Evinced by a large cross-sectional survey, a recent publication by Modecki et al. (2020) asserts that ‘more [smart]phone use was associated with higher parenting quality’. This generalistic concluding statement contradicts an increasing corpus of research. The purpose of this commentary is to highlight some of this relevant corpus and provide some caution to their assertion.

Modecki and colleagues reported on self-reported phone use and data on relationship warmth elicited from a cross-sectional convenience sample conducted by the Australian Broadcasting Corporation in August 2017. However, it is widely recognised that considerable care needs to be exercised when relying on participants’ self-reported estimates of their phone use, as the psychometric and reliability properties of responses is generally low (Lee, Ahn, Nguyen, Choi, & Kim, 2017). Furthermore, the attachment scale used by the authors seeks to measure parent attachment to children but not vice versa and, indeed, is not a measure of parenting quality. This measure of parental warmth, also self-reported, was elicited using a single item of unknown validity. These self-report scales leave unanswered questions about actual parenting quality and practices, the child’s lived experience of parental warmth and the impact on child–parent attachment relationships. Caution may be prudent in interpreting these data; attachment relationships are two-sided phenomena and not straightforward to measure well. Further, one person’s expression of warmth may not be received as such; observational data would be helpful in interpreting the validity of a parent’s assessment of their own warmth.

Their reliance on a convenience sample also limits the external validity of the findings and is likely to introduce important (yet hidden) biases. In accordance with the STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) best practice reporting guidelines (www.strobe-statement.org), a useful addition to the paper would have been the inclusion and stratification of demographic information about the children whose parents were surveyed. Table S2 ‘Participant Demographics’ deals with information about the parents themselves, only alerting us to the range of children’s ages (0–18 years) and the mean age of the youngest child living at home (7.86 years). This omission of detail is crucial. Parenting is a role which varies substantially depending upon the chronological age, developmental stage and individual needs of those being parented. While Modecki and colleagues seek to describe the ‘real effect’ of smartphones on parenting, the age, stage and needs of the children studied remained largely unexplored. Thus, the implicit assumption made by these authors is that there is no effect modification between phone use and parenting behaviour over children’s differing age, stage and needs, but this needs thorough testing.

In addition to the lack of child age stratification, the poor psychometric properties of the primary variables and the questionable external validity of its findings, Modecki and colleagues’ study also likely suffers from residual confounding – where additional important confounding factors beyond parent’s age, relationship status, education, employment status and age of youngest child were not considered or collected. Furthermore, the authors use these cross-sectional data to assess whether ‘smartphone use predict[s] parenting’ employing associative techniques rather than, for example cross-validation methods which utilise both training and test data sets. Cross-validation avoids the optimistic estimates of predictive performance known to exist when the full data set is used for both model specification and prediction assertions. These methodological issues highlight the need for cautious interpretation of the data.

Modecki and colleagues conclude that ‘more [smart]phone use was associated with higher parenting quality’. Yet, parental distraction by smartphones has, for example, been shown to interfere with parental sensitivity (Beamish, Fisher, & Rowe, 2019), itself a precursor to a secure attachment relationship. Elsewhere, parental smartphone use has been linked to risks to child safety, child behaviour problems (Beamish et al., 2019; Newsham, Drouin, & McDaniel, 2018), a parent’s feeling of closeness to their children and fewer voiced interactions – with implications for children’s language learning (Newsham et al., 2018). These and other findings reinforce the need for parental phone use caution.

While Modecki and colleagues claim that ‘more smartphone use is associated with better (not worse) parenting’, it begs the question, better for whom? As previously suggested, ‘parenting’ is not a job description where one-size-fits-all. For example, the greatest concern often pertains to infants and young children, but the specific associations for this age group were not addressed in the paper. The neurobiological needs of infants are different from the needs of older children. Infancy is a time unlike any other, with rapid neurological growth and the setting of trajectories across a variety of developmental domains. For babies, parental smartphone use may be especially impactful, as their early relationships implicate brain growth, the complexity with which they learn language, their future relationships, their physical health, their educational outcomes and their development of morality.

