Brady Eisert and Sarah Coyne, School of Family Life

Introduction

People born in the United States (US) today will be surrounded by media for most of their lives. However, very little work has been done longitudinally to examine how this media saturation is related to development. The existing research typically examines parenting predictors after the child is born and already using media. However, prenatal attitudes and behavior are important in predicting many parenting practices including decisions around feeding (Gregory, et al., 2016; Martin-Anderson, 2013), sleeping (Nowkicki, et al., 2017), daycare (Powell & Karraker, 2017), and circumcision (Montano et al., 2018). Accordingly, it is likely that there are significant prenatal predictors of infant media use or exposure. Therefore, the need for longitudinal research investigating these this association is pressing.

The purpose of this study is to longitudinally examine the relationship between prenatal maternal attitudes surrounding the appropriateness of media use during parental activities (“technoference” attitudes) and actual practices during playtime. Our hypothesis is that prenatal technoference attitudes will predict higher actual technoference during play, over and above the effects of maternal age, baby age, maternal ethnicity, and maternal efficacy.

Methodology

Participants for this study came from the Project MEDIA dataset. Analyses included 165 pregnant women who completed several questionnaires regarding mental health, media use, attitudes toward child media use, and other behaviors (Wave 1). Participants also completed several questionnaires (e.g., child media use, maternal technoference) approximately one year after their child was born (Wave 2). There was a 99% retention rate between the two years. Babies averaged 3.50 months (SD = 1.91), and mothers averaged 31.18 years (SD = 4.80). The largest ethnic group was European American (78.3%), followed by Hispanic/Latino (10.2%), African American (4.8%), and Asian American (4.2%). Study variables are described below.

Play Technoference. Play technoference measured the self-reported likelihood of a parent to use a phone or other device during play with the child. It was assessed using a five-point Likert scale from 1 (I never do this) to 5 (very likely). A higher score is associated with more technoference.

Prenatal technoference attitudes. Mothers’ attitudes about technoference (the appropriateness of using media while engaging with a child) were assessed using a five-item measure. Participants were asked to imagine that their child was 1 year old and rate their agreement with statements regarding media use while their child is awake and present. Participants answered on a four-point Likert scale (1 = strongly disagree; 4 = strongly agree). A higher averaged score indicated greater technoference. Reliability for this measure was acceptable (α = .66).

Control Variables. Maternal and baby age were provided by participants. Maternal ethnicity was dichotomized based on being European American (1) or not (0) due to low variability (78.3% were European American). Maternal efficacy pertains to parents’ perceptions of their competence on basic skills required in caring for an infant. It was measured using an adapted version of the Parenting Sense of Competence Scale (Pedersen, Bryan, Huffman, & Del Carmen, 1989). The scale consisted of sixteen items on a seven-point Likert scale from 1 (not at all representative of me) to 7 (strongly representative of me). Items were averaged. Greater scores were indicative of greater feelings of efficacy. Reliability for this measure was good (α = .78).

Results

A multiple hierarchical regression was run to determine if mothers’ prenatal technoference attitudes significantly predicted play technoference (Model 2) over and above the effects of maternal age, baby age, maternal ethnicity, and maternal efficacy (Model 1). A check for multicollinearity revealed that some predictor variables had moderate intercorrelations and a high condition index. After testing and examination, maternal age and ethnicity were removed due to their associations with each other and prenatal technoference attitudes. A new multiple hierarchical regression was run with only baby age and maternal efficacy as controls in Model 1 and no change to Model 2. No issues with multicollinearity were found in the new model.

Results revealed that the control model (Model 1) was insignificant (R2 = .009, F(2,162) = .718, p = .489). The control variables did not significantly predict technoference during playtime. Model 2, which included our predictor variable, showed both overall significance (R2 = .124, F(3,161) = 7.618, p < .001) and a significant effect over and above that of Model 1 (R2Δ = .116, F(1,161) = 21.239, p < .001). Specifically, prenatal technoference attitudes significantly predicted later play technoference (β = .342, p < .001). Mothers with prenatal attitudes supporting higher technoference had more technoference later during playtime with their babies.

Discussion

Due to multicollinearity issues, our hypothesis could not be tested with all control variables. The results of this study, however, still found that prenatal technoference attitudes predicted higher playtime technoference over and above the effects of the baby’s age and maternal efficacy. This is an important finding because it suggests that prenatal maternal attitudes toward technoference may influence actual practices when the child is born despite the child’s age or parental efficacy. This implies that future technology-related parenting practices may be influenced if individuals change their attitudes surrounding technoference before children are born. This powerful relationship could be harnessed by therapists and educators to help parents create more healthy environments for their developing children.

Conclusion

Children in the US are being saturated with media, but research surrounding the effects of media on development is limited. While this study has produced results connecting prenatal attitudes with postnatal practices, future research is needed to investigate this connection in more depth. For example, more robust analyses could examine associations over a greater time period with added predictors and outcomes. As more research like this emerges, individuals and families could be helped for generations to come.