Understanding consumer spending during the pandemic using RI-LTA
We present a novel latent transition analysis with random intercept (RI-LTA) framework to understand the changes to consumer spending landscape through the pandemic across various product categories and acquisition channels. Using data from 785 individuals from across United States at four different time points (pre-COVID, December 2020, March 2021 and March 2022), the presented model revealed five different behavioral classes of consumer spending and the dynamics of movement between classes as a result of the pandemic. The results from the model also revealed that the pandemic’s single biggest impact was in terms of suppression of demand for dine-in and take-out of food - potentially due to it being a high-risk activity. However, over time a reversal of this behavior is seen, back to an increased dine-in and food take-out activity. The second significant behavioral transition was of movement towards delivery and pick-up services where about 2-3-fold increase in their usage is visible — a large portion of which appears to be stable over time, more so in the case of delivery than pick-up.
Trajectories of telework through the pandemic
The altered work landscape in the U.S as a result of the COVID-19 pandemic is still evolving but there is evidence suggesting that some degree of telework will likely stick long beyond the pandemic. If true, these trends will likely have a strong impact on major metropolitan areas since since most cities are structured around the idea of presence of many individuals at the same place and time. In this study, we undertake a trajectory clustering analysis to reveal and characterize clusters of telework trajectories through and beyond the pandemic. Specifically, using agglomerative hierarchical clustering, we identify four clusters of telework trajectories with distinct adoption of telework patterns, and then use cluster membership modeling to understand occupation and socio-demographic factors associated with these trajectory clusters. We also present a set of binary and ordered probit models to project telework patterns in April 2024, four years since the beginning of the pandemic. The binary model is used to understand the factors associated with work location uncertainty in April 2024. For those with no uncertainty in work location April 2024, ordered probit model is used to understand factors associated with the extent to which an individual is expected to work in-person. We use these results to identify implications for the future of transit and local economy.
For whom did telework not work during the pandemic?
The COVID-19 pandemic required employees and businesses across the world to rapidly transition to work from home over extended periods, reaching what is likely the upper bound of telework in many sectors. Past studies have identified both advantages and disadvantages of teleworking. The pandemic experience offers a unique opportunity to examine employees’ experiences and perceptions of telework given the broad participation duration and extent. While employer strategies will play a major role in defining the future forms and adoption of telework, employee preferences and constraints, such as access to appropriate technology to work from home or the home environment, are also going to be important factors. Using data from a U.S. representative sample of 318 working adults, this study uses a Multiple Indicator Multiple Cause Model (MIMIC) to understand employee satisfaction with telework. The presented model links telework satisfaction with experienced and perceived benefits and barriers related to telework, and hence provide a causal structure to our understanding of telework satisfaction. We also present an ordered probit model without latent variables that help us understand the systematic heterogeneity in telework satisfaction across various socio-demographic groups. The results suggest younger and older aged individuals experienced/perceived lower benefits and higher barriers to teleworking compared to middle aged individuals. The results also suggest a disproportionate impact on Hispanic or Latino and Black respondents as well as on those with children attending online school from home. Accordingly, this study highlights important factors impacting telework adoption that employers and policy makers should consider in planning future work arrangements and policies in a post-pandemic world.
Read more here.
Disentagling social capital - Understanding the impact of bonding and bridging on urban activity participation
Social capital is a critical glue for economic and social development in urban areas. Yet, to effectively guide research and practice, there is a need for careful measurement of social capital and how it links to important aspects of urban system functions. This study is aimed at examining the multi-dimensional nature of social capital and the relationship between these dimensions and travel behavior. Prior research has shown connections between stand-alone social capital concepts, such as resources gathered via social networks, with specific aspects of travel behavior. In this work, we expand the definition of social capital to cover separate dimensions, modeled via multiple indicators. Specifically, we make use of over 1400 observations from the Pew Internet Networks and Community Survey dataset to build a Structural Equation Model dividing social capital into two latent dimensions: bonding and bridging to examine the relationship of both these dimensions with discretionary urban activity participation diversity and frequency. Moreover, broader measures of neighborhood and community engagement are included in the model to explain how such engagement can help with the accumulation of social capital.
Read more here.