Through babies’ eyes: Practical and theoretical considerations of using wearable technology to measure parent–infant behaviour from the mothers’ and infants’ view points

Highlights

• •

  Ecologically valid measurement of mother–infant behaviour is difficult.

• •

  Observations often rely on short snapshots with a researcher present.

• •

  Wearable cameras can be worn by mothers and infants alone at home.

• •

  Home interactions were compared with an observation with a researcher.

• •

  More negative maternal behaviour was recorded using ‘headcams’ alone.

Abstract

Aims

To explore the utility of first-person viewpoint cameras at home, for recording mother and infant behaviour, and for reducing problems associated with participant reactivity, which represent a fundamental bias in observational research.

Methods

We compared footage recording the same play interactions from a traditional third-person point of view (3rd PC) and using cameras worn on headbands (first-person cameras [1st PCs]) to record first-person points of view of mother and infant simultaneously. In addition, we left the dyads alone with the 1st PCs for a number of days to record natural mother–child behaviour at home. Fifteen mothers with infants (3–12 months of age) provided a total of 14 h of footage at home alone with the 1st PCs.

Results

Codings of maternal behaviour from footage of the same scenario captured from 1st PCs and 3rd PCs showed high concordance (kappa >0.8). Footage captured by the 1st PCs also showed strong inter-rater reliability (kappa = 0.9). Data from 1st PCs during sessions recorded alone at home captured more ‘negative’ maternal behaviours per min than observations using 1st PCs whilst a researcher was present (mean difference = 0.90 (95% CI 0.5–1.2, p < 0.001 representing 1.5 SDs).

Conclusion

1st PCs offer a number of practical advantages and can reliably record maternal and infant behaviour. This approach can also record a higher frequency of less socially desirable maternal behaviours. It is unclear whether this difference is due to lack of need of the presence of researcher or the increased duration of recordings. This finding is potentially important for research questions aiming to capture more ecologically valid behaviours and reduce demand characteristics.