Content and construct validation of language acquisition enunciative signs in the second year of life.

Abstract

To analyze the content and construct validity of enunciative signs of language acquisition for children aged 13 to 24 months. The signs created were subjected to an assessment of clarity and relevance by six expert judges in language acquisition from an enunciative perspective. Based on their judgment, an experimental version was produced that was applied to the analysis of videos of mother-baby / examiner interactions, lasting 15 minutes, of 77 mother-baby dyads aged 13 to 18 months and 89 dyads in the age range aged between 19 and 24 months. The validity of reliability and internal consistency was performed by analyzing 10% of the sample by two speech therapists. The construct validation was carried out by the factorial analysis carried out on the total sample. The data were analyzed in Statistica 9.1 and PASW 17.0. All signs were considered clear and relevant by the expert judges. The reliability analysis showed an almost perfect agreement (0.8 ≤ Kappa ≥ 1.0) for most signs. The internal consistency for Phase 3 showed alpha = 0.771 considered high and Phase 4 presented alpha = 0.917 bordering on very high. The factor analysis of phase 3 revealed 2 factors, explaining 59.1% of the total variance and phase 4 revealed 1 factor, explaining 75.7%. Content and construct validity were observed for five of the twelve signs in Phase 3 and for all signs in Phase 4.