Abstract
Given the dependent nature of parent–infant interactions necessary for language development, it is important to understand how context may influence these interactions. This study examines how contextual variables influence communicative, cognitive and social measures of parent–infant interactions. Specifically, how do feedback toys and traditional toys associate with important parent and child measures necessary for learning and development? Here we report evidence that toy type is associated with quality and quantity of parent–child interactions. Condition 1 examined parent–child interactions during two separate play sessions; one contained feedback toys and the other contained traditional toys. Condition 2 combined the toys so dyads had access to both types at the same time. Infants produced higher levels of directed vocalizations and directed gestures, but had shorter durations of sustained attention, when interacting with traditional toys compared to feedback toys. Parents responded more to infants’ vocalizations and gestures when interacting with traditional toys. In general, these results suggest that toy properties can play a significant role in parent–infant interactions.
The infants’ environment plays an important role in language development. There are complex relations among social and contextual factors that, together, impact language development. For example, many studies have demonstrated the importance of the quality and quantity of parents’ communication for infants’ language learning (e.g., Baumwell, Tamis-LeMonda, & Bornstein, 1997; Bornstein, 1989; Hart & Risley, 1995; Hirsh-Pasek et al., 2015; Tamis-LeMonda, Bornstein, & Baumwell, 2001). Similarly, studies have shown that certain interactive contexts, such as book reading, provide optimal input for infants’ language development (Wasik & Bond, 2001; Whitehurst et al., 1988). Therefore, it is important to understand how context influences parent and infant behavior to increase our understanding of the variables important for language development (Gros-Louis, West, & King, 2016). Given that learning opportunities in the first year are most abundant during play (Bruner, Jolly, & Sylva, 1976) and that play is linked to cognitive, social, language, and emotional development (Fisher, 1992; Ginsburg, 2007), the present study examined the effects of different toy types on infant behavior and parent responsiveness in semi-naturalistic interactions.
Studies have often compared behavior across contexts during typical parent–infant interactions that occur throughout the day, such as book reading, free play, caretaking interactions (e.g. feeding, bathing, dressing), and face-to-face interactions. During book reading contexts, parent speech is structurally more complex, contains more questions, fewer directives, and has more topic-continuing replies compared to other contexts such as caretaking interactions (Hoff-Ginsberg, 1991; Lewis & Gregory, 1987). This is in comparison to toy play where caregivers produce more directives (Hoff-Ginsberg, 1991; Lewis & Gregory, 1987) and fewer utterances that elicit conversation with the infant (Hoff-Ginsberg, 1991). In contrast, face-to-face interactions are often characterized by fewer directives and a higher frequency of repetitive language (Lewis & Gregory, 1987). Within toy play, certain categories of toys elicit differential caregiver verbalizations. For example, when playing with feminine toys compared to masculine toys, parents produce more utterances with praise, teaching, and questions; whereas, when interacting with masculine toys, parents produce more comments overall, more negative comments, and more animated sounds compared to when playing with neutral toys (Caldera, Huston, & O’Brien, 1989). During book reading contexts, mothers produce naming responses more frequently while questions are used more frequently during puppet contexts (Gros-Louis et al., 2016). Thus, it appears that across (e.g., book, face-to-face interaction) and within (e.g., book, puppet) various contexts, parent speech is quantitatively and qualitatively different (e.g., complexity, type of responses).
Similar to the relation between context and parent speech differences, there is also some evidence to suggest that context affects infant and toddler communication. Ten-month-old infants produce more vocalizations during contexts without toys (e.g., face-to-face interactions with caregiver) compared to toy or book contexts (Lewis & Gregory, 1987), and within book contexts, infants produce more consonant-vowels than vowel-like sounds (Gros-Louis et al., 2016). Hoff (2010) examined children during the latter part of their second year during interactions with their caregiver in three contexts: mealtime, toy play, and book reading. Children produced a larger repertoire of words and topic-continuing replies during the book context compared to during mealtime and toy play. Together, these studies suggest that interactive context affects the quality and quantity of parent and infant behavior.
