Abstract

Pets play a significant role in improving human health and well-being across the life-span. However,
given the steep increase in dog bites in the UK in recent years (http://www.hesonline.nhs.uk),
the risks of dog ownership also need to be addressed, potential risk factors identified and effective
prevention paths offered. It has been shown that children often confuse a fearful or angry dog
with a friendly one when asked to describe a dog’s emotional expression using still images (Meints,
Racca & Hickey 2010) and that children and adults often do not understand dogs’ signalling (Reisner
& Shofer 2008).
We tested 22 4-year-olds and 24 5-year-olds showing them videos of dogs according to Shepherd’s
(2002) appeasement signalling. We investigated children’s judgements by using a 5-point
face scale ranging from happy (1) via neutral to unhappy (5). We measured if similar misinterpretations
of dogs showing teeth as “happy” occur even when the stimuli are videos and not still
images. As recent studies by Flom et al. (2009) and Pongrácz et al. (2011) suggest that additional
acoustic input may enhance children’s correct interpretation of dog signalling, we also tested if
audible distress signaling (i.e. growling, snarling) influences children’s judgements.
Results show a main effect for Risk Situation (F(3.102) = 29.45; p < 0001) with children distinguishing
high risk dog behaviours from low and medium risk displays, however, they did not
discriminate between medium and low risk. Significantly lower scores were found for videos of
friendly dogs – this shows that children interpreted these stimuli correctly. While scores for the
high risk group are higher than for the other groups, overall, these scores are still lower (average
of 3.2) than expected, possibly due to 81 % of children scoring some of the dog videos as “very
happy”. Children often chose “very happy” when dogs exposed their teeth.
We then compared children on a subset of high risk displays (i.e. growling dogs with exposed
teeth) by showing video stimuli with or without sound to investigate if audible growling increases
children’s correct answers. There were no age group differences, but a significant increase in scores
when stimuli were shown with sound, raising mean scores to 3.8 (out of 5) (F(1.34) = 19.48;
p < 0001).
In sum, this research confirms that children commit errors interpreting dog’s facial expressions:
even when videos instead of still images are used, children still misinterpret highly distressed
displays as “happy” and relaxed. While sounds increase children’s awareness, overall this awareness
does not fully reflect the displayed risk situation. This highlights the urgent need to teach
children to recognise and interpret dogs’ distress signalling appropriately to avoid future injury.
In further research or interventions teaching children dog body language, it would be advisable
to use training stimuli with sound as this seems to enhance correct recognition of dog signalling
for the higher risk signals.