Chapter 17 Writing up results

17.1 Scientific venue

In this video we are not going to talk about how you will frame your paper because that depends on the kind of journal you are aiming to. That said, there are some important things to keep in mind and some key references you will want to remember when you are going to write up your results. So you’ll typically have: a participants section, where I recommend you mention not only the data you are including but all the data you collected, meaning you’ll mention data exclusion (e.g recordings that are too short with respect to others; or recordings that have mostly silence on it, meaning that probably those families put the recorder in a drawer, while exercising their right of withdraw; etc.).

Then you’ll have the equipment section, where you’ll want to say what device you were using and you want to mention also clothing and attachment. Then the next section will be analysis: here you can mention the software you used to analyse (e.g. for LENA, your reference is the LENA technical reports, in particular the one that explains how the hardware and software work, etc.). You’ll probably also want to mention studies on reliability, which is the accuracy of the system. In particular we have a meta analysis that looked at that recently, but that meta analysis was not complete, so if for your particular study you have categories that are most important, for instance if you are interested in the key child, the you’ll need to have the accuracy of key child and this is not something that the meta analysis could look at in much detail because not too many studies report this but there is one paper by me [ALEX] and other people that was published last year, which reports key child’s performance.

In the software section, you will have also manual analysis, so you should provide the link to the human annotation training material that you have, and this is because this section you’ll make about human annotation will effect your data so that should be documented. If You followed the ACLEW documentation scheme version, then there is some reliability data from different labs that you can cite. The said, you cannot assume you got accuracy at the same level, so this is also something you will wan to point out.

17.2 Families

Typically (at least in my lab), we provide summaries with the results of the studies to the families who participated in it, so we thought to give you some tips on how to explain results to families. Some parents want to get their own data, so you can produce an individualised graf which shows the proportion of silence, vocalisations etc., so a sort of overview. With the Lena you can easily do it, by printing a copy of the relevant section.

Some parents want more, so for example a comparison between their quantity of speech and the others and it’s up to you to decide whether it’s reasonable to give them this info. Something we found here in France is that, particularly for methods like this one which is still being developed and where we don’t have a very large normed population on which to base our measures, then providing percentages within a small population might generate some anxiety on parents (for example if you have a community of 30 children, someone is going to be the 30th, and this is perceived as discouraging).

Another thing you might want to do is provide them with group results, which might be a summary of your scientific paper. You can also provide them with the link to the actual paper and this is important because they feel they are represented in that publication. The usual advice is not to use highly technical phrases, instead using a language that is understandable by someone who doesn’t have the technical background.

17.3 Communities

In the piloting stage we mentioned that we like thinking of LFR as a sample of a population, and so we always talk about communities. Communities don’t necessarily have a physical identity but they might be seen as a superset of all of the families that participated in your study. So, if this sounds something that applies to the work you are doing, you can think of ways to get back to the community and not just the single families who participated, by for instance creating a summary of the results or short presentations of your study and what you found so far and distribute it to associations or other entities related to that community. This could be via a newspaper article, if you know a journalist who can help you with that, or it could be an association representing your community (if they have a self identity) so you could provide them with information on your study. So we think that this is important because long-form recordings are capturing people’s lives to a certain extent, and they are not just capturing the single families’ lives but they are samples of a population so it’s important for the identity to which they belong. So in that sense, it is useful to send the message that you have clear who this community is and ideally this is something you are going to discuss in your piloting stage as well, so if the community specifically asks you for some specific goal, this is also something that you need to mention in your write up.

17.4 Resources

Example scientific paper not using LENA: Casillas, Marisa, Penelope Brown, and Stephen C. Levinson. “Early language experience in a Tseltal Mayan village.” Child Development 91.5 (2020): 1819-1835. pdf
Example scientific paper using LENA: Ma, Y., Jonsson, L., Feng, T., Weisberg, T., Shao, T., Yao, Z., … & Rozelle, S. (2021). Variations in the Home Language Environment and Early Language Development in Rural China. International Journal of Environmental Research and Public Health, 18(5), 2671.link
Example scientific paper using LENA looking at methodology: McDonald, M., Kwon, T., Kim, H., Lee, Y., & Ko, E. S. (2021). Evaluating the Language ENvironment Analysis System for Korean. Journal of Speech, Language, and Hearing Research, 64(3), 792-808. link
Example information for parents (French)