Tapan Parikh, Ann & Robert H. Lurie Children’s Hospital of Chicago, United States
This is an open-access article distributed under the terms of the
Technoference, namely parental screen use in the presence of a child, is a widespread phenomenon that has negative effects on parent-child interaction and communication. When parents use screens around their children there are fewer interactions and parents are less contingent and responsive to the child. Additionally, children show more negative behaviors, such as whining, frustration, and outbursts. Communication is also affected—parents speak and gesture less towards their children and, in turn, children are less likely to develop their language abilities. It remains unclear, however, if parental distraction due to screen use affects parent-child interaction and communication more negatively compared to non-digital parental distraction. Fifty-two parent-child dyads (mean child age = 22 months, range 12–36 months) first played for 5 min (Time 1); then (Time 2), the parent was asked to fill out a questionnaire on a tablet (screen condition), on a printed form (paper-pen condition) or was not interrupted (control condition). Interactive quality was assessed at Time 1 and Time 2 using the Coding Interactive Behavior scale. Communication was assessed by coding the number of word tokens and types during Time 1 and Time 2; child gestures were also coded. Results revealed that when parents were distracted—either by the paper-pen or the screen questionnaire—the quality of the interaction significantly deteriorated (
香京julia种子在线播放
Young children rely on interactions with their parents in order to learn and develop their social and emotional skills. When parent-child interactions are disrupted—such as when the parent is being distracted by using a screen—the interaction is negatively affected. In the current study, we test how different parental distractions may disrupt the quality and quantity of interactions between parents and their toddlers. Specifically, we ask whether parental distraction from a screen activity leads to lower quality of the interaction and to less communicative exchanges between the parent and the child than parental distraction from the same activity on paper-pencil.
Digital devices are ubiquitous: people across all ages, all cultures and all socioeconomic backgrounds are using digital media on an everyday basis. For smartphones only, 82% of French families (
Yet, when a parent uses a digital device in the presence of a child, the nature and quality of the parent-child interaction are impacted (
Given that technoference is not an isolated phenomenon, studies have examined the implications of technoference on children. Beyond issues related to safety concerns (
Interactions between parents and infants play a crucial role in supporting various aspects of the child’s development (
A crucial aspect of the quality of parent-child interactions is the ability of parents to detect, acknowledge and respond to child's behavior and communicative bids [i.e., parental sensitivity (
Research shows that when parents use screens in the presence of their child, there are fewer interactions (
Parental screen use during parent-child interactions also affects the child. Technoference is associated with both more internalized behaviors, such as whining or pouting, and more externalized behaviors, such as agitation, frustration and outbursts (
In the first years of life, parent-child interactions provide the foundations of young children's communicative development. Stemming from the Transactional Theory in child developmental (
There is empirical evidence showing that technoference affects parental communicative input towards young children (
Existing evidence shows that technoference is a widespread phenomenon that affects young children's development. Most research, however, has studied parental distraction by a screen compared to situations when parents are not distracted. While these studies highlight issues regarding screen use during parent-child interactions, it is clear that parents are oftentimes distracted in many other ways during interactions with their child, such as attending to another sibling or finishing cooking a meal. In order to determine the effect of parental distraction from using a screen, it is crucial to provide evidence from experimental conditions when parents are distracted from a similar but non-digital activity. To our knowledge, the empirical evidence on the differential effect of technoference compared to other parental distractions is very scarce. In a U.S. study on question-asking during parent-child interactions, Gaudreau et al. (
Accordingly, in this study we ask whether technoference affects the interaction and communication skills of parents and young children above and beyond parental distraction from a paper-pen activity. Based on the literature, we made the following hypotheses:
Both interaction and communication will be negatively affected when parents are distracted compared to when parents are not being distracted. Both interaction and communication will be negatively affected when parents are being distracted in the screen condition compared to when parents are being distracted in the paper-pen condition, given that digital devices may be especially distracting to parents (
Fifty-two parent-child dyads were invited to participated in this study. Two dyads were excluded because of technical problems preventing from coding the data. The final sample consisted of 50 parent-child dyads; however, interaction scores are missing for one dyad due to disturbances during data collection. The G*Power software indicated that the obtained power for correctly rejecting the null hypothesis (1-β err prob) given an effect size of .25 and the study's sample size is .95.
