that cyclists form three groups rather than four: Uncomfortable or Uninterested, Cautious Majority, and Very Comfortable Cyclists. We use survey data from Edmonton, Canada, to update the LTS framework such that levels match the three types of cyclists. Direct infrastructure ratings, binary logistic regression, and route choice data provide the empirical foundation to determine infrastructure characteristics that are suitable for the three types of cyclists. This adjusted framework is called Level of Cycling Comfort (LCC). We apply the framework to Edmonton and compare connectivity outcomes using both LTS and LCC frameworks. Overall, the LCC framework yields more conservative estimates of connectivity.

Introduction

A well-connected bicycle network and direct routes are associated with greater levels of cycling (Schoner and Levinson, 2014), which has been identified as a mode of transportation that can mitigate congestion and reduce air pollution and greenhouse gas emissions in urban environments (Guttenplan et al., 2003, Lindsay et al., 2011). As such, the connectivity of bicycle networks has become an increasing focus of research since the 1990s (Buehler and Dill, 2016). Many metrics, such as connected origin–destination pairs or number of accessible jobs, have been created to measure bicycle network connectivity and its related topics of cycling accessibility and bikeability (e.g. Boisjoly and El-Geneidy, 2016, Lowry et al., 2016, McNeil, 2011, Schoner and Levinson, 2014). These rely on frameworks or equations that define what constitutes acceptable cycling infrastructure which are used to identify segments of the full network that are suitable for cycling. However, other research has shown that cyclists have diverse perceptions of what constitutes a safe and comfortable riding environment (Abadi and Hurwitz, 2018, Larsen and El-Geneidy, 2011, Stinson and Bhat, 2005, Veillette et al., 2019), such that different cyclists will perceive the connectivity of a network differently.

One of only a few tools designed to measure the suitability of infrastructure for cycling while specifically accounting for the diversity of cyclist preferences (as per a cyclist typology) is the Level of Traffic Stress (LTS) framework (Mekuria et al., 2012). It defines infrastructure characteristics that are suitable for cyclist categories known as the Four Types of Cyclists, a typology originally developed by City of Portland bicycle coordinator Roger Geller (Geller, 2006), and later affirmed by Dill and McNeil, 2013, Dill and McNeil, 2016. In the LTS framework, facilities categorized as causing the highest level of cycling stress, LTS 4, are considered suitable only for the Strong and Fearless cyclist type. LTS 3 facilities are matched to the Enthused and Confident, and LTS 2 to the Interested but Concerned. Although the typology contains a fourth cyclist type, No Way No How (i.e., those who will not or cannot cycle), LTS 1 facilities are rather considered to be those suitable for children. Recently, the framework was used on a national scale in the United States to compare job accessibility by bike in 50 major U.S. cities (Owen and Murphy, 2019). The ease of use and limited data requirements have undoubtedly contributed to the widespread application of this framework.

Despite its popularity, the LTS framework has a notable limitation: it was subjectively developed using expert knowledge and existing design criteria (Mekuria et al., 2012) rather than empirical evidence. The Four Types of Cyclists typology on which it is based was also subjectively developed (Geller, 2006). Using survey data collected in Edmonton, Canada, we found that cyclists naturally form three categories rather than four: Uncomfortable or Uninterested, Cautious Majority, and Very Comfortable Cyclists (Cabral and Kim, 2019). We used statistical methods to derive this empirical typology, using variables as similar as possible to the Four Types of Cyclists in order to obtain a typology that is functionally similar, but divides Edmonton’s population into cyclist types that are reflective of its particularities. In this paper, we adjust the LTS framework using empirical data to reflect the comfort of each group in this new typology, thus ensuring connectivity assessments are reflective of the perception of infrastructure for diverse cyclist types. We also compare connectivity, as assessed with LTS, with our updated framework.