We performed two primary analyses. The first was a preliminary analysis investigating the assumption of independence of the utterances produced by the child. This investigation found little evidence of dependence, and so we performed further analyses under the assumption the utterances formed an independent set. The second analysis investigated the research hypothesis by postulating an unobserved variable indicating the child's current cognitive capacity. This capacity structure involved two parts. The first part, the capacity model models the relationship between capacity and the conversational demands while the second part, the capacity function models the relationship between capacity and morpheme length. One can consider how the conversational demands lower the child's unobserved capacity by investigating the capacity model while one can consider how lower capacity is associated with morpheme length by investigating the capacity function.
Our results indicate the conversational technique of pausing to be strongly associated with longer morpheme lengths. There is limited evidence the conversational demands of timing or focus are associated with morpheme length. There is also some evidence that as capacity decreases, the child loses the ability to produce extremely long utterances faster than the child loses the ability to produce medium length utterances.