The paper introducing ‘Replication Studies’ with more details on how to use replications as a training tool has now been published in our recent paper, which you can find here.
If I could poll all behavioral ecologists, or even scientists more generally, my guess is that many if not most of us would say that we all agree that replicating our experiments is a key component to building confidence in our research results. We all know that the world is highly variable and that while demonstrating that some process might be occurring in a given species at a given time, until we’ve seen this same process occurring repeatedly, it’s impossible to know how general this pattern may be. When multiple experiments can document the same or similar things happening across different populations or different contexts, we can begin to be more certain that this process may be playing a broader or more important role in patterning behavioral decisions of animals.
And so while most of us would agree we would like to see more replication of results, we also all lament that this is a pretty tough goal: it’s hard to justify to funders why we need to replicate something that is already “known” and it’s even harder to get such datasets published even if we can collect them since we all know journals often prioritize “novelty” of research.
Well I’m here to suggest one potential way we can alleviate this concerns!
First, I suggest that we incorporate replications into early research training. Have the first project a student does in your lab be a close replication of previous work. This builds research experience and confidence in students allowing them to gain valuable research skills in a lower-stakes scenario before embarking on higher-risk, more independent projects. This is actually something I started doing in my lab purely as a training technique – I found that a major barrier for many first year graduate students was coming up with a ‘novel’ question. Starting researchers are still just getting the lay of the land and so identifying truly novel research gaps can feel very daunting. So instead, I decided I wanted my students to focus on the process of science and research first and not worry too much yet about whether or not their study was ‘cool’ enough.
In this way, I had each of my students find a paper on small freshwater fish (which is what our lab studies) that was dealing with a topic that they think they want to study for their dissertation and then replicate it in the lab! We called these ‘first year experiments’ (for lack of a more exciting moniker :D) and they’ve been an immensely helpful experience! Replicating a previous paper gave my students very helpful guardrails in designing their first projects – they learn how to construct behavioral arenas, how to handle live animals, how to set up cameras, deal with video recordings, analyze data etc etc. Most importantly, without fail, this experience helped then launch all my students into their more independent work. The more experienced researchers know that the best way to come up with new questions is by trying to understand what is going on in your current dataset!
So my lab collected all these great little replication datasets and then I was left to wonder…. what do we do with them? These datasets are valuable because they let us which results have some generality (i.e. the result could be replicated) and which results maybe are a bit more context-specific. At this point, I decided to approach one the journal of my home professional societies, the International Society for Behavioral Ecology and see if they’d be willing to create a home for these important datasets. And they were!
Second, in collaboration with Behavioral Ecology, we have created a new article type – “Replication Studies.” The goal of “Replication Studies” is to be a home for ‘close’ replications (i.e. same or similar animals, same or similar methodologies). Importantly, these articles will also be short-format (3,500 words) as they don’t require extensive motivation in the introduction or detailed methods as the original study already provided that. I think this also presents another great training opportunity for starting researchers as they can get their first paper out with a bit of a lower barrier to writing.
This idea of using replications as research training and the genesis of the Replication Studies section at Behavioral Ecology really just seems like a win-win situation all around! The paper introducing this section and with more details on how to use replications as a training tool has now been published in our recent paper, which you can find here. Details of what counts as a ‘close’ replication are also available in the author guidelines which you can find on Behavioral Ecology’s website here. If you want to hear more about this section and the philosophies behind replications, you can read my interview with Jeremy Fox at Dynamic Ecology too.