However, what is less clear is how within a certain context, differences among certain types of items may impact parent–child interactions. For example, during play contexts, a wide variety of books and toys are often available, ranging from picture books with no words to books with simple sentences, and traditional toys to electronic toys. Recently, there has been a call for additional empirical studies to understand the impact of toys on child outcomes (Radesky & Christakis, 2016), and this is particularly important given the recent surge in electronic toys that manufacturing companies deem to be interactive and supportive of infant learning (Levin & Rosenquest, 2001). 1 Sosa (2016) found that when parents and infants engaged in play with electronic toys, parents responded less, produced fewer words and produced fewer content-specific words than during play with traditional toys or books. Likewise, infants produced fewer vocalizations during play with electronic toys compared to play with books. However, this study did not consider additional infant communicative and cognitive behavior such as gestures (e.g., point, ritualized reaches, palm point) and attention, two essential behaviors that relate to language development (Iverson & Goldin-Meadow, 2005), and that also play a role in parent–infant interactions (Miller & Gros-Louis, 2013). Also missing from recent studies is how infants use their vocalizations and gestures while engaged in various contexts. Given that recent work has demonstrated the way in which infants use their vocalizations during play interactions impacts both learning in the moment and language development, it is important to understand the way in which infants use their communicative behavior when engaged with social partners (Goldstein, Schwade, Briesch, & Syal, 2010; Gros-Louis, West, & King, 2014).
The purpose of the present study was to fill gaps in our understanding of how play with different toy types affects infant communicative and cognitive behavior (i.e., vocalizations, gestures, attention), and parent responses, that are known to be critical to language development. We focused on different toy types, and selected toys based on properties that represent typical infant toys available in stores such as feedback toys (e.g., toys with sound, music, and lights) and traditional toys (e.g., nesting cups) to examine how toy type affects moment-to-moment behaviors in parent–infant interactions that facilitate language development. To extend our knowledge from recent findings by Sosa (2016) and Gros-Louis et al. (2016), we focus on infant behavior directed towards contextual variables (i.e., toys, parent) that are crucial for learning (e.g., Goldstein et al., 2010). Specifically, we asked how contextual variables during parent–infant interaction affect the: (1) frequency of infants’ gestures and vocalizations; (2) frequency and duration of infant attention; (3) frequency of parent contingent feedback to infant vocalizations and gestures; and (4) global levels of parent sensitivity. Two conditions were conducted: Condition 1 examined parent–infant interactions when playing with the toy sets separately (e.g., one play session with feedback toys and one play session with traditional toys), while Condition 2 examined parent–child interactions when playing with a combination of feedback and traditional toys. This was done to examine if toy type independently influenced variables related to parent–child interactions and if differences did exist in Condition 1, if those same results would carry over, regardless of other toy types when combined in Condition 2.
Based on findings from Sosa (2016), we expect feedback toys in the current study to elicit similar behavior as electronic toys. If this is the case, infants will produce fewer vocalizations when interacting with feedback compared to traditional toys; furthermore, since vocalizations and gestures are mechanistically linked (e.g., Miller & Lossia, 2013), infants will also produce fewer gestures when interacting with electronic toys. In addition, a recent commentary by Radesky and Christakis (2016) suggests that toys that make noises or have some other features like lights are effective at commanding attention. If this is the case, children interacting with feedback toys may have more frequent and longer periods of sustained attention than when interacting with traditional toys. Finally, measures of parent responsiveness will result in fewer contingent responses to infant communicative behavior when interacting with electronic toys (Sosa, 2016). We did not have a priori hypotheses about parent sensitive behavior.
Methods
Participants
Fifty 9- to 12-month-old infants (21 females, 29 males; M = 331.96 days, SD = 44.5 days, range = 244–392 days) participated with their parent (45 mother–infant dyads, five father–infant dyads). We first assigned infants to Condition 1 (n = 34; 17 females, 17 males; M = 319 days, SD = 49 days, range = 244–390 days) and a separate set of infants were assigned to Condition 2 (4 females, 12 males; M = 358 days, SD = 12.13 days, range = 348–392 days). Participants were selected from a database that contained birth records published from a suburban area in the Midwestern United States. Data from an additional 11 infants were excluded from the analyses due to non-English speaking parents, equipment failure, or infant crying. In some instances, infants were excluded from Condition 1 while running Condition 2 and additional infants were recruited to participate in Condition 1. Parents reported their race/ethnicity as White (n = 42), Chinese (n = 1), Japanese (n = 1), African American (n = 1), Filipino (n = 1), White and African American (n = 1), White and Korean (n = 1), and Puerto Rican and Mexican (n = 1). All of the participating parents had at least a high-school diploma, with 46 of the parents holding at least one university degree. Demographic information for one parent–infant dyad was not complete. All infants were born full term with no known hearing problems. Infants received a book or toy as compensation for their participation.
