Abstract
Evidence that early parent-child conversation supports early language development suggests a need to understand factors that account for individual differences in parent-child conversation engagement. Whereas most studies focus on demographic factors, we investigated the role of toddler temperament in a longitudinal study of 120 economically strained families. Specifically, we investigated the degree to which toddlers’ negative affectivity and effortful control, considered together as in a composite reflecting challenging temperament, accounted for variability in parent-toddler conversation engagement, and whether the frequency of that engagement mediated associations between toddler temperament and toddler expressive language skills. Toddler challenging temperament (i.e., high negative affectivity and low effortful control) and parent-toddler conversation engagement were measured at 18 and 30 months. Toddler expressive language skills were measured at 18, 24, and 36 months. As expected, a path model indicated inverse relations between toddler challenging temperament and concurrent parent-toddler conversation engagement at both 18 and 30 months. Unexpectedly, there were no direct associations between toddler challenging temperament and toddler expressive language skills either concurrently or longitudinally. However, we found indirect effects of toddler challenging temperament on later toddler expressive language skills via parent-toddler conversation engagement. Findings highlight the importance of considering toddler temperamental characteristics in addition to family demographics as important factors that account for variability in parent-toddler conversation engagement.
Keywords: temperament, conversation, parent-child communication, language
Parent engagement with their young children in contingent, reciprocal conversation is vital to supporting early language development and nurturing children to become competent communicators. Evidence suggests that parent-child conversational turns, over and above parental language input, account for unique variance in child language outcomes (e.g., Donnelly & Kidd, 2021; Zauche et al., 2016; Zimmerman et al., 2009), such that young children who experience more frequent conversation engagement with their caregivers exhibit better expressive and receptive skills later in life. Parent-child conversation engagement may be especially important in the early toddler period (Gilkerson et al., 2018), a developmental stage characterized by rapid growth in verbal abilities (Camaioni, 2001). In the present investigation, we conceptualize conversation engagement as semantically contingent verbal turns between parents and their toddler. The semantic contingency of conversational turns between a caregiver and child is particularly important, supporting language development by tailoring communication to the child’s current attention, interests, and developmental level, and by creating a social feedback loop, in which children are both exposed to language and provided a context to practice and receive feedback about their communication (Beiting et al., 2022; Casillas, 2014; Tamis-LeMonda et al., 2014; Warlaumont et al., 2014).
Given the importance of early conversations for language outcomes, it is critical to understand the factors that account for variability in parent-child conversation engagement. Decades of research have focused on socioeconomic factors, such as family income or parent education. Germinal work by Hart and Risley (1995) and numerous research programs since have generally found that parents with the resources afforded to those of higher socioeconomic status tend to talk their children more frequently, initiate more conversations with their children, and respond more frequently to their children’s verbal communications, in addition to exposing their children to a wider vocabulary (e.g., Hoff, 2006; Sohr-Preston et al., 2013). Yet more recently, this work has been critiqued as highlighting differences between socioeconomic groups and neglecting the variability within these groups (Rowe, 2018; Sperry et al., 2018). To address this gap, in the present study, we examine parent-toddler conversation engagement within an economically strained sample and examine factors that account for variability in conversation engagement within this group. Given that children’s own characteristics are known to contribute to parenting and parent-child interactions (e.g., Belsky, 1984; Taraban & Shaw, 2018), and given evidence of interrelations between young children’s temperamental traits and their language abilities, we consider toddler temperamental traits as important factors that may contribute to parent-toddler conversation engagement and, in turn, toddler language outcomes. Specifically, we examine concurrent and longitudinal relations among toddler language ability, toddler temperament, and parent-toddler conversation engagement from 18 to 36 months.
Toddler Temperament and Relations to Language and Conversation
Temperament refers to “constitutionally based individual differences in reactivity and self-regulation, in the domains of affect, activity, and attention” (Rothbart & Bates, 2006, p. 100). In the present study, we consider two related temperamental traits, negative affectivity and effortful control, which are temperament dimensions reflecting reactivity and regulation, respectively (Rothbart & Bates, 2006). Negative affectivity refers to the frequency and intensity of negative emotional reactions to changes in the environment (Rothbart & Bates, 2006). Effortful control refers to the ability to suppress a dominant response in favor of a nondominant response to modulate emotion, attention and behavior (Kochanska et al., 2000; Rothbart & Bates, 2006). Conceptually and empirically, these traits are intricately related, although often studied separately (Rothbart & Bates, 2006). Negative affectivity and effortful control tend to be inversely associated with one another; young children rated as higher in effortful control tend to be rated as lower in negative affectivity, and are often observed to express anger more slowly, less frequently, and less intensely (Rothbart & Rueda, 2005; Tan et al., 2013; Vroman & Durbin, 2015). Thus, effortful control can be considered to regulate or modulate negative affectivity (Rothbart & Rueda, 2005).
These traits may be particularly important to consider as predictors of communication in the early toddler period (the second and third year of life), when expressions of negative affect peak (e.g., Barry & Kochanska, 2010; Lewis et al., 2004), effortful control emerges (Posner & Rothbart, 2000), and parental expectations for self-regulation and for verbal rather than emotional communication increase (Brownell & Kopp, 2010). There is limited evidence regarding how these two dimensions relate to parent-toddler conversation; however, indirect evidence from two literatures suggests links. First, both dimensions have been linked to early child language abilities; although the mechanisms underpinning these associations are unclear, one possibility is that young children’s temperamental traits related to reactivity and regulation contribute to language outcomes via their impact on conversation engagement. Second, both dimensions have been linked to global parenting qualities such as sensitivity or responsiveness. In the sections below, we summarize what is known about associations of negative affectivity and effortful control with language outcomes, as well as with conversation engagement and related parenting constructs. Notably, although we will consider negative affectivity and effortful control together in the present study, most of this literature treats them separately.
Temperament and Language Outcomes
Consistent evidence suggests that toddlers with higher effortful control exhibit greater concurrent receptive and expressive language skills (e.g., Eisenberg et al., 2010; Kang et al., 2022; Karrass & Braungart-Rieker, 2003; Kucker et al., 2021; Usai et al., 2009, 2020) and develop better language skills over time (Guedes & Cadima, 2022; Spinelli et al., 2018). On the other hand, negative affectivity is generally associated with risk for poorer language outcomes, although the evidence is more mixed. Numerous studies suggest that high negative affectivity is associated with poorer language skills both concurrently (e.g., Garello et al., 2012; Kubicek & Emde, 2012; Kucker et al., 2021; Noel et al., 2008; Nozadi et al., 2013; Pérez-Pereira et al., 2016; Salley & Dixon, 2007) and longitudinally (e.g., Cioffi et al., 2021; Ollas et al., 2020; Pérez-Pereira et al., 2016). However, some studies report no relation between negative affectivity and language (e.g., Ishikawa-Omori et al., 2022; Laake & Bridgett, 2018; Ollas et al., 2020), and a few find relations in the opposite direction, such that higher negative affectivity predicts the development of better language skills (e.g., Moreno & Robinson, 2005; Nozadi et al., 2013; Spinelli et al., 2018). It is possible that wide variability in measurement across studies accounts for some of these mixed findings. For example, unexpected relations between negative emotionality and language outcomes are more common in studies where child negative affect is observed (e.g., Moreno & Robinson, 2005; Nozadi et al., 2013) rather than reported by parents, and in studies measuring emotionality in infancy (e.g., Moreno & Robinson, 2005; Spinelli et al., 2018). Moreover, rather than considering negative affectivity in isolation, it may be important to account for factors such as the temperamental propensity to self-regulate emotional reactions (i.e., effortful control) in conjunction with negative affectivity to understand implications for language development and communication, a possibility we explore in subsequent sections.
Thus, in general, the extant literature suggests that toddlers who are higher in negative affectivity and/or lower in effortful control are at risk for poor language outcomes. Although there may be many mechanisms at play (see Conture et al., 2013), one possibility is that aspects of temperament related to reactivity and regulation may contribute to language outcomes via their impact on parent-toddler conversation engagement.
Temperament and Parent-Toddler Conversation Engagement
To date, there is little investigation of the links between child temperament and parent-child conversation engagement. One hypothesis is that toddlers who are more temperamentally challenging (i.e., are high in emotional reactivity and/or low in self-regulation) may discourage, disrupt, or be more difficult to engage in verbal conversation (Kubicek & Emde, 2012). Although limited evidence exists to test this hypothesis, a robust literature links temperamental traits of reactivity and regulation to more global parenting behaviors, such as sensitivity and responsiveness. Toddlers who are more “difficult” or high in negative emotional reactivity tend to receive less contingent, sensitive responses to their cues from their caregivers (for review, see Kiff et al., 2011; Paulussen-Hoogeboom et al., 2007). On the other hand, higher effortful control is associated with more responsive and sensitive parenting (see Kiff et al., 2011 for review) and more reciprocal parent-child interactions (K. M. Brown et al., 2022). These findings suggest that toddlers with more challenging temperament, that is, toddlers who are higher in negative emotional reactivity (negative affectivity) or lower in temperamental regulation (effortful control) may experience less conversation engagement with their caregivers. The limited direct evidence available suggests that infants rated as higher in negative affectivity receive less language input of lower complexity from their mothers during parent-child interactions (Vernon-Feagans et al., 2008). Moreover, both maternal ratings of negative affectivity and the frequency of toddler negative emotion expressions during interactions predict aspects of parent-toddler conversation engagement (Fields-Olivieri et al., 2020; Ontai & Virmani, 2010), such that higher negative affectivity predicts lower conversation engagement. Thus, the limited existing evidence supports the hypothesis that negative affectivity may disrupt or deter parent-child conversation engagement. No known study has examined the relation of effortful control to conversation engagement, although two studies find no relation between effortful control and parent language input (Kang et al., 2022; Spinelli et al., 2018).
Need for Transactional, Longitudinal Models
In sum, the extant literature suggests relations among parent-toddler conversation engagement, toddler temperament, and toddler language skills. Yet, there is a lack of longitudinal examination of these factors at multiple time points to elucidate transactional relations among them and to test whether and how child temperament and parent-child conversation engagement relate to one another and may shape child language development. Although the research question was slightly different, one known study examined whether toddler emotional reactivity (observed anger) mediated the effect of parenting (sensitivity) on language outcomes at two timepoints; this mediating effect was supported among concurrent associations at 30 months only (Nozadi et al., 2013). While this study considered the impact of parent behaviors and child characteristics in shaping language development over time, there stills remains a gap in the literature around the conceptualization of such interconnected relations and consideration of the role of parent-toddler communication specifically, which shows unique effects on language over and above sensitivity (Hirsh-Pasek et al., 2015). Thus, no known study has examined whether toddler temperamental traits may influence language outcomes via parent-toddler conversation engagement.
