The Relationship Between Media Multitasking Behavior and Executive Function in Adolescence: A Replication Study

Abstract

The proliferation of digital gadgets and increased media consumption among adolescents has raised interest and concern regarding possible effects on cognitive functioning. Research investigating this relationship has yielded mixed results. This study aims to replicate the research conducted by Baumgartner et al. concerning the relationship between adolescent media multitasking and executive functioning on a sample of 296 Romanian early adolescents. The same methodology as the original study was followed and its findings were partially replicated. Mainly, results of regression analyses indicated that more media multitasking with media activities was related to more self-reported executive difficulties, but not with performance on three computerized executive functioning tasks. Media multitasking with non-media activities was unrelated to executive functioning. When comparing extreme groups, however, heavier media multitasking was associated with faster performance on the task-shifting and inhibition computerized tasks. Implications for cognitive development and research methodology are discussed.

Keywords

media and technology cognitive development time use/leisure activities self-regulation

Technological advance in the last 20 years has brought a series of changes in the way people live their day-to-day lives—from smartphones, to smart TVs, to smart cars and smart houses, technology promises to make all things faster and easier than any time before. However, modern media and gadgets facilitate high rates of media use (MU) and media multitasking (MM)—two aspects of media behavior that have raised interest, but also concern, among researchers and professionals in education and psychology. MU refers to the frequency or the time spent by individuals engaging in media activities, such as watching TV or playing video games (Baumgartner, Weeda, van der Heijden, & Huizinga, 2014). Multitasking refers to the simultaneous engagement in two or more activities (e.g., Redick et al., 2016). We speak of MM when at least one of these activities involves a media device or content (Brasel & Gips, 2011; Ophir, Nass, & Wagner, 2009). At least two types of MM have been identified: MM with media activities, when one engages in multiple media activities at the same time (e.g., browsing Facebook while watching TV), and MM with non-media activities, when one engages in a combination of media and non-media activities (e.g., watching entertainment videos while doing homework; Wallis, 2010).

Studies that aim at disentangling MM effects approach this behavior in multiple ways. While some studies analyze MM as a continuous variable (e.g., Baumgartner et al., 2014; İmren & Tekman, 2019), most authors distinguish between heavy media multitaskers (HMMs)—those who report high levels of engagement in MM, and light media multitaskers (LMMs)—those who report low levels of engagement in MM. In this extreme-group approach, participants are divided into these groups by selecting those who report the highest and lowest levels of MM on individual measures of general MM (e.g., Media Multitasking Index; Ophir et al., 2009) by using different cutoff methods (e.g., upper and lower quartiles—Alzahabi & Becker, 2013; upper and lower deciles—Baumgartner et al., 2014; +/–1 SD—Ophir et al., 2009).

Media-Related Behavior in Adolescence

The “Media Use in The European Union Report” (TNS Opinion and Social, 2018) revealed that 84% of EU citizens watch TV (almost) every day. In addition, 65% of them use the internet at least once a day and 42% use social media platforms daily. These trends are on an ascending slope, having increased with 1% to 4% over the course of a year (TNS Opinion and Social, 2018). In Romania, 72.4% of the general population had internet access at home in 2018 (Institutul Național de Statistică, 2018). For Romanian children and adolescents (9–16 years old), Velicu, Balea, and Barbovschi (2019) show that internet access from mobile devices has increased four times since 2013 (21% in 2013 vs. 84% in 2018). Online activities have mainly entertainment and communication purposes in this age group (e.g., 74% of 9–16-year-olds use online socializing networks; Velicu et al., 2019). In the United States, the same age group reported almost 8 hours of MU every day and 29% of that time was spent doing more than one media activity at a time (Rideout, Foehr, & Roberts, 2010). Rosen, Carrier, and Cheever (2013) have also shown that adolescents tend to switch very often between media—about every 6 minutes. Due to these habits, children and adolescents are exposed to double the media content in half the time (Rideout et al., 2010).

In this context, it is important to consider that the beginning of adolescence is marked by increased neural development and organization (Galván, 2014), especially in the prefrontal cortex (PFC) along with other brain regions (Casey, Getz, & Galván, 2008; Crone & Steinbeis, 2017). This goes hand in hand with the development of executive functions (EF), which also follow a protracted developmental course well into young adulthood (Huizinga, Dolan, & van der Molen, 2006). Moreover, recent studies have shown that the brain’s functional connectivity individualizes in response to environmental input, a process that peaks around 14 years of age (Galván, 2017; Kaufmann et al., 2017). Therefore, the increased neural plasticity that characterizes the PFC during adolescence renders it more susceptible to environmental influences (Galván, 2014). In this context, concerns are raised that engaging so much with media and technology during this time might alter the way in which the brain is wired and may affect its functioning (Gardner & Davis, 2013; Wilmer, Sherman, & Chein, 2017). Specifically, this might translate into EF difficulties, poor academic performance, and other adaptive problems in this younger, more vulnerable group (Carr, 2010; Gardner & Davis, 2013). As will be illustrated further, studies that target media-related behaviors in both adult and younger samples show that, indeed, MU and MM are associated with both short-term (e.g., Lillard, Drell, Richey, Boguszewski, & Smith, 2015; Maass, Klöpper, Michel, & Lohaus, 2011) and longitudinal effects on cognition (e.g., Baumgartner, van der Schuur, Lemmens, & te Poel, 2017) and changes in brain structure (e.g., Hutton, Dudley, Horowitz-Kraus, DeWitt, & Holland, 2020).

