Experimentally Induced Increases in Early Gesture Lead to Increases in Spoken Vocabulary

Abstract

Differences in vocabulary that children bring with them to school can be traced back to the gestures they produced at the age of 1;2, which, in turn, can be traced back to the gestures their parents produced at the same age (Rowe & Goldin-Meadow, 2009a). We ask here whether child gesture can be experimentally increased and, if so, whether the increases lead to increases in spoken vocabulary. Fifteen children aged 1;5 participated in an 8-week at-home intervention study (6 weekly training sessions plus follow-up 2 weeks later) in which all were exposed to object words, but only some were told to point at the named objects. Before each training session and at follow-up, children interacted naturally with caregivers to establish a baseline against which changes in communication were measured. Children who were told to gesture increased the number of gesture meanings they conveyed, not only during training but also during interactions with caregivers. These experimentally induced increases in gesture led to larger spoken repertoires at follow-up.

ACKNOWLEDGMENTS

We thank A. Evenson and G. Spharim for help in coding, as well as S. Beilock, J. Iverson, K. Kinzler, and S. Levine for comments on earlier drafts.

Notes

¹A gesture intervention in which a gesture was modeled for the child and the child was then encouraged to use it has been shown to successfully increase gesturing outside of the training context in children with autism (Ingersoll, Lewis, & Kroman, 2007). Our study evaluates whether the strategy also works with typically developing children.

²If we use a standard nonparametric Mann-Whitney test for our four Part 1 analyses, we obtain similar results.

³As specification checks, we reran analyses of effects of instruction to gesture on child gesture in experiment, this time including covariates, and we obtained similar results: a) pretest child speech as a covariate (B = 11.967, SE = 2.244, β = .815, p < .001; randomization test, p < .001); b) pretest child gesture as a covariate (B = 12.424, SE = 2.387, β = .846, p < .001; randomization test, p < .001). For both analyses, we obtained similar results to what we would have obtained if we had also included the residual contrast described in Part 2.

⁴We reran analyses of effects of instruction to gesture on child gesture with caregiver, this time including covariates, and we obtained similar results: a) pretest child speech as a covariate (B = 0.634, SE = 0.253, β = .575, p = .014; randomization test, p = .011); b) pretest child gesture as a covariate (B = 0.530, SE = 0.207, β = .480, p = .013; randomization test, one-tail p = .032, two-tail p = .068). We obtained similar results to those we would have obtained if we had also included the residual contrast in each analysis.

⁵We reran analyses of effects of instruction to gesture on child gesture composite, this time including covariates, and obtained similar results: a) pretest child speech as a covariate (B = 0.475, SE = 0.090, β = .810, p < .001; randomization test, p < .001); b) pretest child gesture as a covariate (B = 0.453, SE = 0.095, β = .773, p < .001; randomization test, p < .001). We obtained similar results to those we would have obtained if we had also included the residual contrast in each analysis.

⁶If we include pretest child speech as a covariate in this analysis, we find an effect of instruction-to-gesture group on child speech with caregiver at follow-up (B = 1.533, SE = 0.812, β = .430, one-tail p = .042, two-tail p = .083; randomization test, one-tail p = .062, two-tail p = .124). If we instead include pretest child gesture as a covariate, there is a marginally significant effect (B = 1.333, SE = 0.889, β = .374, one-tail p = .080, two-tail p = .160; randomization test, one-tail p = .083, two-tail p = .157). Results for both analyses are similar if we also include the residual contrast.

⁷If we use the CDI toddler form, which was administered at follow-up, as our speech outcome measure in Path b, we find similar effects. Child gesture composite is a significant predictor of speech outcome, as measured by the CDI (B = 41.555, SE = 22.521, β = .817, one-tail p = .0475, two-tail p = .095; randomization test, p = .002). The Part 1 analysis of instruction effects is not significant, t(13) = 0.394, p = .350 (randomization test, M_difference = 9.7, p = .366).

⁸We also performed an instrumental variables analysis as a complementary method of testing indirect effects. In this analysis, random assignment is our instrument (in our study, instruction to gesture) and is thought to address the issue of omitted variables that could be confounders (Angrist & Krueger, 2001). We tested significance of the indirect effect using a randomization test, and both one-tail and two-tail tests were marginally significant (one-tail p = .098, two-tail p = .099).

Though not conclusive, these findings suggest a causal effect of manipulated child gesture on child speech outcome, a result confirmed in our indirect-effects analysis.

⁹As reported earlier, if we instead control for our parent-reported pretest child language measures (pretest child spoken vocabulary and pretest child comprehension vocabulary), we obtain similar results.

¹⁰We also ran a randomization test using an alternative method of permuting the data that was similar to the one used in Part 1. We randomly permuted the instruction-to-gesture group but used actual values for all other predictors and outcome. Our findings are unchanged (one-tail p < .001).

Additional information

Notes on contributors

Eve Sauer LeBarton

Eve Sauer LeBarton is now at the Johns Hopkins School of Medicine and the Kennedy Krieger Institute.