Apparatus
Parents and their infants were tested in a semi-naturalistic environment to allow for video and audio recordings. All sessions were conducted in a child-proof playroom (a 3.9 m × 4.6 m). Behavioral interactions were recorded using one or two of four remote-controlled wall-mounted cameras (Sony HDR-HC7, via a Videonics MX-1 NTSC digital video mixer), and the video feed was recorded by an iMAC (2.16 GHz Intel Core 2 Duo) using Final Cut Pro (version 5.1.4). High-quality recordings were obtained by a small wireless microphone and transmitter (Telex Communications FMR-150), sewn into a pair of overalls worn by the infant.
Procedure
We manipulated the contextual variable, specifically, the availability of certain toys, to assess the effect on infant communicative behavior and parent responsiveness. Parents and their infants came to the lab for one 20–30 minute session; for the present study, our focus is on the play session that occurred during the first 20 minutes. During each play session, the parents were instructed to play as they normally would at home, and they were free to move around the room and interact with the toys as they wished. Toys were selected and categorized into one of two sets based on their properties during manipulation (Table 1). Feedback toys were defined as producing a response such as an electronic sound or automatic movement in response to manipulation. Examples include an airplane that produces music and lights when certain pieces are placed in the correct position and a farmhouse that produces sounds. Traditional toys were defined as having no automatic or electronic response after manipulation. Examples include a set of nesting cups and balls. While it is possible that traditional toys could elicit a sound (e.g., banging two nesting cups together), the intended function is to nest the cups, not produce a sound or movement.
Description of feedback and traditional toys.
Each parent–infant dyad participated in one of two conditions. In Condition 1, the play session was divided into two 10-minute periods, with one period for each toy set, feedback and traditional toys. The order of toy sets was counterbalanced across participants (e.g., traditional set followed by feedback set or feedback set followed by traditional set). In Condition 2, feedback and traditional toys were combined for the duration of the play period. Half of the toys were selected from each toy type category to maintain the same total number of available toys for the play session as in Condition 1. See Table 1 for the selection of toys for each condition.
Data coding
Parent–infant behavioral and vocal interactions were coded using Elan (The Language Archive; https://tla.mpi.nl/tools/tla-tools/elan/) software (Lausberg & Sloetjes, 2009, 2015). During the entire play session, the following variables were coded: (1) infant vocalizations, (2) infant gestures, (3) infant attentional shifts, (4) parent contingent feedback to infants’ vocalizations and gestures, and (5) parent sensitivity.
Infant vocalizations
Any sound that infants produced except distress vocalizations (e.g., fussing, crying) and vegetative sounds (e.g., coughing) were coded. Infant utterances were divided into syllables, defined as a rhythmic unit in the vocal stream containing at least one vowel (Nathani & Oller, 2001). Each syllable was categorized as directed (i.e., object, parent) or undirected (Goldstein et al., 2010; Gros-Louis et al., 2014). Object-directed vocalizations were coded when an infant vocalized while looking at a toy that was either within reach or that he/she was holding (Goldstein et al., 2010). Parent-directed vocalizations were coded when an infant vocalized while looking at the parent (Miller & Gros-Louis, 2013). In Condition 2 in which feedback and traditional toys were combined, parent-directed vocalizations were coded during episodes surrounding the particular toy. For example, if the infant was playing with the Fisher Price Amazing Animals Press Go and then looked up at the parent and vocalized while looking at the parent, the vocalization would be coded as parent-directed with feedback toy. Undirected vocalizations were coded when the infant vocalized while looking around the room or without a clear focus.
Infant gestures
The criteria for coding gestures were modified from Iverson and Goldin-Meadow (2005): deictic, conventional, and ritualized reaches. Deictic gestures are defined as referencing something in the immediate environment, such as an object. Three separate deictic gestures were coded: (a) show: holding up an object toward the social partner with the arm extended at an angle greater than 45 degrees from the infant’s torso; (b) palm point: extension of the arm with the palm facing down and more than one finger extended toward an object and/or social partner; (c) index point: extension of the arm with the index finger extended toward an object and/or social partner. Conventional gestures are defined as head, hand and/or arm movements that are culturally defined (e.g., shaking the head ‘no’). Ritualized reaches are one or two arm extensions toward an object and/or social partner. All of the gestures are independent of holding an object or touching a social partner except for show, which requires the use of an object.