Considering Negative Affectivity and Effortful Control Together
Despite their conceptual relatedness, few studies have examined negative affectivity and effortful control together in relation to language outcomes or parent-toddler conversation engagement, although a small number of studies have considered these two temperament dimensions in tandem to predict socioemotional outcomes (Liew et al., 2004; Moran et al., 2013; Muris et al., 2007; Muris & Ollendick, 2005). There is evidence for additive effects of these dimensions on child adjustment (Liew et al., 2004; Muris & Ollendick, 2005), such that the combination of high negative affectivity and low effortful control may be particularly detrimental (Muris et al., 2007; Van Beveren et al., 2019). For example, one study of preschool-aged children found that an “emotion-related effortful control” composite, in which higher scores reflected higher effortful control and lower negative affectivity, predicted more socially appropriate reactions to disappointment and greater social competence (Liew et al., 2004).
In the present study, we take a similar approach, considering negative affectivity and effortful control together as a single construct of challenging temperament, with higher scores reflecting higher negative affectivity and lower effortful control. We labeled the construct as such to distinguish from the early conceptualization of “difficult” temperament (Thomas & Chess, 1977). Although we conceptualize challenging temperament in a similar manner (i.e., considering multiple dimensions of temperament together to identify more challenging combinations of temperamental traits), we do not categorically distinguish children as “easy”, “slow to warm up” or “difficult”. Rather we maintain a dimensional approach by considering toddlers on a continuous scale from less challenging (i.e., low negative affectivity and high effortful control) to more challenging (i.e., high negative affectivity and low effortful control).
The Present Study
In a sample of economically strained families, we examined concurrent and longitudinal associations among parent-toddler conversation engagement, toddler language abilities, and toddler challenging temperament, a construct reflecting high negative affectivity and low effortful control. We assessed parent-toddler conversation engagement and toddler challenging temperament at 18 and 30 months and assessed toddler language abilities at 18, 24 and 36 months. Using path model analysis (see Figure 1 for full conceptual model), we tested four main hypotheses. First, we expected that toddler challenging temperament would be associated with poorer language skills, both concurrently and longitudinally (H1). Second, we expected that toddler challenging temperament would be associated with lower frequency of parent-toddler conversation engagement (H2). Third, we predicted that parent-toddler conversation engagement would significantly predict later toddler language skills (H3). Finally, we expected that parent-toddler conversation engagement would mediate the association between toddler challenging temperament and toddler language skills, that is, that toddler challenging temperament would predict toddler language outcomes indirectly via parent-toddler conversation engagement (H4).
Figure 1. Conceptual Model.
Note. Covariates (parent education, child sex, maternal verbosity and number of children present) not presented in this model
Methods
Participants
Data were drawn from a longitudinal study of 120 economically strained families from rural and semi-rural areas in the Northeastern United States. The data of focus in the present paper were collected from January 2002 to March 2005. Families were recruited at target-child age 18 months. Families were eligible to participate if their total household income, accounting for family size, fell above the federal poverty line but below the national median income. Eligibility was constrained in this way to recruit a sample that was economically strained but was not subject to the risk factors associated with poverty or the resources available to middle class families. Total annual household income (M=$40,572, SD=$14,387) ranged from $15,000 to $70,000, and average income-to-needs ratio was 2.33 (SD=.87) at the time of recruitment. The sample comprised predominantly two-parent families; at 18 months, only two households reported that the target child’s father figure was not involved. Parent education ranged from some high school to completing an advanced degree. Less than half of mothers (42.5%) and less than one third of fathers (31.4%) had completed a college degree. Mothers completed an average of 14.66 years (SD=1.849) of education and fathers completed an average of 14.14 years (SD=1.991) of education.
Target children (78 male, 42 female) were primarily first or only (43.3%) or second-born (40%) children. Most (93.3%) were identified by mothers as White and 6.7% were identified as biracial, consistent with the demographics of the area from which families were recruited.
Procedures
Families were recruited at target-child age 18 months and were assessed semi-annually until age 48 months. Only data from the 18-month, 24-month, 30-month and 36-month assessments are used for the present study. Families participated in a combination of observations in their home and assessments in the laboratory.
At 18 and 30 months, families participated in an unstructured home observation, during which they were observed by a research assistant and asked to behave as normally as possible. Participants were asked to schedule their observation at a time when both parents were expected to be home. Research assistants observed the target child and family in 10-minute intervals (6 at 18 months and 4 at 30 months), and, at the end of each interval, rated child and parent behaviors that are not of interest in the present study. Research assistants also audio recorded these intervals. Following the home observation, audio recordings from two intervals (one with the fewest emotional events and one with the most emotional events, to capture variability in the emotionality of communication segments) were transcribed. Parent-toddler conversation engagement was later coded from these transcripts. Mothers also completed questionnaires assessing toddler temperament at these visits.
At 18, 24, and 36 months, participants were assessed in the laboratory and engaged in a number of challenging and non-challenging laboratory tasks. The “challenging” tasks were designed to assess emotion regulation abilities and are not of interest in the present study. The other tasks are referred to as “non-challenging” given that they were not designed to tax child emotion regulation capacities. Of interest in the present study, speech samples were drawn from three “non-challenging” tasks that target children engaged in with their mother: free play, clean up, and a wordless picture book reading task.
Participants were provided monetary compensation following each visit. Children received a variety of small rewards. The study was approved by the university’s Office of Research Compliance.
Measures
Parent-Toddler Conversation Engagement
Parent-toddler conversation engagement, or the extent to which parents engaged with their toddlers in semantically contingent conversational turns, was coded using a revised version of the Connectedness Coding system (J. R. Brown et al., 1996; Ensor & Hughes, 2008). As the original coding scheme was developed for preschool-aged children, we adapted it by adding codes to account for the less-advanced language abilities of our sample. For example, on average, 40% of toddler utterances at 18 months and 11% of toddler utterances at 30 months were unintelligible; thus, we added a “Gibberish” code to reflect toddlers’ attempts to verbally communicate with their parents that were unintelligible to transcribers or coders. Although this code is not of specific interest in the present study, it is notable because it allows parent turns to be coded as responding to a toddler utterance even if the utterance is unintelligible to the coder.
First, speaker turns were identified. A speaker turn was defined as an utterance of one speaker (mother, father, or target child) bounded by either another speaker’s utterance or by greater than 10 seconds of silence. Only parental utterances directed toward the target child, or target child utterances directed towards either parent, were coded. Once identified, each turn was assigned one of eight codes, which included codes indicating contingent communication (as will be discussed next), and other codes not characteristic of contingent communication, which are not a focus of the present study. Three codes representing contingent communication are of interest in the present study: parent connected turns, parent attempts to understand, and parent initiations. A parent’s turn was identified as a Connected Turn if the parent’s utterance was semantically related to the toddler’s preceding utterance. A parent’s turn was identified as an Attempt to Understand if the toddler’s previous utterance was fully or partially unintelligible, and the parent responded by asking or guessing what the toddler was attempting to communicate. This code was added by our team given the rate of unintelligible utterances at 18 and 30 months. Finally, a parent’s turn was identified as an Initiation if the parent started a new topic either after significant silence or unrelated to the toddler’s previous utterance and successfully elicited a semantically related response from the toddler. Table 1 provides additional explanation and examples of each of these three codes. Notably, a parent utterance following a toddler “Gibberish” turn could be coded as connected if the parent responded to the toddler as if they understood the toddler. Similarly, a parent turn could be coded as an initiation if the subsequent toddler turn was coded as Gibberish. This decision was made given that parents are likely able to understand more of their toddlers’ speech than transcribers or coders, who are not familiar with the child and do not have access to visual cues that may aid interpretation of speech.
Table 1.
Connectedness Codes, Definitions and Examples
| Connectedness Code | Definition | Examples | |
|---|---|---|---|
|
| |||
| Parent Initiation | Parent’s utterance starts a new topic unrelated to the toddler’s previous turn or after significant silence (10+ seconds) and successfully elicits a semantically related response from toddler. | Parent A: Honey, you want a snack now or wait? | Initiation |
| Child A: ‘Nack | Connected | ||
| Note that an utterance that starts a new topic and elicits an unintelligible (Gibberish) response from the toddler can be coded as an initiation if the context cues (i.e., parent’s subsequent turn) indicate the toddler was communicating or attempting to communicate on a semantically related topic. (See Example B) | Parent B: Hey buddy, you all done? | Initiation | |
| Child B: (unintelligible) | Gibberish | ||
| Parent B: (laughs) No, you’re not all done? | Connected | ||
| Parent Connected Turn | Parent’s utterance is semantically related to the toddler’s previous utterance (initiation, connected, or gibberish) | Child C: Doggie! | Initiation |
| Parent C: Yeah, that’s a doggie! | Connected | ||
| Note that a parent’s response to a toddler’s an unintelligible (Gibberish) utterance can be coded as connected if the parent responds as if they understand the toddler (See Example D) | Child D: (unintelligible) | Gibberish | |
| Parent D: Uhhuh, a doggie says woof woof! | Connected | ||
| Parent Attempt to Understand | Parent’s utterance following a toddler’s unintelligible (Gibberish) utterance asks or attempts to guess what the toddler is attempting to say | Child E: (unintelligible) | Gibberish |
| Parent E: What did you say bud? Ball? | Attempt to Understand | ||
Coders were trained to achieve 90% accuracy with the third author and 20% of cases were double coded to evaluate interrater agreement. Interrater agreement ranged from good to excellent for initiations (18m mother ICC=.891, 18m father ICC=.939, 30m mother ICC=.984, 30m father ICC=.983), excellent for connected turns (18m mother ICC=.991, 18m father ICC=.939, 30m mother ICC=.992, 30m father ICC=.993), and moderate to excellent for attempts to understand (18m mother ICC=.779, 18m father ICC=.738, 30m mother ICC=.945, 30m father ICC=.745).
Within time point, initiations, connected turns, and attempts to understand were significantly correlated (range r=.251–.814 for mothers; r= .469–.895 for fathers) and were summed to form a conversation engagement composite for each age. Descriptive statistics for mother, father, and parent initiations, connected turns, attempts to understand, and conversation engagement at 18 and 30 months are presented in Table 2. Intercorrelations among mother, father, and parent conversation engagement are presented in Table 3. Mothers’ and fathers’ conversation engagement codes were unrelated to one another (range r=−.0004–.104, all ps >.05).
Table 2.
