Push or Carry? Pragmatic Opportunities for Language Development in Strollers vs. Backpacks

Abstract

Strollers and backpacks are employed early, frequently and throughout the first year, with parents overwhelmingly using strollers. However, because these transport modalities put infants in different proximities to caregivers, postures, and states of alertness, their use may translate to different opportunities that are of developmental consequence, particularly with regard to language. We used GoPro technology in a within subjects counter-balanced design to measure dyadic vocalizations in strollers and backpacks with 7-to-11-month-old infants. Parent-infant dyads (N=36), who regularly used both transport modes took two 8-minute walks in their own neighborhoods using their own carriers while wearing lightweight head-mounted GoPros. There was significantly more parent speech, infant vocalizations, dyadic conversations and infant-initiated speech in backpacks, as well as more head motion consistent with visual scanning by infants. Backpacks appear to be a practical way to encourage more engaging, language-enriched developmental opportunities in the critical first year.

Parents commonly use strollers and backpacks to transport infants in a variety of settings. Surprisingly, no studies have compared the perceptual, linguistic, or social-emotional affordances provided by these two common modes of infant transport despite the potential developmental consequences of their use. Market research shows that 99% of new mothers own a stroller, while only 17% own a backpack (American Baby Group, 2006). Birken, Lichtblau, Lenton-Brym, Tucker, Maguire, Parkin et al. (2015) report wide use of strollers (e.g., 5.8 million sold in the United Stated in 1998 alone), with families owning two to three strollers on average despite the possibility that strollers increase passivity, restrict movement and reduce physical activity. Although caregivers prefer strollers to backpacks, these different modes of infant transport may not equally optimize developmental opportunities in part because they put infants in different postures relative to the environment (Adolph, 2014), different proximity to caregivers and different states of alertness (Fredrickson & Brown, 1975). Specifically, infants in strollers are typically semi-reclined approximately18 inches off the ground, and are usually facing away from the caregiver, who is 22–24 inches behind and above the infant. By contrast, infants in backpacks are seated upright approximately 12 inches behind the caregiver, and within 8–10 inches of the caregiver’s visual field. This closer proximity and shared visual field may maximize certain developmental opportunities for the infant.

At the most basic level, upright posture facilitates alertness in infants (Fredrickson & Brown, 1975). In addition, being upright fosters opportunities for joint attention to objects or events beyond the dyad (Deak, Triesch, Krasno, de Barbaro, & Robledo, 2013), as well as reciprocal vocalization (Anderson, Vietze, & Dokecki, 1977; Langfur, 2013), and shared social and emotional communication (Clearwater, 2011). Therefore, these experiences might be expected to increase for infants in backpacks. Fogel, Messinger, Dickson, & Hsu (1999) also proposed that being upright and engaged with a caregiver, “…offers entirely new and unanticipated possibilities – such as expansion of the range of visual attention – that move the dyad to discover novel ways of relating to each other” (p. 331). They found that mothers who used postures that allowed for more physical contact with their infants reported more feelings of closeness and affection. Backpacks, but not strollers, afford such physical contact as caregivers can easily touch infants and can jostle or bounce them as part of the ride. In short, backpacks may increase the opportunity for shared perceptual, linguistic, kinesthetic, and social-emotional experiences between caregivers and infants.

Field, Malphurs, Caraway, & Pelaez-Nogueras (1996) reviewed a number of studies showing that infants transported in carriers (as opposed to infant seats) exhibit more smiling and vocalizing, and less crying. Using a within-subjects design, Field et al. (1996) found differences in infants’ behavior when they were carried facing out (toward the environment) than facing in (toward the caregiver). The 3-to-4-month-olds in their sample were eight times more likely to be in an “active awake” state (e.g., significantly more head turning and limb movement, and significantly less likely to fall asleep) in the facing out position, putting the infant “in a state of readiness to be active and explore the environment” (p. 53). Field et al. (1996) predicted regular “facing out” carrying experience in the first year might translate to superior exploratory, motor, and cognitive development by the first birthday, as well as a higher incidence of secure attachment. Both backpacks and strollers require infants to face out, however, backpacks may allow the infant to be more upright, engaged, and kinesthetically stimulated.

