Validity testing a new Babble Checklist.

Amy Bidgood and H Ronayne presented this paper at the ESRC International Centre for Language and Communicative Development Annual Conference in Manchester, UK.


Children’s babble has been shown to be important in their language development, both in terms of the speed of their vocabulary acquisition (e.g. McGillion et al., 2016) and the particular words they produce first (e.g. McCune & Vihman, 2001). Babble production also affects how babies listen to language (e.g. DePaolis et al. 2011). Although some consonants appear more frequently in early babble that others – /d/ and /m/ are more common than /l/ and /s/, for example – there are large individual differences in the amount of babble children produce, the age at which they start to babble, and the consonants they produce first. In order to understand these differences, the majority of research has used recordings of children’s speech, which is then phonetically transcribed. This process is time-consuming and requires specialist training. It can therefore be prohibitively expensive and impractical for many research studies. As part of the Language 0-5 Project, we have developed a new Babble Checklist – a parental report measure that asks about the frequency with which children produce each consonant. If this tool is to be usefully used for research, however, we first need to establish whether parents are able to complete the Checklist accurately. In order to test the validity of the Babble Checklist, we recorded and phonetically transcribed the babble of 20 babies aged between 9 and 12 months. The babies were recorded during a 30-minute lab-based play session, with a parent. Phonetic transcription was carried out following the session. Before recording took place, parents completed the Babble Checklist. We are therefore able to compare the consonants that parents reported their children as producing with what their children said during the recording session.As some previous research studies have suggested that parents are unable to complete similar checklists accurately (e.g. DePaolis et al. 2011), we split parents into two group. Parents in Group A completed the Checklist only once, immediately before the play session. Parents in Group B completed the Checklist three times, each one week apart, with the final Checklist completed immediately before the play session. This allows us to test whether accuracy increases as parents become more familiar with the Checklist. If the validity testing demonstrates that parents are able to accurately complete the Babble Checklist, we will have created a cheap and easy method for researchers and practitioners to collect information about children’s early babble – an important step in their language development.