Conversation Engagement Descriptives
| Combined Parent | Mother | Father | t-test Mother vs. Father | ||
|---|---|---|---|---|---|
|
| |||||
| M (SD) | M (SD) | M (SD) | |||
|
| |||||
| 1. | 18m Initiations | 9.59 (9.52) | 6.68 (8.95) | 2.89 (4.20) | 3.64*** |
| 2. | 18m Connectedness | 20.31 (19.44) | 13.98 (16.72) | 5.76 (8.71) | 4.26*** |
| 3. | 18m Attempts to Understand | 3.31 (4.56) | 2.17 (3.39) | 0.96 (2.11) | 2.79** |
| 4. | 18m Conversation Engagement | 33.21 (30.19) | 22.82 (27.16) | 9.60 (13.85) | 4.12*** |
| 5. | 30m Initiations | 15.44 (10.01) | 10.38 (7.03) | 5.32 (6.28) | 5.25*** |
| 6. | 30m Connectedness | 50.04 (31.84) | 34.38 (22.68) | 15.92 (16.46) | 6.64*** |
| 7. | 30m Attempts to Understand | 2.60 (2.65) | 1.64 (2.00) | 0.93 (1.44) | 2.90** |
| 8. | 30m Conversation Engagement | 66.51 (39.16) | 44.40 (29.52) | 22.17 (22.82) | 6.47*** |
p < .001.
p < .01.
p < .05.
p < .10 (marginal significance).
Table 3.
Intercorrelations among Mother, Father, and Parent Conversation Engagement at 18 and 30 Months
| 1 | 2 | 3 | 4 | 5 | 6 | ||
|---|---|---|---|---|---|---|---|
|
| |||||||
| 1. | 18m Mother Conversation Engagement | --- | |||||
| 2. | 18m Father Conversation Engagement | .002 | --- | ||||
| 3. | 18m Parent Conversation Engagement | .883*** | .458*** | --- | |||
| 4. | 30m Mother Conversation Engagement | .373*** | −.008 | .260* | --- | ||
| 5. | 30m Father Conversation Engagement | .128 | .202M | .186 | .127 | --- | |
| 6. | 30m Parent Conversation Engagement | .350** | .129 | .306** | .819*** | .665*** | --- |
p < .001.
p < .01.
p < .05.
p < .10 (marginal significance).
The parent-toddler conversation engagement composite was created by summing mother and father initiations, connected turns, and attempts to understand. This composite reflects the extent to which parents successfully initiated and sustained semantically contingent conversation with their toddlers. Of the families with useable connectedness data, nine fathers at 18 months and twelve fathers at 30 months were not present during the home observation and therefore did not engage their toddler in conversation. Given that receiving language input from only one caregiver during portions of the day is a common experience even in two-parent families (Bergelson et al., 2019), if only one caregiver was present during the home observation, the composite reflected only one parent’s turns, and we did not treat these data as missing. This decision was supported in part by finding no significant differences in parent-toddler conversation engagement for observations in which both parents were present (18m Mean=32.433, SD=30.51; 30m Mean=68.57, SD=39.53) compared to observations in which only mothers were present (18m Mean= 41.11, SD=26.94; 30m Mean=53.71, SD=35.41), at either 18 months, t(97)=.822, p=.413 or 30 months, t(99)=−1.32, p=.189. However, in post-hoc analyses, we did re-run the path model including only cases for which both mothers and fathers were present, to assess whether relations differ with one versus two parents present.
Notably, although we are interested in the contributions of both toddler and parents to conversation engagement, the immature expressive language skills and the high rate of unintelligible utterances in our sample precludes identification of connected and initiation turns for many toddler utterances. Thus, we focus on parent turns that reflect engagement. Importantly, each of these parent turn types (initiation, connected turn, attempt to understand) implies toddler engagement. That is, parent initiation must, by definition, be followed by a toddler connected turn or gibberish. Similarly, a parent connected turn, by definition, follows a toddler initiation, connected turn, or gibberish, and a parent attempt to understand follows toddler gibberish.
Toddler Language Ability
Estimates of toddler language ability were derived from child speech samples drawn from the “non-challenging” laboratory tasks at 18, 24, and 36 months. Video recordings of the tasks were transcribed using the Codes for Human Analysis of Transcripts (CHAT) system and subsequently analyzed using the Computerized Language Analysis (CLAN) Program (MacWhinney, 2000) to generate two indices of toddler language ability at each age: mean length utterance (MLU) and Number of Different Words (NDW). Toddler MLU reflects the average number of morphemes per child utterance, and is a measure of syntactic complexity (Pezold et al., 2020). Note that if the child made no intelligible utterances, MLU was considered 0. Toddler NDW reflects the number of unique words spoken by the child during the tasks and is an indicator of expressive vocabulary (Pezold et al., 2020). Within each age point, each language indicator was significantly correlated across the three tasks (range r=.370–.842 for MLU; range r=.440–.716 for NDW), so a single indicator for MLU and a single indicator for NDW, calculated across all three tasks, was used at each age. Descriptive statistics and intercorrelations of toddler MLU and NDW at 18, 24, and 36 months are presented in Table 4. MLU and NDW were significantly correlated at each age (range r=.436–.656). To create a language composite at each age, MLU and NDW were first z-transformed and then summed.
Table 4.
Descriptives and Intercorrelations among Toddler Language Measures at 18, 24 and 36 months
| Whole sample intercorrelations | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
|
| |||||||||||
| Whole Sample M (SD) | Male M (SD) | Female M (SD) | Male vs. Female t-test | 1 | 2 | 3 | 4 | 5 | 6 | ||
|
| |||||||||||
| 1. | 18m MLU | 1.16 (.36) | 1.15 (.42) | 1.18 (.28) | −0.40 | --- | |||||
| 2. | 18m NDW | 10.92 (10.23) | 8.60 (7.64) | 13.77 (12.19) | −2.81 | .436** | --- | ||||
| 3. | 24m MLU | 1.53 (.44) | 1.49 (.41) | 1.58 (.47) | −1.08 | .111 | .330** | --- | |||
| 4. | 24m NDW | 39.44 (23.90) | 36.67 (21.00) | 42.83 (26.84) | −1.41 | .164M | .493** | .656** | --- | ||
| 5. | 36m MLU | 2.48 (.55) | 2.40 (.53) | 2.57 (.56) | −1.72M | .005 | .176M | .378** | .322** | --- | |
| 6. | 36m NDW | 96.88 (30.56) | 90.70 (28.12) | 104.37 (31.97) | −2.44* | .145 | .302** | .276** | .364** | .510** | --- |
Note. MLU= Mean Length Utterance; NDW=Number of Different Words
p < .001.
p < .01.
p < .05.
p < .10 (marginal significance).
Toddler Challenging Temperament
Mothers reported on toddler temperament via the Toddler Behavior Assessment Questionnaire-Revised (TBAQ-R; Goldsmith, 1996; adapted by Rothbart, 1996) following the home visits at 18 and 30 months. In this 105-item questionnaire, mothers rated the frequency at which their toddler had engaged in specific behaviors within the past month on a Likert scale from 1 (Never) to 7 (Always). Two scales were used in the present study. The Negative Affectivity scale reflects the sum of the anger, sadness, social fearfulness and soothability (reversed) subscales. The Effortful Control scale reflects the sum of attentional focusing, attention shifting, inhibitory control, low-intensity pleasure, and perceptual sensitivity. Descriptive statistics and intercorrelations of each scale at each age are presented in Table 5. The Negative Affectivity and Effortful Control scales were significantly inversely correlated at both 18 (r=−.385, p<.001) and 30 (r=−.391, p<.001) months.
Table 5.
Descriptive Statistics and Intercorrelations among Toddler Temperament Dimensions at 18 and 30 months
| Whole Sample M (SD) | Male M (SD) | Female M (SD) | Male vs. Female t-test | 1 | 2 | 3 | 4 | ||
|---|---|---|---|---|---|---|---|---|---|
|
| |||||||||
| 1. | 18m Negative Affectivity | 3.45 (.55) | 3.43 (.54) | 3.47 (.57) | −0.33 | --- | |||
| 2. | 18m Effortful Control | 4.51 (.56) | 4.45 (.59) | 4.59 (.52) | −1.39 | −.385** | --- | ||
| 3. | 30m Negative Affectivity | 3.57 (.54) | 3.49 (.55) | 3.65 (.53) | −1.53 | .684** | −.284** | --- | |
| 4. | 30m Effortful Control | 4.77 (.58) | 4.67 (.58) | 4.88 (.56) | −1.93M | −.333** | .737** | −.391** | --- |
p < .001.
p < .01.
p < .05.
p < .10 (marginal significance).
At each age, a challenging temperament composite was calculated by reverse coding Effortful Control, then z-transforming each scale (Negative Affectivity and reverse-coded Effortful Control) and summing the z-scores. Thus, high scores on the challenging temperament composite reflect toddlers who are high in negative affectivity and low in effortful control.
Notably, it is possible that average scores on the challenging temperament composite could reflect: toddlers who are average in both negative effectivity and effortful control, toddlers who are low in both negative affectivity and effortful control, or toddlers who are high in both traits. To examine this possibility and assess the appropriateness of using the composite in subsequent analyses, we ran a latent profile analysis (LPA) using negative affectivity and effortful control to identify groups of toddlers based on these two temperament dimensions at 18 and 30 months. The full details of these analyses are presented in Supplementary Materials. In short, the results do not support the existence of a group of toddlers that are either low in both negative affectivity and effortful control or high in both traits, and instead reflect groups that are either high in negative affectivity and low in effortful control, low in negative affectivity and high in effortful control, or average in both traits. These results make us more confident in our ability to use the challenging temperament composite and interpret average scores.
Covariates
Four covariates were considered for the path models: parent education and child sex (both reported by mothers at study enrollment), the number of children present during the home observation, and maternal verbosity during the non-challenging lab tasks.
Number of children present.
Given that the presence of siblings is known to influence language input and parent-child conversation (Hoff, 2006), the number of children present during the home visit was covaried at 18 months and 30 months. Coders noted the number of children present during each 10-minute observation interval, which was averaged across the visit. The average number of children present (including the target child) at 18 months was 1.77 (SD=.79), and at 30 months was 1.93 (SD=.92).
Maternal verbosity.
Given that the toddler language measures were derived from a laboratory task in which mothers were present, at each language assessment, we calculated a maternal verbosity score, to account for how talkative mothers were during these tasks. The same CHAT transcripts used to generate toddler language estimates were analyzed using the CLAN Program (MacWhinney, 2000) to calculate the total number of word tokens used by mothers across the non-challenging laboratory tasks, which we refer to as maternal verbosity. A repeated measures ANOVA indicated no significant age differences in maternal verbosity, F (2, 218)=.545, p=.581. The average number of word tokens spoken by mothers across the challenging lab tasks was 946.73 (SD=254.44).
Data Preparation
The data were first examined for outliers and for violations of assumptions of normality. Given the positive skew of the conversation engagement variables, these variables were square root transformed to achieve a normal distribution.