Mean child age was 22 months (min = 12, max = 36, SD = 7.37 months); 26 girls (52%) participated. Mean parental age was 34 years (min = 27, max = 49, SD = 4.72 years); 45 mothers (90%) participated. Parents were predominantly married or in a couple (88%), mostly highly educated (67% held a university degree) and the large majority had a professional occupation (68%). Participants were recruited in the metropolitan area of a large French-speaking city in Switzerland by posting flyers in day-care centers and pediatric practices, by word-of-mouth, and through advertisement on social networks.
Data was collected in a quiet laboratory room over the summer of 2021. Upon arrival of the parent-child dyads, an experimenter explained that the purpose of the study is to evaluate parent-child interaction (without mentioning parental distraction), answered eventual questions and obtained written consent from parents. Each parent-child dyad was invited to sit on a foam tile carpet on the floor; several cushions were supplied. A seek-and-find book as well as a wooden bear puzzle were provided for the interactive play session (see
Materials used for the interactive play session.
All dyads were instructed to play for 10 min as they usually would do in a quiet room. Participating dyads were randomly assigned into one of three condition. In the first condition, parent-child dyads interacted for 10 min without distraction (“control group”, 16 dyads, or 32%). In a second condition (“paper-pen distraction”, 17 dyads, 34%), each dyad interacted for 5 min without distraction (Time 1), then was interrupted by the experimenter who asked the parent to fill in a demographic questionnaire on paper while continuing to interact with the child for additional 5 min (Time 2); the experimenter then left the room. In the third condition (“screen condition, 16 dyads, 32%), each dyad interacted for 5 min without distraction (Time 1), then was interrupted by the experimenter who asked the parent to fill in a demographic questionnaire on a digital tablet while continuing to interact with the child for additional 5 min (Time 2); the experimenter then left the room.
In all conditions, after ten minutes of interaction, the experimenter returned to the room and stopped the video recording. Parents were then invited to complete the demographic questionnaire either in the lab or at home. Each session lasted approximately 40 min. At the end of each session, participants were informed about the focus of the study on parental distraction and were given the possibility to retract their participation if they no longer agreed with the aims of the study; no participants retracted. All participants received a CHF20 gift voucher for a bookstore.
A 13-item demographic questionnaire contained questions about the child (e.g., age, nationality, child’s place in siblings and dominant childcare arrangement) and the parent (e.g., age, gender, education and marital status).
The interaction skills of the child, the parent and of the dyad taken together were assessed during the 10-minute observed play using the Coding Interactive Behavior scheme [CIB (
The child scale consists of items divided into three composite scores: social involvement, withdrawal/negative emotionality, and compliance; for the purposes of this study, we focus on the social involvement and withdrawal composite scores. The child social involvement composite score is coded on the following 9 items: child gaze/joint attention, positive affect, child affection to parent, alert, fatigue (reversed score), child vocalization/verbal output, child initiation, competent use of environment, and creative symbolic play. The child withdrawal composite score is coded on the following 4 items: negative emotionality/fussy, withdrawal, emotional lability, and child avoidance of parent.
The parent scale type consists of items divided into four composite scores: sensitivity, intrusiveness, limit setting, and negative mood; for the purposes of this study the two former composite scores are used. The parent sensitivity composite score is coded on the following 10 items: acknowledging child’s signals, elaborating on child's signals, parent gaze/joint attention, positive affect, vocal appropriateness/clarity, appropriate range of affect, resourcefulness, praising of the child, affectionate touch, and parent supportive presence. The parent intrusiveness composite score is coded on the following 4 items: forcing, overriding child's signals, parent anxiety, and criticizing the child.
The dyad scale consists of items divided into two composite scores. The dyadic reciprocity composite score is coded on the following 3 items: dyadic reciprocity, adaptation-regulation, and fluency. The negative state composite score is coded on the 2 following items: constriction and tension.
