Joint Attention and Language Abilities: The Moderating Effect of a Risky Temperament Profile

Document Type


Publication Date



Researchers have demonstrated a persistent relationship between joint attention (JA) and language abilities. For example, 14-month JA is associated with concurrent performance on a word-object association task under control conditions as well as under distraction (Salley et al., 2012). Research has also shown associations between temperament and language. For example, 13-month temperament predicts 20-month productive vocabulary (Dixon & Shore, 1997). It has been suggested that "risky" temperamental profiles, such as when children have high negative affectivity and low effortful control, can especially lead to language delay (Dixon & Smith, 2000). In this investigation, we explored whether temperamental profile might moderate the relationship between JA and language ability. Eighty-three children (32 girls) visited the lab at M = 15.45 months (SD = 1.92 months). Caregivers completed the Infant Behavioral Questionnaire-Revised (IBQR) and the MacArthur-Bates Communicative Development Inventory: Words and Gestures (MCDI-WG). The IBQ-R produced three overarching superdimensions: surgency, negative affectivity, and effortful control, two of which were used to identify children as temperamentally "at-risk" or "buffered." Total receptive vocabulary was derived from the MCDI-WG. Temperamental risk was defined as scoring high in negative affectivity and low in effortful control; while temperamental buffering was defined as scoring low in negative affectivity and high in effortful control. JA was measured using a Brooks and Meltzoff (2005) type gaze-following procedure, with some gaze-following trials subjected to an exogenous distractor (Elmo video playing in the background), and others undistracted. JA was defined as total infant looking time to experimenter-fixated target objects. Overall, receptive vocabulary was correlated with JA in both nondistracted (r = .30, p = .01) and distracted conditions (r = .25, p = .04). Although infants did not differ in either JA or receptive vocabulary as a function of temperamental profile, we found that the correlation between JA and receptive vocabulary did. Specifically, JA was not associated with receptive vocabulary for children with risky temperament (see Table 1). But there was a large and positive association between receptive vocabulary and JA among children with a buffered temperament, regardless of distraction condition. Moderation analyses confirmed that temperamental risk was a significant moderator of the JA-receptive vocabulary relationship (moderator control = -1.09, p = .006; moderator distraction = - 0.75, p = .01). These results are partially consistent with theoretical expectations, although they need be supported by further research. They suggest, for example, that the JAlanguage relationship may be attenuated or enhanced depending on infants' temperament profiles. It may be that children who are low in negative affectivity and high in effortful control can maximize their allocation of attention both in the service of following the gaze of a social partner, and in making word-referent mappings during social exchange. The fact that the JA-receptive vocabulary correlation appeared unaffected by the presence of an exogenous distractor raises the possibility that one means through which a buffering temperamental profile may operate is by desensitizing children to ambient environmental distractions during real-time acquisition of linguistically relevant stimuli.


Philadelphia, PA

This document is currently not available here.