I gave the survey to a small group of people and had them watch the animations I made and note their responses, as I outlined in the last post. In order to interpret their answers, I decided on a framework to analyse the clips according to certain variables in the animation and the way the character behaves, rating each on a scale from 1 (animal) to 5 (human).

Stance – Is the character upright, all fours, or somewhere between?
Movement – How the character moves, the timing of its movements and gait, etc.
Facial expression – Do they emote in an animal-like fashion, or human, or a mixture?
Gesture – The presence or absence of human gestures/mannerisms.

As the one who created the animations and according to my thought processes as I did so, I would assign the clips like so:

Ridley upright – Stance 5, Movement 5, Face 4, Gesture 5
Ridley down – Stance 3, Movement 3, Face 2, Gesture 1
Leopard-Man upright – Stance 4, Movement 4, Face 3, Gesture 3
Leopard-Man down – Stance 1, Movement 1, Face 1, Gesture 1

I am also including ‘Appearance’ as a factor, to estimate how much the overall design of the characters might affect the perception of their behaviour. In which case I’d assign Ridley a 4, being close to but not entirely human; and the Leopard-Man a 2, being close to but not entirely animal.

These ratings are simply to help me interpret people’s responses to the clips – e.g. if a clip with a low Appearance score but a high Movement score reads as particularly human-like to viewers, that suggests that the kind of movement in an animation is a more important aspect than visual design in affecting the audience’s perception of the character. Additionally, the degree to which the participants in the survey agree or disagree with my scores suggest how successful or unsuccessful I was in trying to evoke a particular reaction for each clip.


Questions 1. and 2.:

Some of these results were expected, others were more of a surprise.

Out of both Ridley clips, 80% of respondents preferred her in the clip where she was performing human behaviour. Out of both Leopard-Man clips, 60% prefer the animation with the creature down on all fours and acting like an animal. This suggests that in terms of coherently marrying design with behaviour, I was successful in getting across both Ridley’s original humanity and the Leopard-Man’s animal origin, as those corresponding animations were the most successful to the viewers.

Comparing both clips where the characters are acting more human, 80% still preferred Ridley over the Leopard-Man, which agrees with the above results. Out of both animal-like clips, however, 60% also preferred Ridley. This was unexpected, and may be put down to the design of the characters – participants’ comments included that she appeared more ‘friendly’ – but also perhaps to the way her animations were influenced by my deeper knowledge of Ridley’s character: one responder commented that Ridley seemed to have more personality, leading to more positive feelings about her than the Leopard-Man.

The next question crossed the categories over, matching Ridley walking with the Leopard-Man on all fours, and vice versa. Presented with Ridley acting human with the Leopard-Man acting like an animal, 60% responded in favour of the Leopard-Man, which I did not expect. This may be because the Leopard-Man’s animal like action matched its animal like appearance and made more visual ‘sense’ compared to Ridley’s not-quite-human looks paired with very human movement. The last match-up was between the Leopard-Man walking upright and Ridley acting more like an animal. This went 100% in favour of Ridley’s clip. This perhaps gives further explanation to the previous result, if we bear in mind the Uncanny Valley theory: both examples of a part-human-looking creature behaving more or less like a full human were rated lower in appeal than the same creatures performing animal like behaviour.


Question 3, Human vs animal:

From my scoring system I would expect the clips to be judged, in order from most human to least, like this:

Ridley upright, then Leopard-Man upright, then Ridley down, then Leopard-Man down.

The results from the survey are interesting – while I was correct in my estimation of the most and least human-seeming clips, Ridley on all fours and the Leopard-Man walking upright drew for middle place. This suggests that the stance was not as strong a factor in judging how human or not the characters seemed than I’d anticipated. This could be explained by the appearance of each character – it could be that Ridley’s mostly human design was a stronger element than the fact she was posing and moving mostly like an animal might; and that the Leopard-Man looks inhuman enough that even an upright carriage doesn’t counteract it. This response also surprises me because Ridley’s expression in that clip are pretty weighted towards non-human, however I can think of a possible explanation. As a more expressive character, her clear and strong reactions to things may have read as more recognisably human-like. The Leopard-Man’s reactions are less exaggerated, which coupled with its design may have contributed to participants not seeing it as very human-like.

Question 4. Genres:

I asked participants to choose from a list of genres which they thought each clip might fit with. I gave options for things like horror, fantasy and sci-fi, which I expected the participants to choose, and some other genres like comedy, drama, thriller, etc. which I thought were more likely to go unchosen. The breakdown was as follows:

Ridley ‘s animal clip: Top choices were fantasy, sci fi and action. Horror was not chosen by any participants for this clip. Comedy was at 16%, which again I think is due to Ridley’s friendlier look and her more exaggerated expressiveness. The story Ridley comes from is a fantasy one with action and slight comedy elements, so that being one of the top choices in the respondents indicates the clip was partly successful in evoking the reaction I would want for that story setting.

Leopard-Man walking upright:  Highest choice was fantasy, then sci-fi, then horror and thriller tied for third place. I would have assumed horror might be in first place, especially with the responses previously discussed where this clip was generally the least appealing. Comedy was chosen at 6%, which I think is likely due to an animation choice of having it be hit with the ball, which maybe came across as slightly slapstick. Action was chosen 11% of the time. This clip matches the Moreau narrative in the sci-fi/horror/thriller elements, however the other genres chosen suggest the animation is unclear enough to evoke other responses that make it weaker than it ideally should be.

Ridley’s upright clip: Fantasy came in first, followed by a three-way tie for comedy, drama, and action. Sci fi was chosen at 12%. Horror was picked at 6%, which I did not predict. This does agree with the earlier possibility that while she does appear friendlier than the Leopard-Man, she does still have something uncanny, or potentially so, about her. I’m not displeased by this; in the narrative setting she belongs to, while she needs to be likeable and engaging, having her also be able to be a little disconcerting works for the way the world in that story views her. This clip was therefore reasonably successful in creating the perception of the character I had intended.

Leopard-Man on all fours: Sci-fi came top, which isn’t as much of a surprise considering the H. G. Wells narrative it comes from is a science fiction story. The next highest are fantasy and horror, which is similarly expected, followed by thriller and action. This clip conforms very well to the setting and tone of the story the character originates from, which makes this the most successful clip in terms of how well it fills its narrative purpose.


I will note my overall conclusions from this survey in the next post.