Missing Data
Data on parent-toddler conversation engagement was missing for 21 participants (17.5%) at 18 months and 19 participants (15.8%) at 30 months due to technical difficulties with audio recordings. Toddler temperament data was missing for 2 participants (1.7%) at 18 months and 6 participants (5%) at 30 months. Toddler language data was missing from 2 participants (1.7%) at 18 months, no participants at 24 months, and from 5 participants (4.2%) at 36 months. Little’s MCAR test indicated that these data were missing completely at random, χ2 =87.947, p=.207. To address missing data and maximize statistical power, we conducted path analyses using full information maximum likelihood estimation.
Statistical Power
A sensitivity analysis (Soper, 2023) indicated that with a sample size of 120, we had sufficient power (≥.80) to detect effects as small as .085.
Data Analytic Plan
A path model was performed using full information maximum likelihood estimation with MPlus (Version 8.5) to test the concurrent and longitudinal relations among toddler challenging temperament, toddler language, and parent-toddler conversation engagement. Parent education and child sex were entered as covariates. The effects of maternal verbosity on within-timepoint toddler language estimates, and the effects of number of children present on within-timepoint variables were also controlled for in the model.
In addition to all cross-lagged paths, all autoregressive paths were included in the model to account for stability of variables across age. Within-time correlations between conversation engagement and toddler language, and between toddler language and challenging temperament, were also included. Given that a primary question was whether toddler temperament contributes to later language development via its impact on concurrent parent-toddler conversation engagement, we included direct within-time paths from toddler temperament to parent-toddler conversation engagement at 18 and 30 months. This allowed us to test for indirect effects of toddler challenging temperament on language via concurrent conversation engagement.
Indirect effects were tested using bootstrapping procedures (reiterations of 5,000) to generate 95% bias-corrected confidence intervals (Hayes et al., 2011). Indirect effects were considered significant if their bias-corrected confidence interval did not contain zero.
Next, to rule out an alternative hypothesis, specifically, that there are indirect effects of conversation engagement on toddler language via temperament, the path model was re-run, reversing the concurrent path from toddler temperament to parent-toddler conversation engagement. Indirect effects (i.e., conversation to temperament to language) were examined using the same procedures described above.
Results
Descriptive Statistics
First, age-related changes in all study variables were examined. Parent-toddler conversation engagement increased significantly between 18 and 30 months, t(82)=−7.270, p<.001, nearly doubling in average turn frequency from 33.048 to 64.783. Toddler negative affectivity significantly increased between 18 and 30 months, t(111)=−2.788, p=.006, as did toddler effortful control, t(111)=−7.062, p<.001. A repeated measures ANOVA indicated that there was a significant effect of age on MLU, F(2, 224)=298.921, p<.001, and on NDW, F(2, 224)=566.849, p<.001, such that both MLU and NDW significantly increased between 18 and 24 months and between 24 and 36 months.
Bivariate correlations of toddler negative affectivity and effortful control with conversation engagement and language variables are presented in Table 6. Notably, negative affectivity and effortful control were associated at similar magnitude, but in opposite directions, to parent-toddler conversation engagement at each age. Bivariate correlations among all continuous variables included in the path models (parent education, number of children present, maternal verbosity, the parent-toddler conversation engagement composites, the toddler language composites, and the toddler challenging temperament composites) are found in Table 7.
Table 6.
Correlations of Toddler Temperament Dimensions with Language and Conversation Engagement
| 18m Language | 24m Language | 36m Language | 18m Engagement | 30m Engagement | |
|---|---|---|---|---|---|
|
| |||||
| 18m Negative Affectivity | .003 | −.032 | −.016 | −.223* | −.196M |
| 18m Effortful Control | .140 | .228* | .113 | .226* | .193M |
| 30m Negative Affectivity | −.052 | −.125 | −.058 | −.218* | −.331** |
| 30m Effortful Control | .056 | .197* | .162M | .180M | .367*** |
p < .001.
p < .01.
p < .05.
p < .10 (marginal significance).
Table 7.
Intercorrelations among Variables Included in the Path Model
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | ||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
| ||||||||||||||
| 1. | 18m Parent Education | --- | ||||||||||||
| 2. | 18m # Children Present | −.098 | --- | |||||||||||
| 3. | 30m # Children Present | .001 | .699*** | --- | ||||||||||
| 4. | 18m Maternal Verbosity | .192* | .028 | −.022 | --- | |||||||||
| 5. | 24m Maternal Verbosity | .116 | −.095 | −.236* | .444** | --- | ||||||||
| 6. | 36m Maternal Verbosity | .230* | −.189* | −.192* | .433** | .558*** | --- | |||||||
| 7. | 18m ConversationEngage | .225* | −.252* | −.197M | .236* | .286** | .183M | --- | ||||||
| 8. | 30m ConversationEngage | .145 | −.300** | −.378*** | .188M | .332*** | .275** | .306** | --- | |||||
| 9. | 18m Toddler Language | .140 | −.067 | .053 | .135 | .026 | .037 | .359*** | .209* | --- | ||||
| 10. | 24m Toddler Language | .195* | −.008 | −.091 | .155M | .093 | .028 | .548*** | .149 | .356*** | --- | |||
| 11. | 36m Toddler Language | .146 | .077 | .044 | .135 | .092 | −.184M | .243* | .229* | .210* | .425*** | --- | ||
| 12. | 18m Chall. Temperament | .016 | .232* | .200* | −.229* | −.122 | −.051 | −.267** | −.234* | −.082 | −.156M | −.076 | --- | |
| 13. | 30m Chall. Temperament | −.021 | .169M | .222* | −.159M | −.148 | −.010 | −.235* | −.415*** | −.065 | −.193* | −.133 | .738*** | --- |
p < .001.
p < .01.
p < .05.
p < .10 (marginal significance).
Path Analysis Results
First, a model including all possible paths of all covariates was run. This model was a poor fit to the data, χ2 (32) = 44.115, p = .075, CFI = .941, RMSEA = 0.059, SRMR = .079. In this initial model, three covariates (child sex, 18-month maternal verbosity, and 24-month maternal verbosity) had no significant paths to any variables in the model. These variables were subsequently trimmed. All other covariates and covariate paths were retained in the final model.
The final path model examining concurrent and longitudinal relations among toddler challenging temperament, toddler language ability, and parent-toddler conversation engagement (controlling for parental education, number of children present, and 36-month maternal verbosity) fit the data well, χ2 (20) = 25.193, p = .1941, CFI = .975, RMSEA = 0.048, SRMR = .047, see Figure 2. Although not presented in Figure 2, the significant covariate paths in this final model were: Parent education to 18-month conversation engagement, β=.224, p=.012, number of children present to 18 month challenging temperament, β=.286, p=.004, number of children present to 30-month conversation engagement, β=−.286, p<.001, and maternal verbosity to 36-month toddler language, β=−.282, p=.003.
Figure 2. Longitudinal Relations among Toddler Challenging Temperament, Parent-Toddler Conversation Engagement, and Toddler Language.
Note. Standardized coefficients are reported, with standard errors in parentheses. Only significant paths (p < .05) are displayed. Model controlled for parent education, number of children present, and maternal verbosity; for ease of interpretation, covariate paths are not displayed.
Concurrent Associations
Regarding concurrent relations, toddler challenging temperament was unrelated to toddler concurrent language ability at 18 months, β=−.093, p=.382. However, toddler challenging temperament was significant inversely associated with concurrent parent-toddler conversation at 18 months, β=−.204, p=.046, and at 30 months, β=−.397, p=.001. Finally, toddler language ability was significantly associated with concurrent parent-toddler conversation engagement at 18 months, β=.306, p<.001.
Longitudinal Associations
With the exception of the path from 18-month toddler language to 24-month toddler language, all autoregressive paths were significant, indicating stability in individual differences in toddler challenging temperament, parent-toddler conversation engagement, and later toddler language. After accounting for this stability, the only significant cross-lagged paths were those from 18-month conversation engagement to 24-month language, β=.419, p<.001, and from 30-month conversation engagement to 36-month language, β=.278, p=.004. That is, accounting for earlier toddler language ability, parent-toddler conversation engagement was a significant predictor of toddler’s later language abilities. Toddler challenging temperament at 18 months was not a significant predictor of either 24-month language abilities, β=−.055, p=.551, or 30-month conversation engagement, β=.155, p=.239. Toddler 24-month language abilities did not predict 30-month temperament, β=−.096, p=.209 or 30-month conversation engagement, β=−.052, p=.657. Moreover, 30-month challenging temperament was not a significant predictor of 36-month language, β=.064, p=.517.
Indirect Effects
Although there were no direct effects from challenging temperament to toddler language abilities, there were several indirect effects, which are summarized in Table 8a. First, there was a significant indirect effect of 18-month challenging temperament on 24-month language via 18-month conversation engagement, β=−.085, 95% CI [−.219, −.006]. Moreover, 18-month challenging temperament predicted 36-month language through three significant indirect paths: 1) from 18-month challenging temperament to 18-month conversation engagement to 24-month language to 36-month language, β=−.031, 95% CI [−.092, −.004], 2) from 18-month challenging temperament to 18-month conversation engagement to 30-month conversation engagement to 36-month language, β=−.018, 95% CI [−.056; −.001], and 3) from 18-month challenging temperament to 30-month challenging temperament to 30-month conversation engagement to 36-month language, β=−.078, 95% CI [−.176; −.021]. Finally, there was a significant indirect effect of 30-month challenging temperament to 36-month language ability via 30-month conversation engagement, β=−.110, 95% CI [−.241; −.028].
Table 8.
Summary of Indirect Effects examined in a) the Original Path Model and b) the Alternative Path Model Reversing the Path between Challenging Temperament and Conversation Engagement
| a) Hypothesized Paths | b) Reverse Paths | ||||
|---|---|---|---|---|---|
| Temperament →Conversation →Language | Est. | 95% CI | Conversation →Temperament →Language | Est. | 95% CI |
|
| |||||
| ChallTemp18→Conv18→Lang24 | −.085 | [−.219, −.006] | Conv18→ChallTemp18→Lang24 | .012 | [−.021, .079] |
| ChallTemp18→Conv18→Lang24→Lang36 | −.031 | [−.092, −.004] | Conv18→ChallTemp18→Lang24→Lang36 | .004 | [−.008, .028] |
| ChallTemp18→Conv18→Conv30→Lang36 | −.018 | [−.056, −.001] | Conv18→ChallTemp18→ChallTemp30→Lang36 | −.009 | [−.053, .016] |
| ChallTemp18→ChallTemp30→Conv30→Lang36 | −.078 | [−.176, −.021] | Conv18→Conv30→ChallTemp30→Lang36 | −.005 | [−.036, .008] |
| ChallTemp30→ Conv30→Lang36 | −.110 | [−.241, −.028] | Conv30→ ChallTemp30→Lang36 | −.016 | [−.081, .031] |
Note. Bolded estimates are statistically significant based on inspection of the 95% confidence intervals (CIs). CIs that do not contain zero are considered significant.