Consistent with the latter are studies of Kangaroo Mother Care (KMC), an intervention for pre-term infants in which they are held upright and against their mothers’ skin 24/7 until they reach full-term gestational age. Randomized controlled studies of KMC show long-term effects across a wide variety of domains a full year after the relatively short intervention period (Tessier, Charpak, Giron, Cristo, de Calume, & Ruiz-Pelaez, 2009). Specifically, KMC is related to higher cognitive scores at one year of age compared to standard care. Tessier et al. (2009) proposed that infants in KMC received more kinesthetic and vestibular stimulation, which may have made them more responsive to the environment, feeding back to caregivers, who provide additional stimulation. Although backpacking an infant later in the first year is not equivalent to KMC, these findings do suggest that backpacking may be associated with similar, albeit not as dramatic, effects perhaps providing an “alertness advantage”.

Similarly, Anisfeld, Casper, Nozyce & Cunningham (1990) propose that proximity alone is likely to make caregivers more aware of and responsive to their infants’ needs and states. In the rationale to their longitudinal experiment in which mothers were randomized to carriers (i.e., a device allowing the infant to be transported on the caregiver’s body) or infant seats, they predicted “A mother can more easily recognize prodromal signs of hunger or discomfort in a carried infant than an infant in a crib or stroller at some distance from her” (p. 1618), a prediction which was supported based on measures of secure attachment at 13 months. Anisfeld et al. (1990) also found that infants transported in carriers (vs. infant seats) through the first year exhibited less proximity-seeking behavior (i.e., less vocalizing when alone, later to socially smile, less regular bouts of crying) presumably because their greater proximity did not require them to activate the attachment system. The experimental basis of these findings prompted Anisfeld et al. (1990) to draw a causal relationship between carrier use and secure attachment, whereby carriers promote a “different ecology of interaction” between the dyad, and whereby caregivers become more “contingently responsive” to the infants’ states, needs, and behaviors (p. 1625).

However, it is not just the caregiver who receives feedback about the infant s/he is carrying; the infant also receives feedback about the caregiver. The “yoked” relationship between the caregiver and infant in a backpack may be thought of as analogous to a horse-and-rider, in which the infant receives information about the caregiver’s posture, position, and movement through the environment. Empirical evidence on therapeutic horseback riding has found that intervention pushes the rider’s physical and psychological development by requiring him/her to respond to and adjust his/her own body movements to the constant vestibular and kinesthetic feedback of the horse (Zadnikar & Kastrin, 2011). Backpacks replicate this effect, while strollers do not.

One reason that comparing transport modalities is important is because parents employ them early, frequently, and throughout the first year. Given the importance of the first year to major developmental milestones like language (Kuhl, Conboy, Padden, Nelson, & Pruitt, 2005) and attachment (Bowlby, 1988) and their relationship to later outcomes like IQ, achievement (Colker, 2016) and social-emotional adjustment (Schimmenti & Bulfuco, 2015), identifying accessible and practical cognitive, social and emotional opportunities for infant-caregiver interaction has important implications for development in infancy and beyond.

The purpose of the present study was to compare infant and parent dyadic engagement in strollers and backpacks in a naturalistic within-subjects counterbalanced design, employing participants who regularly use both carrying modes. We hypothesized that both infants and parents would vocalize more, and that infants would more actively scan their surroundings when traveling in backpacks than strollers.

