Changes in consumer purchasing patterns at New York City chain restaurants following adoption of the sodium warning icon rule, 2015-2017
Prasad D, Jasek JP, Anekwe AV, Dominianni C, Mezzacca TA, Sisti JS, Farley SM, Kessler K. (2023) Changes in consumer purchasing patterns at New York City chain restaurants following adoption of the sodium warning icon rule, 2015–2017. PLoS ONE 18(4): e0274044. https://doi.org/10.1371/journal.pone.0274044
This repository contains the dataset, code, and codebook for the paper, "Changes in consumer purchasing patterns at New York City chain restaurants following adoption of the sodium warning icon rule, 2015-2017.” These files are being made publicly available to fulfill the PLOS journals’ data availability requirement (https://journals.plos.org/plosone/s/data-availability).
- nyc_chain_receipt_2015_17. sas7bdat - SAS dataset with records and variables used for analyses shown in paper. Contains 4,844 observations and 24 variables.
- nyc_chain_receipt_2015_17_code.sas – SAS program with code used to produce results included in manuscript.
- nyc_chain_receipt_2015_17_codebook.csv – Codebook for dataset.
The dataset and program files were created in SAS Enterprise Guide 7.1. The dataset file can be imported into statistical software programs other than SAS (e.g., R, Python, SPSS), but the coding language in the program file is specific to SAS software packages (e.g., SAS Enterprise Guide, SAS 9.4).
In 2016, New York City (NYC) began enforcing a sodium warning regulation at chain restaurants, requiring placement of an icon next to any menu item containing ≥2,300 mg sodium. As shifts in consumer purchases are a potential outcome of menu labeling, we investigated whether high-sodium purchases from NYC chains changed following policy implementation. Using receipts for verification, consumer purchases were assessed at 2 full-service (FSR) and 2 quick-service (QSR) chain restaurants in NYC and Yonkers, NY, which did not implement sodium menu labeling, in 2015 and 2017. Primary outcomes included the proportion of respondents purchasing high-sodium item(s) (containing ≥2,300 mg sodium) and mean sodium content of purchases; changes were assessed by difference-in-difference regression models, adjusted for demographic and location co-variates. At both FSR and QSR, there was not a significant change in the proportion of NYC respondents purchasing 1 or more high-sodium items, relative to Yonkers (FSR difference-in-difference: -4.6%, p=0.364; QSR difference-in-difference: -8.9%, p=0.196). Among NYC FSR respondents, mean sodium content of purchases significantly declined compared to Yonkers (difference-in-difference: -524 mg, p=0.012); no changes in mean sodium were observed among QSR participants (difference-in-difference: 258 mg, p=0.185). Although there was a reduction in mean sodium content of purchases among NYC FSR patrons following sodium warning icon implementation, the mechanism behind the relatively larger NYC decline is unknown.
John Jasek, jjasek@health.nyc.gov