To address the alternative hypothesis that parent-toddler conversation engagement may influence language via effects on toddler temperament, we re-ran the path model reversing the paths from toddler temperament to conversation engagement, and then inspected the indirect effects from conversation engagement to toddler temperament to toddler language. The indirect paths inspected are summarized in Table 8b. Although the paths from conversation engagement to toddler challenging temperament remained significant at 18 months, β=−.214, p=.001, and 30 months, β=−.249, p=.001, there were no significant indirect paths from conversation engagement to language via temperament, as evidenced by all confidence intervals containing zeros.
Finally, in addition to the direct paths from parent-toddler conversation engagement to subsequent time-point toddler language abilities (i.e. from 18-month conversation engagement to 24-month language and from 30-month conversation engagement to 36-month language), there was a significant indirect path from 18-month parent-toddler conversation engagement to 36-month language via 24-month language, β=.154, 95% CI [.066, .278], and via 30-month parent-toddler conversation engagement, β=.085, 95% CI [.014, .192].
Summary of Results
The results of the path analysis did not support H1; there were no significant concurrent or longitudinal direct effects between toddler challenging temperament and toddler language. However, H2 was partially supported; more challenging temperament was associated with lower concurrent parent-toddler conversation engagement at both 18 months and 30 months. However, there were no longitudinal relations between temperament and conversation engagement, that is, 18-month challenging temperament was not significantly associated with 30-month conversation engagement. H3 was supported; parent-toddler conversation engagement was associated with later toddler language skills. Finally, H4 was partially supported. Given a) the lack of direct effects of toddler challenging temperament on language outcomes, and b) the lack of cross-lagged paths from toddler temperament to later conversation engagement, mediation was not supported. However, there were significant indirect effects of 18-month challenging temperament on 24-month and 36-month language via 18-month conversation engagement, as well as a significant indirect effect of 30-month challenging temperament on 36-month language via 30-month conversation engagement. The reverse paths (indirect effects of conversation engagement on language via temperament) were not supported.
Post-Hoc Analyses
To examine whether results differ based on the number of parents present during observation of conversation engagement, we re-ran the final path model excluding visits in which only mothers were present. This resulted in a sample of 109, which included cases for which both parents were present for at least one of the home observations. The resulting model fit the data well, χ2 (20) = 24.069, p = .2394, CFI = .978, RMSEA = 0.044, SRMR = .046. The relations remained largely the same, except that the path from 18-month challenging temperament to 18-month conversation engagement weakened and became nonsignificant, β=−.186, p=.100, as did the indirect effects from 18-month challenging temperament to 24-month and 36-month language via conversation engagement. However, the path from 30-month challenging temperament to 30-month conversation engagement remained significant, β=−.404, p=.002, as did the indirect effects from 30-month challenging temperament to 36-month language via 30-month conversation engagement, β=−.103, 95% CI [−.241, −.020].
In addition, although comparison of mother and father conversation engagement was not a focus of the present study, we did re-run the model using a) only mother conversation engagement data and b) only father conversation engagement data to explore whether the conversation engagement of a particular parent may be driving the current findings. For these two models, we included only observations in which both parents were present. Our model was a poorer fit to the mother-only data and to the father-only data compared to the combined parent data, as indicated by a mix of acceptable, marginal and inadequate model fit indices for mothers, χ2 (20) = 29.184, p = .0842, CFI = .947, TLI = .871; RMSEA = 0.066, SRMR = .048, and poor model fit for fathers based on all but one model fit index, χ2 (20) = 38.670, p = .0073, CFI = .888, TLI = .727; RMSEA = 0.095, SRMR = .064. Thus, we do not report on these separate mother-only and father-only models further.
Discussion
Parents’ engagement with their young children in reciprocal, contingent verbal communication is a critical process supporting early language development (e.g., Zauche et al., 2016). Most research examining sources of individual differences in parent-child conversation engagement has focused on demographic factors such as family income or parent education, despite wide variability in parent-child conversation engagement within socioeconomic groups (Rowe, 2018; Sperry et al., 2018). We addressed this gap by considering— within a sample restricted to include only families who are economically strained– whether toddlers’ temperamental characteristics reflecting emotional reactivity and self-regulation (negative affectivity and effortful control, respectively) account for individual differences in parent-toddler conversation engagement. Each of these temperamental traits has been linked to early language abilities (e.g., Kucker et al., 2021; Noel et al., 2008; Salley & Dixon, 2007), yet the mechanisms explaining these associations are largely unexplored. We considered these two inversely related traits together, as a composite reflecting challenging temperament, examining whether challenging temperament accounted for variability in conversation engagement and language outcomes and whether conversation engagement may mediate associations between challenging temperament and language skills. We found partial support for our hypotheses. Although mediation was not supported, we did find significant concurrent relations between challenging temperament and parent-toddler conversation engagement, and indirect effects of challenging temperament on later language skills via conversation engagement. These findings, their implications, and the study strengths and limitations are discussed in the following sections.
Parent-Toddler Conversation Engagement and Language Development
First, the results add to a literature demonstrating the importance of early parent-child conversation engagement for supporting language development. The frequency with which parents engaged their toddlers in semantically contingent verbal conversation predicted subsequent toddler expressive language skills, including expressive vocabulary and syntactic complexity, while controlling for earlier language abilities. These longitudinal relations were found between 18 and 24 months, and between 30 and 36 months. Notably, 18-month conversation engagement also accounted for variability in 36-month language skills via a unique indirect pathway, through 24-month language. In other words, 18-month conversation engagement contributed to variability 36-month language not only because it predicted stability in conversation engagement from 18 to 30 months, but also because it supported 24-month language skills, which contributed to 36-month skills independent of 30-month conversation engagement. These findings speak to the importance of early conversation engagement during the developmental period when expressive language skills are emerging. Moreover, they complement findings from a study showing that conversational turns between 18–24 months, but not other developmental windows, predict subsequent language abilities (Gilkerson et al., 2018). Thus, these early conversations in the second year of life appear to build a foundation of communication skills that support subsequent language development.
Challenging Temperament, Conversation, and Language
Existing research shows associations of toddler temperamental traits with both concurrent language skills and subsequent language development. Specifically, although there are some inconsistencies in the literature, generally, high toddler negative affectivity or low toddler effortful control are associated with poorer language skills (Eisenberg et al., 2010; Guedes & Cadima, 2022; Kucker et al., 2021; Noel et al., 2008; Salley & Dixon, 2007). There is a dearth of literature examining mechanisms that may explain these relations; however, associations between these temperamental traits and parent sensitivity and responsiveness suggest that one mechanism may be the impact of more challenging toddler temperamental characteristics on parent-child interactions. We tested the possibility that toddlers with more challenging temperament (i.e., toddlers who are higher in emotional reactivity and lower in self-regulation) may deter or disrupt the parent-toddler conversation processes that are known to support language development.
Supporting our hypotheses, parents of toddlers with more challenging temperament engaged with their toddlers in verbal conversation less frequently at both 18 and 30 months, even when accounting for toddler language abilities. These findings are consistent with the small number of previous studies linking high negative emotion expression or negative affectivity with lower parental language input or less frequent parent-toddler conversation engagement (Fields-Olivieri et al., 2020; Ontai & Virmani, 2010; Vernon-Feagans et al., 2008) as well as with a much larger literature linking high negative affectivity or low effortful control with lower parental sensitivity or responsiveness (Kiff et al., 2011; Paulussen-Hoogeboom et al., 2007). Notably, 24-month language ability did not account for variability in 30-month conversation engagement, which was instead predicted by 18-month conversation engagement and by concurrent challenging temperament. Although this may be explained partially by the 6-month gap between measurement of language and conversation engagement, it may also reflect advances in toddler language abilities. As toddler language skills mature, parent-toddler conversation engagement may be less constrained by language abilities by 30 months, and instead appears to be shaped by toddler temperament as well as by family history of conversation engagement. These findings reinforce the importance of early conversation in laying the groundwork for later conversation engagement. Moreover, they suggest that conversation between parents and their young children are driven not only by the toddler’s language skills but also by their temperamental traits.
Why might challenging temperament predict less conversation engagement between parents and their toddler? One possibility is that negative emotions themselves interfere with engagement. Children in the early stages of expressive language acquisition have difficulty speaking while experiencing negative emotions (Bloom & Beckwith, 1989). Therefore, toddlers who are high in emotional reactivity and low in self-regulation may make fewer attempts to communicate verbally or it may be more difficult for their parents to engage them in conversation. Previous work from our group suggests that, at 18 months, effects of temperament and emotion expressions on parent-toddler conversation engagement may be driven largely by effects on toddlers’ attempts to communicate (Fields-Olivieri et al., 2020). Yet in addition to effects on toddler communication, challenging toddler characteristics such as high emotional reactivity and low self-regulation contribute to parental fatigue and undermine responsive parenting, both when the child is distressed and when the child is not expressing negative emotions (Bailes & Leerkes, 2023; van den Boom & Hoeksma, 1994). Even in the absence of toddler negative expressions, parents of highly reactive and poorly regulated toddlers may be less likely to respond to their toddler’s verbal cues or may be hesitant to initiate conversation with their toddler, fearing that such engagement may evoke negative emotions. Indeed, evidence from naturalistic recordings suggests that negative emotions are often elicited by parent-toddler conversation (Fields-Olivieri & Cole, 2022). Evaluation of these hypotheses requires measurement of toddler temperament as well as direct observation and dynamic analysis of in-the-moment toddler emotion expressions and parent and toddler verbal behaviors, an issue we discuss further below.
Unexpectedly, toddler challenging temperament was not associated with concurrent language abilities and there were no direct effects of challenging temperament on subsequent language outcomes. Although challenging temperament at both 18 and 30 months was related to 24-month language at the bivariate level, these relations weakened once accounting for other factors such as previous language abilities and parent-toddler conversation engagement. Although these results are surprising, other studies fail to find relations of negative affectivity or effortful control with language abilities concurrently (Nozadi et al., 2013; Usai et al., 2009) or longitudinally (e.g., Ishikawa-Omori et al., 2022; Laake & Bridgett, 2018). Moreover, a few studies suggest that the effect of temperamental characteristics on language abilities may depend on parental sensitivity (Karrass & Braungart-Rieker, 2003; Laake & Bridgett, 2018). Although we tested a mediation model in the present study, the relations among language, temperament, and parenting are likely complex; toddler temperamental traits may evoke certain parenting behaviors and yet parenting behaviors may also moderate associations between toddler traits and developmental outcomes. Future work should consider investigating the complex interplay among these factors and consider both mediating and moderating processes, including the role that parents’ own emotional and self-regulatory characteristics may play.