Method

Participants

The present study was employed APA ethical guidelines for conducting research with human participants, with written informed consent obtained from parents for each infant before any data collection. The Institutional Review Board at Johnson State College approved all procedures. Infants (N=36; 21 males) ranged from 7-to-11 months (M=33.7 weeks, SD=5.2). We purposely recruited infants in this age range: 1) to insure they had adequate neck strength to be carried in a backpack, 2) because most backpack manufacturers do not recommend use until 6-months due to neck strength, and 3) because the second half of the first year has been identified as a sensitive period for language development (Kuhl et al., 2005). Most infants (97%) participated with their mothers (M= 31.8 yrs, SD= 4.7), who tended to be married (94%), college-educated (83%) and of middle-income (M=$81.5k, SD=$39.7). All parents owned and regularly used both strollers and backpacks, with roughly half preferring strollers (n=16), half preferring backpacks (n=14), and the rest reporting no preference (n=6). There was no significant difference in parents’ regular weekly use of strollers (M=3.76, SD=2.96) and backpacks (M=4.55, SD=4.25), t(39)= −0.90, p>.05.

Materials

GoPro Hero 4

Lightweight (2.4 oz.) GoPro Hero4 video cameras were affixed to parents’ and infants’ heads via ProGear Adjustable Kids’ head mount for GoPro Hero 4/3 & 3/2/1 Ages 3-to-14, which fit adult participants as well.

Measures

Self-report questionnaires

Parents were e-mailed demographic questions several weeks in advance of the procedure to be collected at the time of the experimental protocol, which asked them to indicate on a 7-point bipolar scale (1=stroller, 4=both the same, 7=backpack) their perceptions regarding differences in their own and their infants’ stroller/backpack behavior, as well as their differential use of and preference for strollers and backpacks. For example, parents indicated where on the scale they and their infants each appeared “more talkative” (1=stroller, 7=backpack). Parents were also asked which transport mode they preferred, and specifically how many times per week they used each mode.

GoPro audio/video

We had originally intended to synchronize the video feeds, but this was unnecessary as infants’ cameras easily picked up both parents’ and infants’ vocalizations due to the high quality of the video and audio. Therefore, frequency and duration of infants’ and parents’ vocalizations were coded from the audio feed of the camera worn by the infants.

Specifically, we coded the frequency and duration (in seconds) of infant and parent vocalizations, infant-initiated and parent-initiated conversations, concurrent/overlapping vocalizations, as well as conversational units. The latter were defined as vocal exchanges that were initiated by one member of the dyad and reciprocated immediately by the other member. A conversational unit ended when more than one second elapsed with neither member vocalizing. Thus, a conversational unit was continuous with no gaps in order to be consistent across measurement.

Infants’ visual scan was also coded on a scale of one to three. Specifically, the footage was divided into 30-second intervals, with the first and last five seconds of footage coded on either end of each interval. Those 10-seconds of footage were coded on the degree to which the infant’s head moved across the visual field (1 = did not move; 2 = moved somewhat; 3 = moved greatly). Since head- mounted cameras only indicate head movement and cannot accurately indicate the direction of the infants’ gaze (Smith, Yu, Yoshida, & Fausey, 2015), these data were collected as only a rough approximate of whether infants may have been more likely to be scanning the environment in strollers or backpacks.

Two trained research assistants worked in tandem to code variables from the GoPro video and audio feeds. Given that video footage revealed the experimental conditions, it was not possible for video coders to remain blind to the conditions. However, they were unaware of the directional hypotheses. Cohen’s kappa across language categories and visual scan based on a random selection of 25% of the videos ranged from .89 to .99.

Procedure

Participants were recruited using public birth records from the Vermont Department of Health to locate infants who born 6-to-12-months earlier and residing in one of three counties proximate to the PI’s laboratory. To protect the hypotheses, prospective participants were mailed a flyer for a study investigating “infants’ experiences in strollers and backpacks”. The flyer was also posted on the laboratory’s social media page. Infants had to have been full-term at birth and 6-to-12-months old upon enrollment, and parents had to own and regularly use a structured-frame front-facing backpack and a front-facing stroller. Interested parents emailed the PI, who obtained informed consent and mailed them the questionnaires to be collected before the procedure several weeks later. Despite the self-report measures, none of the parents gave any indication of having detected the hypothesis, many reported not having considered whether their infants behaved differently in either modality, and were only attuned to which device suited convenience at the time of its use.

