Skip to main content
. 2017 Nov 16;14(11):1394. doi: 10.3390/ijerph14111394

Table 5.

Instrumental variable analysis results using toilet paper as the instrument and two-stage residual inclusion models for the healthy-to-unhealthy price ratio, healthy food price, and unhealthy food price (n = 2765) a.

Price Outcome n n Events Stage 1 Stage 2
t-Value F-Statistic Odds Ratio Lower CL b Upper CL b
Healthy-to-unhealthy price ratio 2765 543 -33.36 1113 0.82 0.57 1.19
Healthy food price 2765 543 59.86 3583 1.15 0.91 1.45
Unhealthy food price 2765 543 98.59 9720 1.10 0.93 1.30

a Covariates in each stage included age, gender, geographic region, wealth/income index, race, smoking status, neighborhood SES, supermarket density, population density, and cost of living index; b Confidence limits were obtained from a bootstrapped analysis using 10,000 replications.