Despite the null findings regarding direct effects of challenging temperament on language abilities, we did find evidence for indirect effects of challenging temperament on subsequent language skills via parent-toddler conversation engagement. That is, more challenging temperament was associated with lower concurrent parent-toddler conversation engagement, and lower conversation engagement predicted poorer language skills at subsequent ages. Thus, although we did not find strong support for direct relations between toddler temperament and their language abilities, our findings suggest that toddler challenging temperament plays an important role in the communication processes that support early language development.
It should be noted that, in post-hoc analyses including only observations in which both parents were present, the associations of toddler challenging temperament with parent-toddler conversation engagement weakened at 18 months, whereas the concurrent relations between challenging temperament and parent-toddler conversation engagement at 30 months remained. A few explanations are possible. First, a decrease in sample size may have decreased the power to detect significant effects. An additional possibility is that the inverse association between challenging temperament and parent-toddler conversation engagement at 18 months was driven by cases in which only a single parent was present. That is, at 18 months, these detrimental effects of challenging temperament on parent-toddler conversation engagement may be particularly strong when a caregiver is parenting alone. In contrast, at 30 months, the association between challenging temperament and parent-toddler conversation engagement was more robust, and remained significant even when excluding cases in which only one parent was present. Given increases across the toddler period in parental expectations for emotional self-regulation and verbal rather than emotional communication (Brownell & Kopp, 2007), by 30 months, parents, regardless of who is present, may more often disengage from toddlers they view to be highly emotionally reactive and low in self-regulation.
Finally, it is notable that parent education accounted only for variability in 18-month, but not 30-month, conversation engagement. Income was restricted in an effort to achieve an economically strained sample that did not benefit from the resources afforded to the middle class, but also did not experience the potentially confounding risk factors associated with poverty. Within this constrained sample, there was variability in parent education; however, constraining income in this way could have decreased the strength of the effects of SES factors. With this caveat in mind, the results may also speak to the importance of considering non-SES-related factors that account for variability in conversation engagement, and address calls to examine sources of individual differences within socioeconomic groups (Rowe, 2018; Sperry et al., 2018). Our findings suggest that as toddler language skills are emerging at 18 months, parent education (along with temperament) accounts for variability in the extent to which parents initiate and sustain conversation with their toddlers. However, as toddler language skills solidify, by 30 months, conversation engagement appears less driven by parent education and instead is shaped by the family history of conversation engagement and by toddler temperamental traits.
Strengths, Limitations and Future Directions
In sum, we find that toddler temperamental traits, namely the combination of emotional reactivity and regulation, contribute to individual differences in parent-toddler conversation. Thus, in addition to the role that socioeconomic factors play in understanding between-person differences in parent-toddler conversation engagement, we find that child factors also play a role.
A key strength is the use of observational measurement of semantically contingent conversation within naturalistic settings across a critical period of development. Whereas many studies consider only temporal contingency of parent-child conversational turns (e.g., Gilkerson et al., 2018; Zimmerman et al., 2009), we measured semantic and, to a lesser extent, temporal contingency of turns. Semantic contingency between toddler cues and parent responses is particularly important for language development because it provides toddlers with input related to their current focus of attention and provides opportunity for the toddlers to practice and receive feedback on their expressive skills (Casillas, 2014; Tamis-LeMonda et al., 2014; Warlaumont et al., 2014). Moreover, the ecological validity of our measurement of conversation engagement is another strength. Estimates of parent-child communication differ significantly when measured in structured settings compared to when sampled naturalistically in the home (Bergelson et al., 2019; Tamis-LeMonda et al., 2017), and effect sizes of parental linguistic input are larger when measured in the home versus structured settings (Anderson et al., 2021). Thus, our sample of parent-toddler conversation engagement likely more closely resembles toddlers’ everyday communication experiences than if measured in a lab setting.
One potential limitation to our approach to quantifying conversation engagement is that we did not impose a strict time limit on response latency when identifying connected turns or attempts to understand, other than the overall rule that all speaker turns were bounded by another speaker’s utterance or by 10 seconds of silence. Although there is no known systematic review of parental verbal response latencies, particularly in naturalistic settings, several studies of parental responsiveness (e.g., Pretzer et al., 2019; Van Egeren et al., 2001) and a recent review of infant response latency (Nguyen et al., 2022) suggests that most turns occur within much shorter time frames (i.e., within 1–2 seconds). Although our coders report that most parental responses followed a toddler turn within a few seconds, given the nature of the transcription and coding process (which was not timestamped), it is not possible to report descriptive statistics on average response latencies of parental connected turns or attempts to understand. Because we did not impose of strict latency rule to distinguish “connected” turns from other non-contingent codes, it is possible that our coding system may have overestimated contingent communication. Despite this potential limitation, parental conversation engagement predicted subsequent toddler language ability, suggesting that the coding system does capture an aspect of contingent communication that supports early language development. It is possible that with more stringent latency criteria, the results may have been further strengthened.
One drawback to our naturalistic observational approach is that, in 9% of cases at 18 months and 12% of cases at 30 months, fathers were not present for the observation. Although engaging in conversation with only one caregiver is a common experience for young children that may happen frequently throughout everyday life (Bergelson et al., 2019), for some participants, we lacked information about what conversation engagement may look like with both parents present. Given the naturalistic nature of the observations, we chose not to treat these data as missing. However, we recognize that there may be both within-parent and between-parent differences in communication during interactions with both parents versus only one parent present, including mother-father differences being accentuated in triadic interactions (Bingham et al. 2013). When interacting with their young children together, mothers and fathers appear to accommodate to each other’s contributions to communication, both speaking less than when parenting in dyadic contexts (Bingham et al., 2013; Jean & Elizabeth, 2022). Consistent with these findings, in our sample, toddlers received comparable levels of conversation engagement regardless of whether fathers were present, suggesting that mothers speak more when parenting alone but less when parenting together. Our naturalistic observations were not designed to systematically compare mothers’ and fathers’ parenting; therefore, we chose to take a “whole-family” approach, accounting for the presence of siblings and combining mother-toddler and father-toddler conversation engagement to examine whether toddlers’ challenging temperament relates to their experience of being engaged in contingent communication with any caregiver. Our post-hoc analyses support our decision to take this “whole-family” approach, given better model fit for combined parent conversation engagement data compared to separate mother-only and father-only models. Additional post-hoc analyses do suggest that at 18 months, challenging temperament may be especially impactful on conversation engagement when only mothers are present but less so when both parents are available to communicate. Future research might supplement naturalistic observations with more structured observations including different partner configurations to more systematically compare effects of challenging temperament on mother-toddler versus father-toddler versus triadic conversation engagement.
Another strength of the current study is the use of observational measures of toddler language abilities; we measured toddler language skills from spontaneous speech samples recorded during laboratory tasks at 18, 24, and 36 months. The metrics derived from these samples (MLU, NDW) are widely used measures of language ability and have comparable psychometric properties to standardized measures of language (DeThorne et al., 2011). We used these measures in part because they allowed us to utilize identical measures of language at each time point. One limitation of this approach is that toddlers were interacting with their mothers during the laboratory tasks. Thus, our measure of language is somewhat confounded with our measure of conversation engagement, given that mothers’ communicative behavior in the laboratory tasks likely influenced toddler speech. However, controlling for maternal verbosity addressed concerns that our results might be driven by effects of maternal communicative behavior on language estimates. Moreover, MLU appears unaffected by conversational partner, and sampling speech in the presence of a caregiver may most accurately capture the true extent of young children’s expressive abilities (Bornstein et al., 2000). Future studies might supplement language measures derived from speech samples in the presence of a parent with speech samples in the presence of other adults or with standardized measures of language.
Toddler temperament was assessed via a widely used parent-report measured of temperament. Although parent-report measures may be subject to biases compared to direct observation of emotional reactivity or self-regulation in the lab, parents’ responses to questionnaire-based temperament measures reflect patterns of reactivity and regulation across time and contexts, whereas lab-based measures provide a brief snapshot of in-the-moment reactivity or regulation within a specific task (Rothbart & Bates, 2006). Moreover, parents’ perceptions of toddler temperament matter: parents with more positive perceptions of their children report communicating more with their children (Aring & Renk, 2010), and parent ratings of temperament tend to relate more strongly to parenting behaviors than do laboratory assessments of temperament (Paulussen-Hoogeboom et al., 2007). Our findings suggest that parents who perceive their toddler to be more challenging, that is, higher in negative emotional reactivity and lower in the ability to self-regulate, are less likely to initiate and sustain conversation with their toddlers. Whether these relations would replicate with observational measures of toddler temperament remains to be explored.
A related issue is that we did not measure the in-the-moment effects of toddlers’ emotion expressions or regulatory behaviors on parent-toddler conversation engagement. Parent report of toddler temperament often relates weakly to observational measures of emotion expressions (e.g., Kubicek & Emde, 2012; Tan et al., 2013), likely because multiple factors other than temperament contribute to emotional expressions at a given moment (Strelau, 2001). Given our desire to examine longitudinal relations among temperament, conversation, and language, we chose to focus on a parent-report measure of temperament that reflects toddler emotional reactivity and self-regulation across situations. Yet because of this, we could not assess whether, for example, toddlers’ in-the-moment negative emotion expressions discouraged or disrupted parent-toddler conversation engagement, or whether parental perceptions of the toddler’s temperament shaped their tendency to engage in conversation with their toddler, regardless of their toddlers’ in-the-moment emotion expressions. Relatedly, given the immature language skills of the toddlers in our sample, our measure of conversation engagement focuses on parent contributions (although implies toddler contributions). Thus, we were unable to disentangle whether effects on engagement were driven by parent or child communicative behaviors, or on the tendency of partners to initiate versus respond to conversational turns. Future research taking a dynamic approach that examines sequential patterns of toddler emotion and parent and toddler verbal communicative behaviors will help elucidate the complex processes driving effects of temperament on parent-toddler conversation engagement.
Importantly, the effect of challenging temperament on parent-toddler conversation engagement is just one possible mechanism explaining associations between temperament and language skills. For example, effortful control involves modulation of not only emotions but also attention (Rothbart & Bates, 2006); similarly, the experience and expression of negative emotion may interfere with attentional processes (Salley & Dixon, 2007). Perhaps challenging temperament shapes language development not only through effects on parent-toddler conversation engagement, but also by interfering with attention towards linguistic input within these interactions. Future work should test this hypothesis to further disentangle effects of temperament on language and to elucidate unique effects of conversation engagement.
Finally, our approach of considering individual differences in parent-toddler conversation engagement within a single socioeconomic group is a strength, enriching a literature that has historically focused on between-group differences. Yet whether these findings would replicate in other groups is an open question. There is some evidence that the associations of negative affectivity with parenting quality may be stronger in families from lower socioeconomic backgrounds (e.g., Jenkins et al., 2003; Paulussen-Hoogeboom et al., 2007). It is possible that the effects of toddler challenging temperament on conversation engagement may be especially strong in families exposed to stressors associated with poverty, whereas these relations may weaken for families who have access to more resources. Exploring these processes across culturally and socioeconomically diverse samples is an important future direction.