Each gesture was categorized as directed (i.e., object, parent) or undirected (Miller & Lossia, 2013). Object-directed gestures were scored when an infant gestured to an object while caregiver-directed gestures were scored when an infant gestured to the parent. In Condition 2, parent-directed gestures were coded during episodes surrounding the particular toy. For example, if the infant looked at the Battat Toys Snug Bugs and then pointed at the parent, the point would be coded as parent-directed at a traditional toy. Undirected gestures were scored when an infant gestured to neither the parent nor an object in the room.
Infant visual attention
Infant attention (i.e., visual attention) was coded following the criteria of Miller, Ables, King, and West (2009) and Miller and Gros-Louis (2013). The frequency and duration of infants’ visual attention was coded as looking at a toy (look toy), looking at the parent (look parent), or looking around the room without a clear focus (look undirected). Both the visual gaze and body orientation of infants were used as a proxy for visual attention.
Parent contingent feedback to infant communicative behavior
Parent contingent responses to infant vocalizations and gestures were coded within 2s of the offset of the infant’s behavior (Goldstein & Schwade, 2008; Goldstein, West, & King, 2003; Miller, 2014). Responses were not coded to cries, raspberries or vegetative sounds.
Parent sensitivity
Sensitive behavior is defined as an appropriate and contingent response to the infant’s behavior or focus of attention (Bornstein, 1989). Redirective behavior is defined as contingent, but inappropriate to the infant’s behavior or focus of attention by interrupting or redirecting the flow of the infant’s behavior or attention (Baumwell et al., 1997). Sensitive and redirective parent behavior was coded as: vocalizations, manipulations, and vocal–manipulation combinations (i.e., co-occurrence of the two behaviors).
Vocalizations
Every utterance of the parent was coded as either sensitive or redirective (Miller et al., 2009; Miller & Gros-Louis, 2013). Sensitive vocalizations were coded when the parent commented on what the infant was visually attending to or playing with or about the state of the infant. Redirective vocalizations were coded when the parent commented on an object the infant was not visually attending to or playing with. For another vocalization to be coded, a break in speech streams had to be greater than or equal to 0.5s (Miller et al., 2009; Miller & Gros-Louis, 2013). For example, if a child is attending to and playing with a truck, a parent responding sensitively may say ‘That’s a firetruck,’ while a parent responding redirectively may refer to another object or topic in the room, and say ‘Let’s play with these balls,’ which could potentially redirect the infant’s attention.
Manipulations
Every parent interaction with an object was coded as either sensitive or redirective. A sensitive manipulation was coded when the parent touched an object the infant was attending to and a redirective manipulation was coded when the parent touched an object the infant was not attending to. Note that it was not enough for the infant to just be holding the object, the infant had to attend to the object and in most cases, the infant was also engaged with the object. For example, if the infant was attending to and interacting with the truck, a sensitive manipulation may include the parent showing the infant how to manipulate certain parts of the truck. A redirective manipulation includes the parent touching another object in the room, such as a ball, which could redirect the infant’s attention. Additional manipulations were coded when the parent had their hands off the objects for 0.5s or more and subsequently touched the object again (Miller et al., 2009; Miller & Gros-Louis, 2013).
Vocal–manipulation combinations
A combination of vocalizations and manipulations was coded when the onset of the vocalization and manipulation occurred within 0.5s of one another. For example, if a child is attending to the truck, a parent responding sensitively may say ‘That’s a firetruck,’ while showing the infant how to manipulate certain parts of the truck. In the same scenario, a parent responding redirectively may say ‘Let’s play with these balls,’ while at the same time manipulating the balls.
In Condition 2, a distinction was made to differentiate between interactions with feedback and traditional toys, similar to coding infant directed vocalizations and gestures.
Intraclass correlation coefficient is a common method in behavioral video analysis to assess inter-rater reliability (e.g., Wu & Gros-Louis, 2014). In the present study, a second coder recoded 20% of the parent–child interactions and a two-way random effects model with absolute agreement as the type was used (Shrout & Fleiss, 1979). A high degree of reliability was found across the five measures. The average intraclass correlation and 95% confidence intervals are reported for each measure: infant vocalizations: .993 (95% confidence interval: .966–.999), infant gestures: .970 (95% confidence interval: .950–.982), duration of infant attention: .998 (95% confidence interval: .997–.999), frequency of infant attention: .953 (95% confidence interval: .847–.982), parent contingent feedback: .999 (95% confidence interval: .997–.999), and parent sensitivity: .993 (95% confidence interval: .989–.996). The reliabilities in the present study are consistent with other studies reporting similar kinds of behavioral coding (e.g., Goldstein & Schwade, 2008; Wu & Gros-Louis, 2014).
