Impact model parameter estimates (percentage changes).

<p>Notes: *** <i>p</i> < .01, ** <i>p</i> < .05, * <i>p</i> < .1. Standard errors in parentheses. The July 2018 (Phase 2) estimates are relative to the Phase 1 trend; for example, a nonzero July 2018 slope change indicates a change in the slope of the trend from phase 1. Adjusted for seasonality (month dummies), household composition (number of household members by age and sex group), SES (4 categories), head of household education level (less than high school, high school, more than high school), region-quarter unemployment rate, number of public holidays in the month, October 2014 beverage tax changes (pre-post dummy), and unobserved time-invariant household characteristics. <i>N</i> = 138,367 household-month obs. (2,842 households).</p> <p>(DOCX)</p>