Research assistants contacted participants via text, phone, or email (as preferred by participants) to schedule the procedure in participants’ typical walking locales (i.e., neighborhood, local public park, recreational path, etc.) using participants’ own familiar strollers and backpacks to mimic naturalistic conditions. Parents and infants were each outfitted with head-mounted GoPro Hero 4 video cameras. These devices are lightweight (2.4 oz.), weighing the approximate equivalent of 2 slices of bread or one half of a small apple. Infants tolerated the head-mounted cameras well, and only two infants attempted to pull the camera off. In both cases, the camera was repositioned and the procedure restarted without incident. (See Figure 1.)

Figure 1.

GoPro Hero 4 cameras as affixed to an infant and mother in the backpack condition.

Researchers provided parents with a timer set for four minutes and instructed them to start the timer, begin walking as they typically would until the timer alerted them to turn around, and to retrace their steps back to the research assistant. Once back, the infant was switched to the other mode of transport, the timer was reset, and parents took the second walk following the same route and again turning around and retracing their steps when the timer indicated. Thus, there were two 8-minute walks in each modality. The order was counterbalanced. No other persons (or pets) were allowed to be present on the walk. The first and last minute of each walk were not coded to insure infants were not distracted by the presence of the research assistant, who remained stationary at the starting and ending point, leaving 6 minutes of footage available to code per modality.

Results

Independent samples t-tests on the self-report questionnaire data were performed to compare parents who preferred strollers (n=16) to parents who preferred backpacks (n=14). Regardless of their self-reported preferences, parents perceived both themselves and their infants as significantly more talkative in backpacks, as well as being significantly more likely to touch their infants or to be approached by strangers when carrying them in backpacks. (See Table 1.)

Table 1.

Parent-reported Perceptions of Behavior x Parent-Preferred Carrier Type

	Parent-Preferred Carrier Type
	Stroller (M, SD)	Backpack (M, SD)	t(28)	d
Self-report measures
Infant more talkative	3.8 (1.6)	5.5 (1.1)	3.27**	−1.23
Parent more talkative	4.2 (.83)	5.6 (1.5)	−3.12**	−1.16
Touch Infant	4.7 (1.5)	5.9 (1.5)	−2.27**	−.83
Approached by Stranger	2.8 (1.3)	4.1 (1.3)	−2.65*	−9.68

Note.

^*= p ≤ .05,

^**= p ≤ .01

Note: Ratings are on a 7-point bipolar scale (1= stroller, 4 = no difference, 7=backpack)

Audio and video data harvested from infants’ GoPro’s corroborated the self-report data. Paired samples t-tests showed that parents talked significantly longer when using backpacks (M=65.9_seconds, SD=69.14) than strollers (M=52.37_seconds, SD=63.94), t(40)=−2.26, p<.05, d = −.36. No other differences in duration were found in the vocalization categories. However, when frequency was used as the metric, infants vocalized significantly more often (f) in backpacks (M=7.83, SD=8.13) than strollers (M=5.02, SD=7.33), t(40)= −2.60, p<.001, d = −.41, and there were significantly more dyadic conversations in backpacks (M=3.98, SD=4.70) than strollers (M=2.10, SD=3.41), t(40)= −3.00, p<.001, d = −.49. Infants also initiated significantly more conversations in backpacks (M=2.98, SD=3.58) than strollers (M=1.37, SD=2.03), t(40)= −3.36, p<.001, d = −.59, and appeared to visually scan their surroundings more in backpacks (M=1.77, SD=.31) than strollers (M=1.33, SD=.28), t(40)= −7.15, p<.001, d = −1.12. (See Table 2.) Significant results were subsequently tested for order and age effects, but none were found.

Table 2.