Data analysis
To address the question of how toy properties affect infant and parent behavior, we examined variables that have been shown to be important for language development. Therefore, undirected behaviors were not included in analyses. As a result, only infant directed behavior to toys and parents when playing with feedback and traditional toys were used in the remaining analyses. In particular, we were interested in whether toy properties had an effect on three infant behaviors (directed vocalizations, directed gestures, joint attention) and two parent behaviors (contingent feedback to vocalizations and gestures, and sensitivity).
Infant communicative behavior resulted in two broad categories. First, the frequency of object-directed and parent-directed vocalizations was tabulated separately for each toy type resulting in total scores of combined directed vocalizations for feedback and traditional toys. Second, due to the low number of deictic, conventional, and ritualized reaches produced, gesture categories were combined. In addition, the frequencies of object-directed and parent-directed gestures were combined resulting in scores of directed gestures for feedback and traditional toys. The rate of each behavior was calculated per minute. For infant attention, we were specifically interested in the average duration of attention to particular kinds of toys. The average duration of attention was calculated for engagement with toy (total msec of attending to toy category/frequency of attentional shifts) resulting in average duration of attention to feedback and traditional toys.
Parent responsiveness resulted in two categories. First, the frequency of contingent feedback to infants’ directed vocalizations and gestures was calculated separately for feedback and traditional toys. Second, two scores were calculated to obtain global parent sensitive and redirective behavior. A sensitive score was obtained by adding the frequency of sensitive vocalizations, sensitive manipulations, and a combination of both. Likewise, a redirective score was obtained in a similar fashion but by adding the frequency of the same redirective categories.
Descriptive statistics are provided for frequency of infant communicative behavior, infant attention behavior, parent sensitive behavior, and parent contingent feedback to each type of infant communicative behavior (Table 2).
Descriptive statistics of infant and parent behavior.
Means (SD) presented without data transformation.
Results
A mixed repeated-measures ANOVA was conducted to examine the effect of Toy Set (traditional, feedback) on frequency of infants’ directed vocalizations with Condition (Condition 1, Condition 2) as the between-subjects factor. Infants produced more directed vocalizations when interacting with traditional toys (M = 3.881) than when interacting with feedback toys (M = 2.822), F (1, 48) = 7.702, p = .008, η p 2 = .138, observed power = .776. There was no interaction between directed vocalizations and condition, F (1, 48) = 3.095, p > .7, η p 2 = .016, observed power = .138.
A mixed repeated-measures ANOVA was conducted to examine the effect of Toy Set (traditional, feedback) on the frequency of parent responsiveness to infants’ directed vocalizations with Condition (Condition 1, Condition 2) as the between-subjects factor. Parents responded to infants’ directed vocalizations when interacting with traditional toys (M = 1.327) more than when interacting with feedback toys (M = 1.003), F (1, 48) = 6.401, p = .015, η p 2 = .118, observed power = .698. There was no interaction between the conditions, F (1, 48) = 1.144, p > .2, η p 2 = .023, observed power = .182.
A mixed repeated-measures ANOVA was conducted to examine the effect of Toy Set (traditional, feedback) on the frequency of infant gesture production with Condition (Condition 1, Condition 2) as the between-subjects factor. Across toy sets, infants produced more directed gestures when interacting with traditional toys (M = 0.300) than when interacting with feedback toys (M = 0.179), F (1, 48) = 5.051, p = .029, η p 2 = .95, observed power = .596. There was also an interaction between directed gestures and condition, indicating that infants produced different levels of gestures between the conditions, F (1, 48) = 5.416, p = .024, η p 2 = .101, observed power = .626. Specifically, infants in Condition 1 produced more directed gestures than infants in Condition 2 when interacting with both traditional (MCondition 1 = 0.400, MCondition 2 = 0.087) and feedback toys (MCondition 1 = 0.220, MCondition 2 = 0.090).