Each item is coded on a scale from 1 to 5, as follows: no manifestation of the item's behavior is observed (1), some manifestations are present but not frequent or constant during the interaction (3), manifestations of the item's behavior are frequent and constant throughout the interaction (5). Codes of 2 or 4 can be used to indicate a tendency towards a low (2) or high (4) level. Accordingly, higher scores indicated higher child social involvement, higher parental sensitivity, but also higher child withdrawal, and higher parental intrusiveness.
For the purposes of this study, the 10 min play session was divided into the first 5 min of the interaction (Time 1, no distraction) and the remaining 5 min of interaction (Time 2, no distraction for the control group, paper-pen distraction, or screen distraction). Each time (Time 1 and Time 2) was coded separately.
Coders were trained to use the CIB scheme by a licensed coder who obtained reliability with Ruth Feldman’s team. Once coders reached reliability with the licensed coder on a different video dataset, they coded the video data from the present study. A randomly selected 20% of the video data was double-coded; inter-rater agreement was 80%, indicating that both coders gave the same score or a score that differed by maximum one point at the 5-point Likert scale in 80% of the cases.
Verbal communication skills were assessed using the word tokens and word types produced by the child and the parent during the 10 min play interaction. Additionally, non-verbal communication skills were evaluated in children through the gestures they produced during the 10 min interaction.
Word tokens and types scores were based on the transcription of the speech produced by the child and the parent. Speech was transcribed following an adaptation of transcribing conventions from Hoff (
We further coded for children's nonverbal communication skills during the parent-child interaction, following earlier work (
The dependent variables in our analyses were the following: score for child social involvement during the interaction, score for child withdrawal during the interaction, total number of gestures produced by the child, total number of word tokens produced by the child, total number of word types produced by the child, score for parental sensitivity during the interaction, score for parental intrusiveness during the interaction, total number of word tokens produced by the parent, total number of word types produced by the parent, score for dyadic reciprocity during interaction, and score for dyadic negative states during interaction; all dependent variables showed a non-normal distribution. In order to answer our first research question, namely to determine whether parental distraction—independently of whether it is a paper-pen distraction or a screen distraction—alters parent-child interaction and communication, we performed a Wilcoxon signed-rank on the dependent variables between Time 1 and Time 2 for the paper-pen and screen conditions taken together. In order to determine whether parental distraction from filling out a questionnaire on a
Comparing parent-child interaction and communication between Time 1 (i.e., no distraction) and Time 2 (paper-pen or screen parental distraction) shows that parental distraction matters (see
Descriptive statistics (means; standard deviation scores in parentheses) for the dependent variables across the two assessment times (Time 1, Time 2) for the paper-pen and screen conditions taken together. The last column displays the
Time 1 | Time 2 | |||
---|---|---|---|---|
Interaction | Parental sensitivity | 4.46 (.29) | 3.76 (.55) | .001*** |
Parental intrusiveness | 1.44 (.45) | 1.41 (.53) | .532 | |
Child social involvement | 3.85 (.42) | 3.43 (.52) | .001*** | |
Child withdrawal | 1.28 (.58) | 1.51 (.67) | .01** | |
Dyadic reciprocity | 4.46 (.70) | 3.21 (.84) | .001*** | |
Dyadic negative states | 1.34 (.54) | 1.98 (.44) | .001*** | |
Communication | Child word tokens | 44.53 (37.60) | 41.74 (41.63) | .350 |
Child word types | 21.50 (17.43) | 21.06 (18.49) | .610 | |
Child gestures | 9.06 (8.02) | 7.63 (6.95) | .299 | |
Parent word tokens | 389.82 (149.02) | 278.49 (140.63) | .001*** | |
Parent word types | 152.70 (48.36) | 129.83 (48.63) | .012* |
*
Regarding the quality of the parent-child interaction, our results showed that all of the examined variables, except for parental intrusiveness, differed between Time 1 and 2. Specifically, when parents were asked to fill out a questionnaire at Time 2 (either on a paper-pen or screen format), the quality of the interaction was significantly altered compared to Time 1 when they were only instructed to interact with the child: parents were less sensitive to their children's communicative signals and needs (
In terms of communication scores, however, results showed that only parental speech changed between Time 1 and Time 2, such that parents talked less to their children—both in terms of word quantity (
Taken together, the analyses showed that when parents were distracted by filling out the questionnaire—independently on whether it was on a paper-pen or a screen format—the quality of the interaction as well as the child-addressed parental speech deteriorated.
