2 weeks of regular use of website on child dog safety developed for this research

2 weeks of regular use of publicly-available website on child transportation safety

use of dog safety website at home for about 2 weeks

Coded behavior using objective criteria during a semi-structured interaction with a live therapy dog. Specifically, we examined behavioral patterns for 15 tasks/activities/decisions the child made with the live dog. Sample tasks were when and how the child touched the dog, the extent to which the child was close or intimate to the dog, whether the child handled the dog's toys, and whether the child interrupted the dog during its "rest time". 7 of those hung together in factor analysis. Those 7 were standardized and then averaged to create the scale. It was transformed with linear transformation so all values are positive. Higher numbers indicate higher risk-taking. Theoretically the scale is 0-infinity; in practice most children scored between 0-4. The individual items had an average intercorrelation of .50 and Cronbach's alpha of .65.

Coded behavior in dollhouse simulation. Specifically, in 7 simulated scenarios using a dollhouse that included child and dog characters, furniture, yard, etc., children heard a scene and explained/used the dolls to act what would happen next. For example, the experimenter acted a child doll playing in the kitchen near dog food and the doll dog entered, saw the child, and approached the food bowl. The experimenter said, "[Child's Name] is playing around in the kitchen near [Dog name's] food. [Dog's name] comes into the kitchen and sees [Child's Name] near his/her food bowl making him/her upset and start to growl. What will happen next?" The task was coded using objective coding criteria to score the child's response as safe (1 point), safe but not optimal (0.5 points), or unsafe (0 points). Scores across the 7 scenarios were summed to yield a single score; possible range = 0=7. Higher scores indicate better safety. Inter-rater reliability on 30% of the sample was good; kappa = .90.

2 weeks of regular use of website on child dog safety developed for this research

dog safety: use of dog safety website at home for about 2 weeks

Coded behavior using objective criteria during a semi-structured interaction with a live therapy dog. Specifically, we examined behavioral patterns for 15 tasks/activities/decisions the child made with the live dog. Sample tasks were when and how the child touched the dog, the extent to which the child was close or intimate to the dog, whether the child handled the dog's toys, and whether the child interrupted the dog during its "rest time". 7 of those hung together in factor analysis. Those 7 were standardized and then averaged to create the scale. It was transformed with linear transformation so all values are positive. Higher numbers indicate higher risk-taking. Theoretically the scale is 0-infinity; in practice most children scored between 0-4. The individual items had an average intercorrelation of .50 and Cronbach's alpha of .65.

In the dog safety group, only children with known noncompliance were analyzed for this measure

2 weeks of regular use of website on child dog safety developed for this research

dog safety: use of dog safety website at home for about 2 weeks

There is a discrepancy in participants at risk compared to the participant flow module because some children were not compliant to the intervention. They did not use the internet website and therefore were excluded from analyses. They did not have adverse events; they merely were not exposed to the intervention and therefore were excluded from analysis.

no adverse events occurred; we monitored increased risk for injury, negative medical consequences including dog bites, breach of confidentiality, and other adverse situations to participants and their families