A mixed repeated-measures ANOVA was conducted to examine the effect of Toy Set (traditional, feedback) on the frequency of parent responsiveness to infants’ directed gestures with Condition (Condition 1, Condition 2) as the between-subjects factor. Parents responded to infants’ directed gestures when interacting with traditional toys (M = 0.230) more than when interacting with feedback toys (M = 0.117), F (1, 48) = 3.883, p = .055, η p 2 = .075, observed power = .486. There was also an interaction between the conditions, F (1, 48) = 4.712, p = .035, η p 2 = .089, observed power = .566; parents in Condition 1 responded to more gestures than parents in Condition 2 when interacting with both feedback (MCondition 1 = 0.274, MCondition 2 = 0 .053) and traditional toys (MCondition 1 = 0.141, MCondition 2 = 0.059).
A mixed repeated-measures ANOVA was conducted to examine the effect of Toy Set (traditional, feedback) on average duration of infant attention with Condition (Condition 1, Condition 2) as the between-subjects factor. Infants had a longer average duration of attention when interacting with feedback toys (M = 74.492) than when interacting with traditional toys (M = 49.37), F (1, 48) = 34.641, p < .001, η p 2 = .419, observed power = 1.000. There was no effect of condition, F (1, 48) = 3.308, p > .07, η p 2 = .064, observed power = .430.
A mixed repeated-measures ANOVA was conducted to examine the effect of Toy Set (traditional, feedback) on total levels of parent sensitive and redirective behavior with Condition (Condition 1, Condition 2) as the between-subjects factor. There were no differences in parent sensitive behavior when interacting with either set of toys, F (1, 48) = 0.425, p > .5, η p 2 = .009, observed power = .098. However, parents displayed a higher level of redirective behavior when interacting with traditional toys (M = 3.723) than when interacting with feedback toys (M = 2.900), F (1, 48) = 8.241, p = .006, η p 2 = .147, observed power = .803. There was no effect of condition, F (1, 48) = 1.716, p > .19, η p 2 = .035, observed power = .250.
Discussion
There were four main findings in this study: (1) infants varied their vocal and gestural behavior as a function of the toys present in the play session; (2) moment-to-moment measures of parent contingent feedback to infant vocal and gestural behavior differed, most likely as a result of the differences in infant communicative production; (3) infants’ cognitive measures varied as a function of the toys present in the play session; and (4) global measures of parent responsiveness differed based on toys present in the play session.
The present study revealed that toy type influenced the quality and quantity of parent–child interactions. Previous studies on toy type have typically focused on frequency of infant vocalizations (e.g., Sosa, 2016) and measures of parent language production (Caldera et al., 1989; Gros-Louis et al., 2016; Sosa, 2016). This study extends the recent findings of Sosa (2016) and Gros-Louis et al. (2016) to depict a detailed behavioral analysis of both parent and infant behavior during play sessions with different toy types: those that produce feedback and those that do not. This is significant as quality and quantity of parent–child interactions are associated with language outcomes (Baumwell et al., 1997; Hoff, 2006; Lewis & Gregory, 1987; Tamis-LeMonda et al., 2001).
Infants produced more directed vocalizations and gestures when interacting with traditional compared to feedback toys. These findings add to a growing number of studies that have documented the relation between contextual variables and language measures in children (Gros-Louis et al., 2016; Lewis & Gregory, 1987). Sosa (2016) found that infants in the same age range produced fewer vocalizations when interacting with electronic toys compared to interacting with books, but did not find any differences in infant vocal production between electronic and traditional toys. One difference between our study and Sosa’s (2016) study is the way infant vocalizations were coded. Specifically, we examined communicative behavior surrounding the play session; vocalizations directed either toward the object or parent were our primary focus. In the Sosa (2016) study, all vocalizations, regardless of how the infant directed the vocalization, were analyzed. That is, undirected vocalizations (i.e., those produced without a clear directed focus) were also included in the analysis. It is quite possible that if the usage of directed vocalizations was taken into account, the results would align with the present study. Another difference between the two studies is the selection of toy sets. The electronic toys in Sosa (2016) consisted of three battery-operated toys that produced sound, lights, words, and phrases while the present study consisted of eight toys that produced lights, sound, and actions upon manipulation. Nonetheless, there were more infant vocalizations during the book context compared to play with electronic toys, which, similar to our findings, suggests that infants produce more communicative behavior when interacting with more traditional toys (Sosa, 2016).