Having established that parental distraction alters parent-child interaction and communication, we asked whether the nature of the distraction—digital vs. non-digital—matters as well.
Results showed that, in terms of interaction quality, parental sensitivity,
Descriptive statistics (means; standard deviation scores in parentheses) for the dependent variables at Time 2 by condition. The last column displays the
No distraction | Paper-pen distraction | Screen distraction | |||
---|---|---|---|---|---|
Interaction | Parental sensitivity | 4.46 (.33) | 3.77 (.51) | 3.77 (.60) | .001*** |
Parental intrusiveness | 1.56 (.60) | 1.35 (.44) | 1.51 (.62) | .733 | |
Child social involvement | 3.78 (.27) | 3.37 (.60) | 3.56 (.35) | .041* | |
Child withdrawal | 1.28 (.36) | 1.37 (.49) | 1.67 (.82) | .465 | |
Dyadic reciprocity | 4.38 (.73) | 3.28 (.79) | 3.29 (.88) | .001*** | |
Dyadic negative states | 1.45 (.53) | 2.01 (.39) | 1.95 (.52) | .001** | |
Communication | Child word tokens | 50.44 (56.74) | 34.94 (29.63) | 45.82 (51.03) | .912 |
Child word types | 26.44 (24.40) | 19.35 (16.57) | 20.82 (19.52) | .753 | |
Child gestures | 9.63 (8.55) | 8.53 (8.04) | 7.12 (5.83) | .805 | |
Parent word tokens | 368.75 (116.92) | 267.23 (140.02) | 287.82 (148.76) | .054† | |
Parent word types | 148.25 (34.21) | 131.06 (50.30) | 125.76 (48.29) | .259 |
†
In this study we asked whether technoference affects the interaction and the communication between parents and their children during a 10 min play. Observing 50 parent-child dyads, we found that parental distraction matters for the quality of interaction and for the quantity of communicative acts, independently on whether parents were distracted by a paper-pen questionnaire or by a questionnaire on a screen.
Our first main result shows that when parental focus of attention is shifted from the child to another activity, the quality of interaction is negatively impacted. Specifically, when parents were distracted during the parent-child play, parents were less sensitive to children's communicative signals, children showed lower social involvement, and the dyads showed less reciprocity and more negative states in their exchanges. These findings align with previous theoretical and empirical work highlighting that children, especially young children, need the attention of their parents during moments of interaction and play. When parents pay attention to their children, they acknowledge and respond to the child's behavior and needs, which, in turn, contributes to the child's early social, communicative and emotional development (
Importantly, our study suggests that the fact that parents are not focused on the interaction with their child has a negative effect not only on themselves (i.e., parents being less sensitive to the child), but also on the child
Our findings showed that parental distraction also affects the communicative exchanges between the parent and the child during the 10 min play. Specifically, parents produced less word tokens (i.e., measure of verbal quantity) and less word types (i.e., measure of verbal variability) in the conditions when parents were distracted. This result confirms our hypothesis according to which parental distraction negatively influences the communicative bids between the parent and the child. Interestingly, we did not find differences in the verbal or non-verbal communication that children were addressing to parents across the undistracted (Time 1) vs. distracted condition (Time 2). It is possible that when parents were distracted, children continued to produce communicative bids in order to regain the attention of their parent, thus maintaining a similar level of communication. This hypothesis remains to be further examined in future studies.
