Imagine you are at a soccer game. Players sprint up and down the field, passing the ball, and ultimately scoring. The count on the scoreboard increases by one. The game continues. And the team with the most goals wins.
Now imagine you are at a basketball game. Again, players sprint up and down the court, dribbling the ball, and ultimately scoring. But this time, not all baskets are worth the same amount of points. How do we know that a player should get three points, versus just one? The answer is a line, clearly labeling the point designation. And the team with the most baskets wins.
Finally, imagine you are watching Olympic figure skating. There are no baskets, no goals. But still the person with the most points wins. How is that determined? The answer this time is a clear and exact rule book, detailing each and every technique fault and deduction, the level of difficulty of each and every move, the points awarded for musicality and grace.
Now, I am not a soccer player, nor a basketball player, nor am I – believe it or not – an Olympic figure skater. I am, however, a member of Harvey Mudd’s Social Insect Behavior Lab. Right now, we are studying the interactions between a native (Cephalotes varians) and invasive (Pseudomyrmex gracilis) species of ant. Specifically, we are currently chucking one C. varians and one P. gracilis into a petri dish and watching them battle it out, in the name of Science! (Or more specifically, in the name of Marissa, whose thesis this data is going towards.) And while another lab member Lexie Stinson and I were charged with scoring these ant fights without knowing the hypothesis (we were hypothesis-blind to remove biases), we learned a whole lot about the ants, experimental procedure, and scientific research.
I know Lexie described our ant fights as wrestling in the last blog post “Wrestling Ants: How We Score Insect Aggression Like WWE Moves,” but I like to think of the ants as mini figure skaters. (Though, ultimately we make the same point.) In her post, she outlined how we score the ant fights. The procedure is simple and easy to follow. But what happens when we need replicable data? Or when we need to train someone else to score these fights? Science depends on replicable data. How could we provide this for Marissa’s thesis?
As Lexie and I learned the hard way this past month, you cannot just make it up as you go if you want your results to be replicable or consistent. After all, you wouldn’t get a bunch of randos who have no experience in figure skating to judge the Olympics. Their scores would be based on no criteria, and end up all over the place. (For example, they could do what my sister and I used to do, and judge solely based on the amount of sparkles on their outfits.) But no, figure skating has a set of guidelines and trained judges. Guidelines that make one particular judge give consistent and comparable scores to skater after skater.
But how do they make sure every judge’s scores are consistent and comparable with each other? Again, guidelines. A rule book that is so specific, there is very little room for variation.
We did not have this a month ago… Lexie and I were asked to score a multitude of one-on-one ant interactions for aggression. Below is the extent of the guidelines given to us:
What is ‘avoidance’ or ‘antennation’? What qualifies as ‘short term’ and ‘prolonged’? We quickly discovered that when guidelines are not specific enough, we end up with extremely varied results.
While each of us were consistent between our own scores, compared to each other, they were wildly different. For example, look at this specific experiment:
Avoidance = the number of times one ant avoided or ran away from the other (number of category 1 incidences).
Aggression = a scaled representation of the number of categories 2-4 incidences (category 2 = +1, category 3 = +2, category 4 = +3). So, for example, 2 times antennation + 1 times quick bite + 1 times long bite = 7
In this case, I (HB) found the ant much less avoidant and much less aggressive than Lexie (LS). But it did not consistently go like this. I did not always have lower scores than Lexie. In the next video we watched, we got the following results:
Here, I scored way higher, all around. Yet we both thought we were being consistent in our scoring, which highlights a big problem: how can the experiment results be at all conclusive when we are being so unpredictable? Not only were our results inconsistent with each other, our findings would lead us to different conclusions about which ant is more aggressive. Since we were hypothesis-blind, we did not accidentally skew the results one way or another, but that also means neither of us produced consistent results. Basically, us being bad at consistent scoring could have messed up Marissa’s whole thesis.
This problem drove us to a solution: a clear and specific set of guidelines, just like figure skating!
Drumroll please… INTRODUCING: HOW TO SCORE ANT FIGHTS (A Guide to Ant Scoring by Lexie Stinson and Hanna Bibbler Parker)!!!!!!
In this guide, we have taken the original set of instructions, and elaborated on them. First, we have added comprehensive descriptions of each type of behavior. Second, we collected videos of each of them to show exactly what we are referring to. Third, we created a step by step guide on how we recommend going about scoring each video, suggesting things like what order to watch each species’ video and how to set up a score sheet.
This project served two purposes. Primarily, we have reached consensus on how to score each observed behavior, and put that towards our research and Marissa’s thesis. We were able to re-score videos and come to a more accurate and consistent conclusion – one that could be replicated by others who watch the same videos. This data could now be used to draw a conclusive response to Marissa’s hypothesis (which, we later found out, was that the C. varians would avoid and P. gracilis would attack).
Additionally, this guide can be used in the future. The hope is that, someday, some brave new soul will be tasked with scoring these fights – or more fights like this, for a different experiment – and in their moment of uncertainty, will turn to our trusty guide, which will teach them the ways of the ant scoring, and so they will go about their research, confident and enlightened and ready to tackle whatever comes their way. (Or at the very least, they will know what ‘antennation’ means.)
Further Reading:
https://hmcbee.blogspot.com/2025/05/wrestling-ants-how-we-score-insect.html
No comments:
Post a Comment