Skip to main content
. 2019 Dec 5;14(12):e0220874. doi: 10.1371/journal.pone.0220874

Fig 1.

Fig 1

Results for AIC analyses for the most parsimonious models (ΔAIC < 2) to predict the number of dogs per household, and (A) associated relative variable importance and (B) model averaged coefficients with 95% confidence intervals (red bars). While models presented that household income had the greatest relative importance when predicting the number of dogs per household, variable coefficients overlapping zero suggest that these were not useful predictors.