As one mechanism for better understanding the impact of parental smartphone use on parent:infant relationships, Myruski et al. (2018) studied the parallels between the Still Face Paradigm (SFP) and contemporary parents’ use of smartphones in the presence of their infant. The original seminal SFP study (Tronick, Als, Adamson, Wise, & Brazelton, 1978) had mothers cease their usual responsive interactions and instead assume a ‘still face’ with blank affect. Their babies’ distress and eventual withdrawal was seen as evidence of their need for relational reciprocity and their awareness of interruptions to the natural flow of interactions.

In their modified version, Myruski and colleagues observed 50 mother and baby pairs (mean age 15.4 months, range: 7.2–23.6 months) and sought to mimic ‘typical disruptions in parent–infant interactions that may occur in daily life’ (p1). Mothers were tasked with interacting with their babies, then their smartphones, before returning their attention to their babies. The researchers found that while parents were using their phones, they were likely to assume a still face. They also reported that greater self-reported parental device use was associated with less positive affect from babies and a reduction in the successful repair of interactions following disruptions by smartphones. This is potentially problematic, as babies have been shown to be sensitive to disruptions in the flow of natural interactions (Bigelow & Best, 2013) and unpredictable parental signals are associated with negative cognitive outcomes and risk of mental illness for children. While the study by Myruski and colleagues benefited from observations of parent–infant dyads, this work was limited by a reliance on self-reported patterns of typical phone use. Further, this study varied from the original SFP study, as it used a shortened reunion phase, and had variations in the provision of toys and access to movement for the children.

However, their findings regarding the relationship between self-reported rates of parental phone use and the babies’ affect and reunion behaviours were reinforced in a subsequent study by Stockdale et al. (2020). Stockdale and colleagues sought to amend the variations in the SFP test conditions used in Myruski and colleagues’ study by, instead, duplicating the original model. While their work was still reliant on self-reported parental phone use, they found that ‘Parent beliefs regarding the appropriateness of using media while present with their child was a stronger predictor of infant behavior … than parental reports of their actual technoference behaviors’ (p18). This finding suggests that parental beliefs about the appropriateness of smartphone use have an influential role on their children, while Modecki and colleagues’ assertion that more smartphone use is associated with better (not worse) parenting suggests that smartphone use in the presence of their children is, indeed, appropriate. Promotion of Modecki and colleagues’ assertion that increased smartphone use leads to better parenting outcomes could contribute to parental behaviours that may be problematic in terms of the quality of their caregiving.

Conversely, recommendations that parents limit smartphone use in their children’s presence may not only serve babies but are likely to support parents themselves. Mothers’ overuse of smartphones has been found to contribute to deleterious outcomes such as increased self-reported loneliness (Mandai, Kaso, Takahashi, & Nakayama, 2018), or symptoms of depression (Newsham et al., 2018). In their work, Newsham et al. (2018) include a Public Policy Relevance Statement, in part to encourage parents to limit their own use of media when interacting with children. In addition to the risks to infants and their mothers, levels of parental media use are associated with children’s own levels of media use, later – which are themselves associated with potentially negative effects (Beamish et al., 2019), although the evidence on this remains controversial.

As Modecki and colleagues indicate, there is a need to ask nuanced questions about the impacts of our increasingly technologised world. Answering these nuanced questions and asserting an understanding of any real effect warrants objectively gathered data, high-quality research designs capable of identifying causal effects, and convergent findings from a range of sources of evidence.

We agree that addressing nuanced questions will likely require avoiding the ‘generalized narratives of family risk’ as cautioned by the authors, while acknowledging the potential risks identified in the current literature as well. These risks deserve closer scrutiny – especially to infants and the developmentally vulnerable – and, we believe, ought to be acknowledged. Without stronger research designs and more robust measurement, we do not believe it is warranted to conclude that mobile phone use improves parenting and call for more high-quality research on the potential risks and benefits.

The authors wish to acknowledge and sincerely thank the Joint Editor, Prof. Pasco Fearon, for his constructive and critical comments and review. The authors have declared that they have no competing or potential conflicts of interest.