Conclusions
Our findings address calls to examine non-sociodemographic factors that account for individual differences in the frequency with which parents verbally communicate with their young children. We found that toddler challenging temperament (reflecting high negative affectivity and low effortful control) accounts for variability in the extent to which parents engage their toddlers in reciprocal, semantically contingent conversation, such that toddlers with more challenging temperament experience less frequent conversation engagement with their parents. Moreover, despite finding no direct associations of toddler challenging temperament with toddler language skills, we found indirect effects of challenging temperament on subsequent language skills via parent-toddler conversation engagement. The findings contribute to the existing literature linking toddler temperament with language skills, which thus far lacks exploration of possible mechanisms explaining these associations. Future directions in this line of work include dynamic examination of toddler and parent emotional and verbal communicative behaviors, to further illuminate the complex processes by which toddler temperamental traits may shape parent-toddler conversation and in turn toddlers’ developing communication skills.
Supplementary Material
Highlights.
Toddler temperament accounts for variability in parent-toddler conversation
High negative affectivity/low effortful control predicts less conversation
Indirect effect of toddler temperament on toddler language via conversation
Acknowledgments
This research is supported by the National Institute of Mental Health (RO1 MH61388). It was approved by the Penn State Office of Research Compliance, IRB# 990642.
Footnotes
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
CRediT authorship contribution statement
Fields-Olivieri: Conceptualization; Formal Analysis; Writing- Original Draft; Thinzar: Formal Analysis; Writing- Original Draft; Writing- Review and Editing; Cole: Conceptualization; Investigation; Resources; Funding Acquisition; Supervision; Writing- Review & Editing Roben: Project Administration; Writing- Review and Editing
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Data Availability Statement
Deidentified numerical data and manuals are available upon request from the lead author. The consent forms, which were signed by participants prior to current open science guidelines, do not provide for open availability of all data.
References
- Anderson NJ, Graham SA, Prime H, Jenkins JM, & Madigan S (2021). Linking quality and quantity of parental linguistic input to child language skills: A meta-analysis. Child Development, 92(2), 484–501. 10.1111/cdev.13508 [DOI] [PubMed] [Google Scholar]
- Bailes LG, & Leerkes EM (2023). Transactional associations between infant negative emotionality and maternal sensitivity: Maternal emotion dysregulation as a moderator. Journal of Family Psychology, 37(3), 369–379. 10.1037/fam0001060 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Barry RA, & Kochanska G (2010). A longitudinal investigation of the affective environment in families with young children: From infancy to early school age. Emotion, 10(2), 237–249. 10.1037/a0018485 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Beiting M, Alper RM, Luo R, & Hirsh Pasek Kathy. (2022). Keep the ball rolling: Sustained multiturn conversational episodes are associated with child language ability. American Journal of Speech-Language Pathology. 10.1044/2022_AJSLP-21-00333 [DOI] [PubMed] [Google Scholar]
- Belsky J (1984). The determinants of parenting: A process model. Child Development, 55(1), 83–96. 10.2307/1129836 [DOI] [PubMed] [Google Scholar]
- Bergelson E, Amatuni A, Dailey S, Koorathota S, & Tor S (2019). Day by day, hour by hour: Naturalistic language input to infants. Developmental Science, 22(1), e12715. 10.1111/desc.12715 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bingham GE, Kwon K-A, & Jeon H-J (2013). Examining relations among mothers’, fathers’, and children’s language use in a dyadic and triadic context. Early Child Development and Care, 183(3–4), 394–414. 10.1080/03004430.2012.711590 [DOI] [Google Scholar]
- Bloom L, & Beckwith R (1989). Talking with feeling: Integrating affective and linguistic expression in early language development. Cognition & Emotion, 3(4), 313–342. [Google Scholar]
- Brown JR, Donelan-McCall N, & Dunn J (1996). Why talk about mental states? The significance of children’s conversations with friends, siblings, and mothers. Child Development, 67(3), 836–849. [PubMed] [Google Scholar]
- Brown KM, Ram N, & Lunkenheimer E (2022). The influence of children’s effortful control on parent–child behavioral synchrony. Journal of Family Psychology. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Brownell CA, & Kopp CB (2010). Socioemotional development in the toddler years: Transitions and transformations. Guilford Press. [Google Scholar]
- Camaioni L (2001). Early language. Blackwell Handbook of Infant Development, 404–426. [Google Scholar]
- Casillas M (2014). Turn-taking. Pragmatic Development in First Language Acquisition, 53–70. [Google Scholar]
- Cioffi CC, Griffin AM, Natsuaki MN, Shaw DS, Reiss D, Ganiban JM, Neiderhiser JM, & Leve LD (2021). The role of negative emotionality in the development of child executive function and language abilities from toddlerhood to first grade: An adoption study. Developmental Psychology, 57(3), 347–360. 10.1037/dev0000972 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Conture EG, Kelly EM, & Walden TA (2013). Temperament, speech and language: An overview. Journal of Communication Disorders, 46(2), 125–142. 10.1016/j.jcomdis.2012.11.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- DeThorne LS, Deater-Deckard K, Mahurin-Smith J, Coletto M-K, & Petrill SA (2011). Volubility as a mediator in the associations between conversational language measures and child temperament. International Journal of Language & Communication Disorders / Royal College of Speech & Language Therapists, 46(6), 700–713. 10.1111/j.1460-6984.2011.00034.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- Donnelly S, & Kidd E (2021). The longitudinal relationship between conversational turn-taking and vocabulary growth in early language development. Child Development, 92(2), 609–625. 10.1111/cdev.13511 [DOI] [PubMed] [Google Scholar]
- Eisenberg N, Vidmar M, Spinrad TL, Eggum ND, Edwards A, Gaertner B, & Kupfer A (2010). Mothers’ teaching strategies and children’s effortful control: A longitudinal study. Developmental Psychology, 46(5), 1294–1308. 10.1037/a0020236 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ensor R, & Hughes C (2008). Content or Connectedness? Mother-child talk and early social understanding. Child Development, 79(1), 201–216. [DOI] [PubMed] [Google Scholar]
- Fields-Olivieri MA, & Cole PM (2022). Toddler negative emotion expression and parent-toddler verbal conversation: Evidence from daylong recordings. Infant Behavior and Development, 67, 101711. 10.1016/j.infbeh.2022.101711 [DOI] [PubMed] [Google Scholar]
- Fields-Olivieri MA, Cole PM, & Roben CKP (2020). Toddler emotion expressions and emotional traits: Associations with parent-toddler verbal conversation. Infant Behavior and Development, 14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Garello V, Viterbori P, & Usai MC (2012). Temperamental profiles and language development: A replication and an extension. Infant Behav Dev, 35(1), 71–82. 10.1016/j.infbeh.2011.09.003 [DOI] [PubMed] [Google Scholar]
- Gilkerson J, Richards JA, Warren SF, Oller DK, Russo R, & Vohr B (2018). Language experience in the second year of life and language outcomes in late childhood. Pediatrics, 142(4), e20174276. 10.1542/peds.2017-4276 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Goldsmith HH (1996). Studying temperament via construction of the Toddler Behavior Assessment Questionnaire. Child Development, 67(1), 218–235. [PubMed] [Google Scholar]
- Guedes C, & Cadima J (2022). The bidirectional interplay between self-regulation and expressive vocabulary during toddlerhood. Developmental Psychology, 58(9), 1652–1664. 10.1037/dev0001062 [DOI] [PubMed] [Google Scholar]
- Hart B, & Risley TR (1995). Meaningful differences in the everyday experience of young American children. Paul H Brookes Publishing. http://psycnet.apa.org/psycinfo/1995-98021-000 [Google Scholar]
- Hayes AF, Preacher KJ, & Myers TA (2011). Mediation and the estimation of indirect effects in political communication research. Sourcebook for Political Communication Research: Methods, Measures, and Analytical Techniques, 23(1), 434–465. [Google Scholar]
- Hirsh-Pasek K, Adamson LB, Bakeman R, Owen MT, Golinkoff RM, Pace A, Yust PKS, & Suma K (2015). The contribution of early communication quality to low-income children’s language success. Psychological Science, 26(7), 1071–1083. PsycINFO. 10.1177/0956797615581493 [DOI] [PubMed] [Google Scholar]
- Hoff E (2006). How social contexts support and shape language development☆. Developmental Review, 26(1), 55–88. 10.1016/j.dr.2005.11.002 [DOI] [Google Scholar]
- Ishikawa-Omori Y, Nishimura T, Nakagawa A, Okumura A, Harada T, Nakayasu C, Iwabuchi T, Amma Y, Suzuki H, Rahman MS, Nakahara R, Takahashi N, Nomura Y, & Tsuchiya KJ (2022). Early temperament as a predictor of language skills at 40 months. BMC Pediatrics, 22(1), 56. 10.1186/s12887-022-03116-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jean Q, & Elizabeth N (2022). Parent child directed speech in dyadic and triadic interaction: Associations with co-parenting dynamics and child language outcomes. Early Childhood Research Quarterly, 58, 125–135. 10.1016/j.ecresq.2021.09.005 [DOI] [Google Scholar]
- Jenkins JM, Rasbash J, & O’Connor TG (2003). The role of the shared family context in differential parenting. Developmental Psychology, 39(1), 99–113. 10.1037/0012-1649.39.1.99 [DOI] [PubMed] [Google Scholar]
- Kang D, Yim D, Kang D, & Yim D (2022). The Influence of Children’s Temperament, Phonological Working Memory, Mother Related Factors, and Language Environment on Vocabulary Development in Korean Monolingual and English-Korean Bilingual Children. Communication Sciences & Disorders, 27(2), 281–296. 10.12963/csd.22886 [DOI] [Google Scholar]
- Karrass J, & Braungart-Rieker JM (2003). Parenting and temperament as interacting agents in early language development. Parenting, 3(3), 235–259. 10.1207/S15327922PAR0303_03 [DOI] [Google Scholar]
- Kiff CJ, Lengua LJ, & Zalewski M (2011). Nature and nurturing: Parenting in the context of child temperament. Clinical Child and Family Psychology Review, 14(3), 251–301. ProQuest Social Sciences Premium Collection. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kochanska G, Murray KT, & Harlan ET (2000). Effortful control in early childhood: Continuity and change, antecedents, and implications for social development. Developmental Psychology, 36(2), 220–232. 10.1037/0012-1649.36.2.220 [DOI] [PubMed] [Google Scholar]
