This month saw the first issue of a brand-new journal: Language Development Research, co-founded by LuCiD’s own Ben Ambridge. In this blog, Ben tells us why he started the journal.
It might seem like children’s language development is quite a niche area. But, in fact, as well as more general journals that publish research on this topic, there are – by my count – at least four that specialize in child language. So why do we need another one?
Well, the problem with all of the existing child language journals is that they publish only the “best” papers they receive. Wait – isn’t that a good thing? It depends what “best” means. In practice, papers are often rejected for not being “exciting” enough (e.g., simply replicating or extending an old study), because they didn’t find what they expected to find (or they sort of did, but the results were messy), or just because the journal thinks not enough readers will be interested. In case you don’t believe me, here are a few quotes from journal rejection letters I’ve received over the years (this is no criticism of the individual Editors; this type of “selectivity” is the policy of the respective journals):
The paper does not cover enough new ground to warrant publication as a fully-fledged research article.
It is not clear what new inferences are supported by these data in light of the prior evidence for this effect.
Reviewer 2 questions the utility of these null results on their own.
[from a multiple study paper] Study 1 generated the most interest from the reviewers…and Reviewer 1 went so far as to suggest that it should stand alone as a brief report [i.e., dropping the other studies from the paper – BA]
From a publisher’s perspective, this all makes sense – Who wants to publish articles that readers will find boring? But from the perspective of science, it’s a disaster. When you make it clear that you’ll only publish “exciting” results, what do researchers do? They decide not to submit papers with “null” findings at all, or to drop the studies that “didn’t work”. They oversell their findings – drawing conclusions that go well beyond the data. They run lots of different statistical analyses and report only the ones that “work”.
None of these things sound particularly bad on their own, but drop-by-drop, they poison the scientific literature. To see how, let’s take an example from a different field altogether: antidepressant drugs. In the figure below (from de Vries et al, 2018, creative commons license), each dot represents a study.
If we look at the raw data (panel a), 50% of studies find a beneficial effect of antidepressant drugs, 50% don’t. But by the time the “null” findings of no benefit for the drugs have been hidden away in a file drawer (panel b), dropped from multiple-study papers (panel c) or given a dose of spin (panel d), it looks like the evidence for these drugs is overwhelming. To cap it all, the few null findings that do make it into publication are cited less often (panel e).
For child language research, the situation is probably even worse. All drug trials are at least held on a central register, which is how we know about the “missing” studies in the first place. But there is no such register for child language: Studies are going missing, and we don’t even know it.
Language Development Research is designed to address this problem, by committing – in its policies and procedures – to “publish any empirical or theoretical paper that is relevant to the field…and that meets our criteria for rigour, without regard to the perceived novelty or importance of the findings”. To discourage “questionable research practices” we also require (unless there are confidentiality reasons) “all experimental materials, data and analysis code to be made available in a public repository prior to publication”.
Perhaps most important of all, the journal is free to both readers and authors (all of the other major child language journals charge open-access fees – typically £1000 or more – or else restrict access to subscribers only). This means that we have no financial bottom line that could affect decisions regarding what kinds of articles to accept, or how many of them.
How can we run a journal with no income? It’s much easier than the traditional publishers would have you believe: The journal runs on the open-source Janeway platform and – thanks to my co-founder, Brian MacWhinney – is hosted for free by Carnegie Mellon University’s Library Publishing Service. Of course, Carnegie Mellon are bearing some costs, but the amounts pale into insignificance compared with the thousands or, in some cases, millions that Universities spend on journal subscriptions. We don’t pay our reviewers or action editors of course; but then that is true for most traditional journals too (though some do pay action editors a small annual stipend).
So, colleagues, please support this new model by submitting your research to Language Development Research. Not only will you be getting the same peer review that you get from other journals – and for free – you’ll also be supporting a new and – we think – better model of scientific publishing.
To check out our first issue, or to submit your paper please visit www.LanguageDevelopmentResearch.com