Our findings also extended those of Sosa (2016) and Gros-Louis et al. (2016) by including gestures, which at this age are just starting to emerge in the communicative repertoire of the infant. Importantly, both gestures and vocalizations, and responses to these prelinguistic behaviors, are associated with language outcomes (Bates & Dick, 2002). This is important because infants who produce more object-directed communicative behavior have more opportunities to learn about their environment. For example, parents produce more information after an object-directed vocalization, which has lasting consequences on language measures during the second year of life (Goldstein et al., 2010). Likewise, the number of different objects that a toddler points to at 14 months of age predicts vocabulary size at 42 months and is actually a better predictor than maternal speech input (Rowe, Ozcaliskan, & Goldin-Meadow, 2008). This is related to the fact that mothers ‘translate’ what the apparent referent of the point is (Goldin-Meadow, Goodrich, Sauer, & Iverson, 2007); in turn, what the child gestured to the prior month predicts the verbal lexical item in subsequent months (Iverson & Goldin-Meadow, 2005).
Parents also responded with higher levels of contingent feedback to infants’ directed communicative behavior (e.g., gestures, vocalizations) when interacting with traditional compared to feedback toys. This replicates the findings of Sosa (2016), who also found that parent responses were higher in the book and traditional toy contexts compared to when playing with the electronic set of toys. Similarly, Gros-Louis et al. (2016) found that parents responded more during a book context than when playing with a barn that contained manipulative parts such as buttons to push and animals to squeeze that produced a squeak, which in the present study would be classified as a feedback toy.
Parents who provide contingent feedback to infants’ communicative behavior create opportunities for infants to learn about the consequences of their own behavior (Goldstein & Schwade, 2008). In a naturalistic study, parents’ contingent labeling responses to their infants’ object-directed vocalizations were positively associated with language development at 15 months (Goldstein & Schwade, 2010). Furthermore, an experimental study found that infants who received a label contingent on their object-directed vocalization learned the association between the object and label, whereas infants who received a label contingent on a look alone, did not learn the association (Goldstein et al., 2010). A growing number of studies have demonstrated that increasing levels of contingent feedback to infants’ communicative behavior increases the quality and quantity of prelinguistic vocalizations (Goldstein & Schwade, 2008; Goldstein et al., 2003) and gestures (Miller & Lossia, 2013).
A third finding of the present study was differences in infant cognitive measures across toy sets. Infants had, on average, shorter periods of sustained attention when interacting with traditional compared to feedback toys. While previous studies have argued that the development of attention is related to internal factors (Richards & Anderson, 2004), recent studies have suggested that sustained attention is actually influenced by parent–child social interactions (Miller et al., 2009; Miller & Gros-Louis, 2013; Yu & Smith, 2016). The current results add to the growing literature on external factors influencing sustained attention and suggest that attention also may be influenced by situational factors such as toys. The sounds and lights emitted by electronic toys are effective at activating infants’ orientating reflex, and given that infants are likely to orient to novel stimuli, differences in sustained attention may have resulted from either the toy properties, novelty, or a combination of both (Radesky & Christakis, 2016).
In addition, parents produced less redirective behavior when interacting with feedback compared to traditional toys. The results for infant attention and parent sensitivity parallel those from studies on experimental manipulations of social partner sensitivity. That is, when levels of redirective behavior increase, infants shift their attention more frequently and decrease their duration of visual attention (Miller et al., 2009; Miller & Gros-Louis, 2013). What is unclear, however, is whether the toy type or increased redirective behaviors of parents influenced infant attentional behavior. Future experimental studies manipulating one while holding the other constant are necessary to illuminate the effects on infants’ attention. The average duration of attention to traditional toys, however, is comparable to that in other studies in which traditional toys were used (Miller & Gros-Louis, 2013). This suggests that electronic toys increase infants’ attentional focus and decrease attentional shifts.