- Kubicek LF, & Emde RN (2012a). Emotional expression and language: A longitudinal study of typically developing earlier and later talkers from 15 to 30 Months. Infant Mental Health Journal, 33(6), 553–584. 10.1002/imhj.21364 [DOI] [PubMed] [Google Scholar]
- Kucker SC, Zimmerman C, & Chmielewski M (2021). Taking parent personality and child temperament into account in child language development. British Journal of Developmental Psychology, n/a(n/a). 10.1111/bjdp.12379 [DOI] [PubMed] [Google Scholar]
- Laake LM, & Bridgett DJ (2018). Early language development in context: Interactions between infant temperament and parenting characteristics. Early Education and Development, 29(5), 730–746. 10.1080/10409289.2018.1436366 [DOI] [Google Scholar]
- Lewis MD, Zimmerman S, Hollenstein T, & Lamey AV (2004). Reorganization in coping behavior at 1½ years: Dynamic systems and normative change. Developmental Science, 7(1), 56–73. PsycINFO. 10.1111/j.1467-7687.2004.00323.x [DOI] [PubMed] [Google Scholar]
- Liew J, Eisenberg N, & Reiser M (2004). Preschoolers’ effortful control and negative emotionality, immediate reactions to disappointment, and quality of social functioning. Journal of Experimental Child Psychology, 89(4), 298–319. 10.1016/j.jecp.2004.06.004 [DOI] [PubMed] [Google Scholar]
- Moran LR, Lengua LJ, & Zalewski M (2013). The interaction between negative emotionality and Eeffortful control in early social-emotional development. Social Development (Oxford, England), 22(2), 340–362. 10.1111/sode.12025 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Moreno AJ, & Robinson JL (2005). Emotional vitality in infancy as a predictor of cognitive and language abilities in toddlerhood. Infant and Child Development, 14(4), 383–402. 10.1002/icd.406 [DOI] [Google Scholar]
- Muris P, Meesters C, & Blijlevens P (2007). Self-reported reactive and regulative temperament in early adolescence: Relations to internalizing and externalizing problem behavior and “Big Three” personality factors. Journal of Adolescence, 30(6), 1035–1049. 10.1016/j.adolescence.2007.03.003 [DOI] [PubMed] [Google Scholar]
- Muris P, & Ollendick TH (2005). The role of temperament in the etiology of child Psychopathology. Clinical Child and Family Psychology Review, 8(4), 271–289. 10.1007/s10567-005-8809-y [DOI] [PubMed] [Google Scholar]
- Nguyen V, Versyp O, Cox C, & Fusaroli R (2022). A systematic review and Bayesian meta-analysis of the development of turn taking in adult–child vocal interactions. Child Development, 93(4), 1181–1200. 10.1111/cdev.13754 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Noel M, Peterson C, & Jesso B (2008). The relationship of parenting stress and child temperament to language development among economically disadvantaged preschoolers. J Child Lang, 35(4), 823–843. 10.1017/S0305000908008805 [DOI] [PubMed] [Google Scholar]
- Nozadi SS, Spinrad TL, Eisenberg N, Bolnick R, Eggum-Wilkens ND, Smith CL, Gaertner B, Kupfer A, & Sallquist J (2013). Prediction of toddlers’ expressive language from maternal sensitivity and toddlers’ anger expressions: A developmental perspective. Infant Behavior and Development, 36(4), 650–661. 10.1016/j.infbeh.2013.06.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ollas D, Rautakoski P, Nolvi S, Karlsson H, & Karlsson L (2020). Temperament is associated with the use of communicative gestures in infancy. Infant and Child Development, 29(3), e2166. 10.1002/icd.2166 [DOI] [Google Scholar]
- Ontai LL, & Virmani EA (2010). Predicting elements of early maternal elaborative discourse from 12 to 18 months of age. Early Childhood Research Quarterly, 25(1), 98–111. 10.1016/j.ecresq.2009.08.001 [DOI] [Google Scholar]
- Paulussen-Hoogeboom MC, Stams GJJM, Hermanns JMA, & Peetsma TTD (2007). Child negative emotionality and parenting from infancy to preschool: A meta-analytic review. Developmental Psychology, 43(2), 438–453. 10.1037/0012-1649.43.2.438 [DOI] [PubMed] [Google Scholar]
- Pérez-Pereira M, Fernández P, Resches M, & Gómez-Taibo ML (2016). Does temperament influence language development? Evidence from preterm and full-term children. Infant Behavior and Development, 42, 11–21. 10.1016/j.infbeh.2015.10.003 [DOI] [PubMed] [Google Scholar]
- Pezold MJ, Imgrund CM, & Storkel HL (2020). Using computer programs for language sample analysis. Language, Speech, and Hearing Services in Schools, 51(1), 103–114. 10.1044/2019_LSHSS-18-0148 [DOI] [PubMed] [Google Scholar]
- Posner MI, & Rothbart MK (2000). Developing mechanisms of self-regulation. Development and Psychopathology, 12(3), 427–441. 10.1017/S0954579400003096 [DOI] [PubMed] [Google Scholar]
- Pretzer GM, Lopez LD, Walle EA, & Warlaumont AS (2019). Infant-adult vocal interaction dynamics depend on infant vocal type, child-directedness of adult speech, and timeframe. Infant Behavior and Development, 57, 101325. 10.1016/j.infbeh.2019.04.007 [DOI] [PubMed] [Google Scholar]
- Rothbart MK (1996). Children’s behavior questionnaire. University of Oregon. [Google Scholar]
- Rothbart MK, & Bates JE (2006). Temperament. In Handbook of child psychology: Social, emotional, and personality development, Vol. 3, 6th ed (pp. 99–166). John Wiley & Sons, Inc. [Google Scholar]
- Rothbart MK, & Rueda MR (2005). The development of effortful control. In Mayr U, Awh E, & Keele S (Eds.), Developing individuality in the human brain: A tribute to Michael I. Posner (pp. 167–188). American Psychological Association. [Google Scholar]
- Rowe ML (2018). Understanding socioeconomic differences in parents’ speech to children. Child Development Perspectives, 12(2), 122–127. 10.1111/cdep.12271 [DOI] [Google Scholar]
- Salley BJ, & Dixon WE (2007). Temperamental and joint attentional predictors of language development. Merrill-Palmer Quarterly (Wayne State University. Press), 53(1), 131. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sohr-Preston SL, Scaramella LV, Martin MJ, Neppl TK, Ontai L, & Conger R (2013). Parental socioeconomic status, communication, and children’s vocabulary development: A third-generation test of the family investment model. Child Development, 84(3), 1046–1062. 10.1111/cdev.12023 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sperry DE, Sperry LL, & Miller PJ (2018). Reexamining the verbal environments of children from different socioeconomic backgrounds. Child Development. 10.1111/cdev.13072 [DOI] [PubMed] [Google Scholar]
- Spinelli M, Fasolo M, Shah PE, Genovese G, & Aureli T (2018). The influence of early temperament on language development: The moderating role of maternal input. Frontiers in Psychology, 9. 10.3389/fpsyg.2018.01527 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Strelau J (2001). The concept and status of trait in research on temperament. European Journal of Personality, 15(4), 311–325. [Google Scholar]
- Tamis-LeMonda CS, Kuchirko Y, Luo R, Escobar K, & Bornstein MH (2017). Power in methods: Language to infants in structured and naturalistic contexts. Developmental Science, 20(6), e12456. 10.1111/desc.12456 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tamis-LeMonda CS, Kuchirko Y, & Song L (2014). Why is infant language learning facilitated by parental responsiveness? Current Directions in Psychological Science, 23(2), 121–126. [Google Scholar]
- Tan PZ, Armstrong LM, & Cole PM (2013). Relations between Temperament and Anger Regulation over Early Childhood. Social Development, 22(4), 755–772. 10.1111/j.1467-9507.2012.00674.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- Taraban L, & Shaw DS (2018). Parenting in context: Revisiting Belsky’s classic process of parenting model in early childhood. Developmental Review, 48, 55–81. 10.1016/j.dr.2018.03.006 [DOI] [Google Scholar]
- Usai MC, Garello V, & Viterbori P (2009). Temperamental profiles and linguistic development: Differences in the quality of linguistic production in relation to temperament in children of 28 months. Infant Behavior and Development, 32(3), 322–330. 10.1016/j.infbeh.2009.04.001 [DOI] [PubMed] [Google Scholar]
- Usai MC, Viterbori P, Gandolfi E, & Zanobini M (2020). The relationship between temperamental dimensions and inhibitory control in early childhood: Implications for language acquisition. Infant Behavior and Development, 61, 101495. [DOI] [PubMed] [Google Scholar]
- Van Beveren M-L, Mezulis A, Wante L, & Braet C (2019). Joint contributions of negative emotionality, positive emotionality, and effortful control on depressive symptoms in youth. Journal of Clinical Child & Adolescent Psychology, 48(1), 131–142. 10.1080/15374416.2016.1233499 [DOI] [PubMed] [Google Scholar]
- van den Boom DC, & Hoeksma JB (1994). The effect of infant irritability on mother-infant interaction: A growth-curve analysis. Developmental Psychology, 30(4), 581. [Google Scholar]
- Van Egeren LA, Barratt MS, & Roach MA (2001). Mother-infant responsiveness: Timing, mutual regulation, and interactional context. Developmental Psychology, 37(5), 684–697. 10.1037//0012-1649.37.5.684 [DOI] [PubMed] [Google Scholar]
- Vernon-Feagans L, Pancsofar N, Willoughby M, Odom E, Quade A, & Cox M (2008). Predictors of maternal language to infants during a picture book task in the home: Family SES, child characteristics and the parenting environment. Journal of Applied Developmental Psychology, 29(3), 213–226. 10.1016/j.appdev.2008.02.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Vroman LN, & Durbin CE (2015). High effortful control is associated with reduced emotional expressiveness in young children. Journal of Research in Personality, 58, 46–54. 10.1016/j.jrp.2015.06.006 [DOI] [Google Scholar]
- Warlaumont AS, Richards JA, Gilkerson J, & Oller DK (2014). A social feedback loop for speech development and its reduction in autism. Psychological Science, 0956797614531023. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zauche LH, Thul TA, Mahoney AED, & Stapel-Wax JL (2016). Influence of language nutrition on children?s language and cognitive development: An integrated review. Early Childhood Research Quarterly, 36, 318–333. 10.1016/j.ecresq.2016.01.015 [DOI] [Google Scholar]
- Zimmerman FJ, Gilkerson J, Richards JA, Christakis DA, Xu D, Gray S, & Yapanel U (2009). Teaching by Listening: The Importance of Adult-Child Conversations to Language Development. PEDIATRICS, 124(1), 342–349. 10.1542/peds.2008-2267 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Deidentified numerical data and manuals are available upon request from the lead author. The consent forms, which were signed by participants prior to current open science guidelines, do not provide for open availability of all data.