Table 2.
Linear mixed-effects model of log-transformed activity as a function of dog-related variables, whilst controlling for owner age and location (urban or rural), using the final dataset.
| Variables | Estimate | Error | 2.5% CL | 97.5% CL | p-value | Marginal mean (LM) |
2.5% CL | 97.5% CL | Marginal mean (BART) | 2.5% CL | 97.5% CL |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Intercept | 34.866 | 0.017 | 33.689 | 36.083 | < 0.001 | – | – | – | – | – | – |
| Dog age−2 | 2.432 | 0.020 | 2.303 | 2.565 | < 0.001 | – | – | – | – | – | – |
| Owner age | 0.033 | 0.002 | 0.028 | 0.037 | < 0.001 | – | – | – | – | – | – |
| Breed size | |||||||||||
| Toy | −0.242 | 0.014 | −0.263 | −0.221 | < 0.001 | 37.0 | 36.4 | 37.6 | 39.4 | 38.2 | 40.3 |
| Small | −0.202 | 0.016 | −0.227 | −0.176 | < 0.001 | 37.7 | 37.0 | 38.5 | 41.9 | 40.6 | 43.2 |
| Medium-small | −0.051 | 0.015 | −0.077 | −0.023 | 0.001 | 40.4 | 39.6 | 41.2 | 42.6 | 41.4 | 43.7 |
| Medium-large | Reference category | 40.3 | 39.7 | 40.9 | 41.7 | 41.0 | 42.5 | ||||
| Large | 0.054 | 0.014 | 0.026 | 0.083 | < 0.001 | 41.2 | 40.4 | 41.9 | 43.6 | 42.5 | 44.8 |
| Giant | −0.031 | 0.024 | −0.076 | 0.015 | 0.872 | 35.5 | 34.4 | 36.6 | 38.0 | 36.4 | 39.7 |
| Weight status | |||||||||||
| Ideal weight | Reference category | 38.7 | 38.1 | 39.2 | 42.0 | 41.6 | 42.5 | ||||
| Overweight | −0.002 | 0.002 | −0.005 | 0.002 | 0.896 | 38.6 | 38.1 | 39.1 | 40.7 | 40.1 | 41.3 |
| Neuter status | |||||||||||
| Intact | Reference category | 39.0 | 38.0 | 39.9 | 42.6 | 40.4 | 44.4 | ||||
| Neutered | −0.017 | 0.013 | −0.042 | 0.008 | 1.000 | 38.3 | 38.0 | 38.6 | 41.6 | 41.1 | 42.0 |
| Sex | |||||||||||
| Male | Reference category | 39.5 | 39.0 | 40.1 | 42.5 | 41.9 | 43.2 | ||||
| Female | −0.045 | 0.006 | −0.056 | −0.034 | < 0.001 | 37.8 | 37.2 | 38.3 | 40.6 | 39.9 | 41.1 |
| Location | |||||||||||
| Rural | Reference category | – | – | – | – | – | – | ||||
| Urban | −0.004 | 0.006 | −0.015 | 0.008 | 1.000 | – | – | – | – | – | – |
| Breed size-age interaction | |||||||||||
| Toy | 0.010 | 0.002 | 0.007 | 0.014 | < 0.001 | – | – | – | – | – | – |
| Small | 0.005 | 0.002 | 0.001 | 0.008 | 0.241 | – | – | – | – | – | – |
| Medium-small | −0.013 | 0.002 | −0.017 | −0.010 | < 0.001 | – | – | – | – | – | – |
| Medium-large | −0.023 | 0.001 | −0.025 | −0.020 | < 0.001 | – | – | – | – | – | – |
| Large | −0.028 | 0.002 | −0.031 | −0.025 | < 0.001 | – | – | – | – | – | – |
| Giant | −0.039 | 0.002 | −0.046 | −0.032 | < 0.001 | – | – | – | – | – | – |
The details include intercept, fixed effects coefficients and their 2.5 and 97.5% confidence limit (CL) and p-values. Estimated marginal mean activity is displayed for values of categorical variables for the linear model (LM) and the BART model, where all other variables are adjusted for mean values (dog age of 6.0 years and owner age of 41.7 years). The model had an R2 of 0.515 (SD 0.0108) and a RMSE of 0.462 (SD 0.0154) on log-transformed active minutes, whilst pet ID as a random effect accounted for 0.221 of the variance in the data.