Our second main result is that technoference, namely parental distraction due to using a digital device, does not affect parent-child interaction, nor communication more so than non-digital parental distraction. More specifically, we failed to found a difference between parental distraction due to using a screen compared to parental distraction due to a non-digital activity (i.e., paper-pen condition). This result is in conflict with our hypothesis—we expected that technoference will show a more pronounced negative effect on parent-child interaction quality and communication compared to the non-digital distraction, given that screens are especially distracting to parents (
A first explanation lies within the existing evidence. While the vast majority of studies on technoference point to negative effects, including on parent-child interaction and communication, these studies are either qualitative, thus failing to provide comparisons from experimental conditions, or compare conditions of parents using screens (i.e., technoference) to conditions of parents not using screens [i.e., paying attention to the child; for a review, see (
A major strength of the present study is to provide comparisons of parent-child interaction and communication across three experimental conditions: undistracted parents, parents distracted by a non-digital activity (i.e., completing a paper-pen questionnaire) and parents distracted by an activity on a screen (i.e., technoference). Such comparisons suggest that parents are not more distracted by a screen than by another non-digital activity. Recent literature providing evidence from similar comparisons show similar findings. For example, comparing parent-child interactions while parents used a cell phone to parent-child interactions while parents completed a paper survey, Gaudreau et al. (
Taken together, our results add to the existing evidence showing that parents do not seem to be more distracted when using a screen compared to other types of distraction. This might be so because screens are so ubiquitous in today's society, that both parents and children have become accustomed to such devices. Stockdale et al. (
It is also likely that digital devices prompt a joint attention phenomenon; namely, children could find screens attractive, thus increasing the likelihood of joining in the parent's focus of attention drawn from the screen—possibly more so than when parents are involved in a non-screen activity. There is evidence that when parents and children are co-viewing and especially when parent use this co-viewing as an opportunity for interaction, conversation and sharing (
While the present cross-sectional study adds important findings to the existing literature, it also presents a number of limitations. First, the sample represents a limitation in terms of size and representativeness. It includes 50 dyads (considered a high number for observational studies) and is relatively homogeneous in terms of parent gender (vast majority of mothers), parent education (highly educated parents), and marital status (over-representation of married parents/couples). Second, ecological validity is affected by the experimental setting, which is not representative of everyday parent-child interactions in a natural environment. This may result in a motivational bias with regard to participation, with parents in difficulty with their child avoiding taking part in the study. This may also manifest itself as a social desirability bias through the desire to satisfy social expectations in terms of child rearing. Importantly, the type of parental digital distraction in this study (i.e., completing a questionnaire on a tablet) is different in many ways from the real world, everyday ways in which parents use screens in the presence of their children. Specifically, in our study, parental digital distraction did not include the personal or professional context of screen use, neither any emotional aspect in the nature of the digital distractor, such as for example when consulting work emails or responding to personal messages. In such naturalistic situations of screen use, parents are likely more compelled to use screens and presumably they are more absorbed in the screen use and distracted from the interaction with the child.
Last, parental gestures were not coded due to personnel shortage. It would have been interesting to have the coding of parents’ gestures in order to better understand certain effects observed in the interaction. In fact, several authors (
Drawing from these limitations, future studies are needed in order to provide better understanding of the effect parental use of screens might or might not have on young children. Importantly, longitudinal studies controlling for a number of significant confounds are truly needed in order to capture possible causal links. These include, but are not limited to, the level of immersion (
This study suggests that technoference does disrupt parent-child interaction, but in similar ways than an equivalent non-digital parental distraction. The finding adds to the extensive literature on the importance of parental involvement for the quality of the parent-child interaction. It also allows to dedramatize a certain “moral panic” (
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
The studies involving humans were approved by Commission interdisciplinaire d'Ethique de la Faculté des sciences sociales et politiques de l'Université de Lausanne (E_SSP_042021_00002). The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation in this study was provided by the participants’ legal guardians/next of kin.
SC: Formal Analysis, Investigation, Methodology, Writing – original draft. AF: Formal Analysis, Investigation, Methodology, Writing – original draft. NM: Formal Analysis, Investigation, Methodology, Writing – original draft. FQ: Formal Analysis, Investigation, Methodology, Writing – original draft. ND: Formal Analysis, Investigation, Methodology, Writing – original draft, Conceptualization, Funding acquisition, Project administration, Resources, Supervision.
The author(s) declare financial support was received for the research, authorship, and/or publication of this article.
This study received partial funding from the University of Applied Sciences of Western Switzerland (HES-SO) as well as from Foundation Action Innocence (Switzerland).
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.