Bar camp: Making use of accumulated data in infant language acquisition research

Submit Your Questions

News:

Slides of our presentations are posted in our file cabinet.

Here is a practical introduction to creating a community-augmented meta-analysis.

The bar camp took place 8. July 2014 at the Erasmus building of the Radboud University Nijmegen

How to get to Radboud University?

Questions have been collected here: Questions

About:

Two PhD defenses on the topic of early language acquisition take place in Nijmegen in early July (Christina Bergmann defends on 7. July, Sho Tsuji on 9. July), and we are therefore planning a workshop around these events.

The format will be an interactive “bar camp” on the gains and pains of data accumulation & meta-analysis in infant language acquisition research: You suggest questions/topics that you’d like to discuss, and we prepare a workshop structure.

Workshop theme

Data gathered from infant studies are often noisy, because possible experimental manipulations are limited and only a small number of data points can be gathered in each study. Moreover, comparing results across studies is difficult given differences in methods, age-groups, and language backgrounds.

In order to gain both an overview of and additional insights into the current state of language acquisition research, we have recently created two databases accumulating infant language acquisition experiments (on speech segmentation: https://sites.google.com/site/inworddb/home; on vowel discrimination: https://sites.google.com/site/inphondb/home). Meta-analyses based on these accumulated data have shown that effect sizes vary by method (e.g., headturn preference versus central fixation), suggesting that extra caution is needed when interpreting infants’ success and failure patterns in different paradigms. Additionally, each dataset confirmed some widespread assumptions (e.g., native vowel discrimination increases with age, words at sentence edges are better recognized), and challenged others (non-native vowel discrimination decreases with age, switch from familiarity to novelty preferences with age).

In addition to answering overarching questions, our database tool can be leveraged to decide on the design of a specific study, for instance by power calculations. This characteristic gains importance in times where the replicability of effects and new ways to report statistical results are discussed in the field of psychological sciences.

Your questions

Despite a lot of positive feedback, usage of and contributions to our extant databases are still sparse. We would like to hear your opinion on this approach in general, as well as specific topics you’d be interested to get an answer to. Questions can range from

“How important are meta-analytic approaches, replications, careful, power-analysis-based experimental designs in the field of language development?”

to

“How can I find out which contrast is easy to discriminate by 6-month-old infants?”

Questions do by no means have to be restricted to what is covered in the current databases, but can also serve as a guidelines to creating extensions.

The day of the workshop

We will prepare slots in which the most popular topics are discussed, as well as hands-on slots where we directly use our database and meta-analytic tools to answer specific questions. We also aim to formulate a set of new questions that should be tackled in additional meta-analyses together with the workshop participants.