The attention findings are somewhat at odds with the communication findings; that is, the typical relation is that longer sustained attention is related to frequency of communicative behavior such as gesture and vocal production (Miller & Gros-Louis, 2013). However, we found that infants had longer sustained attention with feedback toys, yet produced fewer communicative behaviors. It should be noted that in prior studies we have always used a majority of traditional toys, thus the findings seem to reflect the effect of electronic toys. The flashing lights and sounds of the feedback toys clearly capture and hold attention and, as noted above, ‘fill the auditory space’ which decreases infant communicative behaviors. One possible explanation for the attention differences as a function of toy type may be a novelty factor. Fantz’s (1964) seminal paper demonstrated that after repeated exposure to a visual stimulus, infants selectively attend to a novel stimulus compared to a familiar stimulus. It is possible that infants had less exposure to the toys in the feedback group, resulting in a higher level of novelty. In the present study, we asked parents prior to the play session if they owned any of the toys to assess familiarity. There was a significant association between toy type and exposure,
Taken together, one possible explanation for the reduced production of behavior during parent–infant interactions with feedback toys might be overall levels of toy stimulation. Many of the feedback toys produced noises, such as songs (e.g. Fisher Price Little People Lil’ Movers Airplane) or movement (e.g. Fisher Price Amazing Animals Press Go) that may have interfered with joint attention. Joint attention typically falls into two categories, responding to joint attention or initiating joint attention, but the common thread is a shared focus on an object, event or experience between two partners (Mundy & Newell, 2007). During periods of joint attention, both the parent and infant produce more utterances and engage in longer conversations. More importantly, maternal references to objects during joint attention episodes are positively related to vocabulary development at 21 months of age (Tomasello & Farrar, 1986). It could be that the extra stimulation from the feedback toys distracts from important characteristics of interaction that occur during joint attention episodes and contribute to language development. Other studies with electronic toys have also noted that toy features are often the likely culprit for filling the auditory space. Sosa (2016) suggested that parents may let the electronic toys do the talking for them. Likewise, Wooldridge and Shapka (2012) noted that during play sessions, multiple electronic toys were emitting noises simultaneously, making it more challenging to engage in conversational turn-taking.
A few limitations of the present study should be noted. First, we focused exclusively on moment-to-moment (i.e., contingent feedback to infant communicative behavior) and global (i.e., sensitive, redirective behavior) levels of parent responsiveness. Previous studies have explored the type of parent language that may differ as a result of contexts (Sosa, 2016). While it is clear that the type of parental vocalization is important to language development (Tamis-LeMonda et al., 2001), both of the parent responsiveness measures used in the present study are also important in language development (Dunst, Gorman, & Hamby, 2010; Tamis-LeMonda Shannon, Cabrera, & Lamb, 2004). Second, our sample is homogeneous and most of the parents are from middle and upper income households. Previous studies have shown that socioeconomic status (SES) is one source of variation in measures of language development (Hoff, 2013). Children from lower-SES homes show lower levels of language skills than children from mid- and high-SES homes on measures of language processing, language comprehension, and language production starting as young as 18 months (Fernald, Marchman, & Weisleder, 2013). Parent speech also differs as a function of demographic factors. Higher-SES parents talk more than lower-SES parents and there are SES differences in child-directed speech (Hoff-Ginsberg, 1998), and amount of vocabulary and syntax (Huttenlocher, Vasilyeva, Cymerman, & Levine, 2002). Future studies should explore parent–infant interactions using electronic and traditional toys in a more diverse sample. Third, there was a significant association between toy and exposure; that is, infants were more likely to have the traditional toys at home. Future studies should ensure that all toys, regardless of context, are novel. This could be achieved by removing batteries from electronic toys and using them also as the non-feedback set. Finally, we did not measure language over time; the current study aimed to examine parent–infant interactions in everyday interactions with various toy properties to understand how contextual variables contribute to factors that influence language learning in moment-to-moment social interactions. Future studies should follow children longitudinally with a particular emphasis on contextual variables such as toys to examine the relation between toy properties and language development.
The present study suggests that contextual variables contribute to important parent and infant variables responsible for language development during transactional social and communicative interactions (Fiese & Sameroff, 1989). Because social processes, such as parent responsiveness to infants’ attention and behavior, are related to many facets of an infant’s development (Baumwell et al., 1997; Tamis-LeMonda et al., 2001, 2004), it is important to understand the contextual factors that influence learning and development in the social environment where parent–infant interactions occur (Gros-Louis et al., 2016). Results from the present study suggest that in the common context of play, different toys may provide different learning opportunities: infants vocalized and gestured more frequently when playing with the traditional toys compared to the feedback toys, parents responded with more contingent responses in the traditional toy context, and infants’ sustained attention was longer in the feedback toy context, but they communicated less. Future work should examine if these results persist over developmental time within contexts to determine if the properties of feedback toys are truly interfering with communicative interactions (Sosa, 2016). In addition, studies should examine how contextual variables relate to real-time learning; that is, how do changes in parent–infant interactions within contexts relate to outcome measures (Goldstein et al., 2010). Future work should continue to take an ecological approach to the study of language by incorporating contextual variables into the study of development. By understanding the relation between contexts and changes in parent and infant behavior in everyday interactions, we can begin to understand how moment-to-moment interactions influence both real-time learning and long-term developmental outcomes (Gros-Louis et al., 2016; Spencer & Perone, 2008).
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.