Observed Behavior x Carrier Type

	Carrier Type
	Stroller (M, SD)	range	Bkpk (M, SD)	range	t(40)	d
Go-Pro measures
Parent Vocalizations (seconds)	52.37 (63.94)	0–304	65.90 (69.14)	0–312	−2.26 *	−.36
Parent Vocalizations (f)	16.34 (14.87)	0–60	20.00 (15.70)	0–65	−1.65
Infant Vocalizations (seconds)	25.83 (42.15)	0–199	29.56 (44.69)	0–209	−0.49
Infant Vocalizations (f)	5.02 (7.33)	0–35	7.83 (8.13)	0–35	−2.60***	−.41
Dyadic conversations (seconds)	23.98 (41.28)	0–167	34.80 (43.58)	0–187	−1.74
Dyadic conversations (f)	2.10 (3.41)	0–16	3.98 (4.70)	0–20	−3.00***	−.49
Infant-initiated conversations (f)	1.37 (2.03)	0–9	2.98 (3.58)	0–16	−3.36***	−.59
Parent-initiated conversations (f)	0.66 (1.41)	0–7	0.76 (1.14)	0–4	−0.43
Overlapping vocalizations (seconds)	5.22 (12.25)	0–61	4.63 (8.80)	0–48	−0.26
Overlapping vocalizations (f)	2.46 (4.67)	0–19	3.15 (4.81)	0–21	−0.78
Visual Scan (1–3 likert)	1.33 (.28)	1–2	1.77 (.31)	1.5–2.6	−7.15***	−1.12

Note.

^*= p ≤ .05,

^***= p ≤ .001

To further explore relations between measures, correlations were run between infant vocalizations (f), parent vocalizations (seconds), infant-initiated vocalizations (f), dyadic conversations (f), and visual scanning in backpacks (since that’s where the effects were found). Only two significant associations emerged but both were strong: the frequency of infant vocalizations with infant-initiated vocalizations [r(41) = 0.83, p <.000], and the frequency of infant vocalizations with dyadic conversations [r(41) = 0.88, p <.000].

Discussion

Strollers and backpacks are used early, frequently, and throughout the first year of life but may provide different affordances that are of consequence given the critical developmental milestones like language (Kuhl et al., 2005) and attachment (Bowlby, 1988) emerging within that time frame. Parents show an overwhelming preference for strollers (American Baby Group, 2006; Birken et al., 2015). However, strollers may not maximize opportunities for verbal practice and social-emotional connection simply because strollers decrease infants’ proximity to parents (Anisfeld et al., 1990), encourage postures that lower infants’ alertness (Fredrickson & Brown, 1975), prevent a shared kinesthetic experience (Tessier et al., 2009), and do not allow for easy and frequent reciprocal exchanges (Fogel et al., 1999).

This is the first within-subjects experiment to compare infants’ and parents’ dyadic engagement when using strollers and backpacks. Importantly, this study did not compare stroller users with backpack users, but employed a sample that regularly used both types of infant transport modes with roughly half preferring one mode to the other. One might have expected parents to be motivated to report that they and their infants were more engaged using the modality preferred by parents. However, this was not the case. Even parents who reported a preference for strollers perceived themselves and their infants as more engaged in backpacks, particularly with regard to vocalizing. These perceptions were corroborated by the audio and video data, which showed parents talk to their infants and infants vocalize to their parents, as well as appear to scan the visual field significantly more when using backpacks. Correlational analyses suggest that “chattier” infants did not have chattier parents; rather infants took more vocal initiative in backpacks, which resulted in more dyadic conversations.

This study was somewhat limited by a self-selected, middle-class, racially homogenous sample. Follow-up studies might include eye-tracking and/or kinesthetic data to examine if parents and infants share these specific elements in backpacks, as well as a physiological metric such as heart rate to objectively measure alertness. Longitudinal studies of attachment and language development as a function of the frequency of stroller and backpack use would also clarify the role of transport modalities in these outcomes. Despite the limitations of this study, backpacks appear to provide infants with pragmatic opportunities for language exposure at a time critical to language development, and to provide parents and infants with more occasions for shared social-emotional and pre-conversational experiences that may contribute to attachment (Anisfeld et al., 1990).