Media-Related Behavior and Its Relationship With EF

The concept of EF is widely used as an umbrella term to refer to a series of cognitive processes, such as inhibition, working memory (WM) and shifting (Miyake & Friedman, 2012). They are interrelated, but independent processes, and are relevant for guiding goal-directed behavior (Barkley, 2012; Friedman et al., 2008; Huizinga et al., 2006; Miyake & Friedman, 2012). EFs have been found to play a major role in the successful regulation of emotions (e.g., Zelazo & Cunningham, 2007), cognition (Del Missier, Mäntylä, & Bruine de Bruin, 2010; Schmeichel, 2007; Wiley & Jarosz, 2012), and behavior (e.g., Gestsdottir & Lerner, 2008; E. K. Miller & Cohen, 2001). Because of their essential role in overriding highly automated or habitual responses, these processes are also strongly associated with academic outcomes (Best, Miller, & Jones, 2009; St. Clair-Thompson & Gathercole, 2006), job performance (M. Miller, Nevado-Montenegro, & Hinshaw, 2012), and health (Williams & Thayer, 2009). EFs have also been shown to be highly involved in the ability to multitask (Courage, Bakhtiar, Fitzpatrick, Kenny, & Brandeau, 2015; Himi, Bühner, Schwaighofer, Klapetek, & Hilbert, 2019; Pollard & Courage, 2017; Redick et al., 2016; Szumowska & Kossowska, 2016) and are consistently related to MM (e.g., Alzahabi & Becker, 2013; Cain & Mitroff, 2011; filtering—Cardoso-Leite et al., 2016; inhibition and shifting—Ophir et al., 2009; Ralph & Smilek, 2017; attentional lapses—Ralph, Thomson, Cheyne, & Smilek, 2014; WM—Uncapher, Thieu, & Wagner, 2016).

Among the first studies to investigate MM and its relation to EF in young adults was conducted by Ophir et al. (2009). Results showed that HMMs had difficulty suppressing irrelevant external stimuli (inhibitory difficulty) and internal information (WM difficulty), as compared with LMMs. HMMs also had worse performance on task-shifting tests, as indicated by higher switch costs and slower overall reaction times than LMMs. This was interpreted as indicating that HMMs might have a “breadth-biased cognitive control”—a tendency to process and mentally represent more information from the environment than is necessary for the task at hand and difficulty restraining attention on a single task. As a result, they are more likely to be subjected to distraction and interference during task performance (Ophir et al., 2009). Since this study was published, several conceptual and direct replications have been carried out on adult samples. Some results support initial findings (e.g., Cardoso-Leite et al., 2016; Ralph, Thomson, Seli, Carriere, & Smilek, 2015; Uncapher et al., 2016; Wiradhany & Nieuwenstein, 2017), others contradict them (e.g., Alzahabi & Becker, 2013; Minear, Brasher, McCurdy, Lewis, & Younggren, 2013; Ralph et al., 2015). Both positive (e.g., Lui & Wong, 2012) and negative effects of MM (e.g., Adler & Benbunan-Fich, 2015; Ralph et al., 2014; Wood et al., 2012) have been reported, showing a mixed pattern of results.

Cross-sectional studies on younger populations mirror the findings in adult samples. In a theoretical review, van der Schuur, Baumgartner, Sumter, and Valkenburg (2015) indicated that young HMMs report more difficulties filtering out irrelevant information and have variable performance in task-shifting tests. Some studies indicate poorer performance for HMMs on complex WM tasks (Ralph & Smilek, 2017; Sanbonmatsu, Strayer, Medeiros-Ward, & Watson, 2013) while others find equivalent WM performance for HMMs and LMMs (Baumgartner et al., 2014; Minear et al., 2013). Baumgartner et al. (2017) have also shown a cross-sectional positive relationship between MM and attention problems in adolescence, irrespective of age, but stronger for girls.

Longitudinal studies on MM and MU tend to point more consistently toward negative effects on cognitive functioning, especially for younger age-groups (e.g., Baumgartner et al., 2017; Johnson, Cohen, Kasen, & Brook, 2007; Swing, Gentile, Anderson, & Walsh, 2010). For example, HMM predicted more attentional problems 3 months later for early adolescents, but not for middle adolescents (Baumgartner et al., 2017). Results concerning MU show that excessive media exposure (e.g., more than 3 hours a day of TV viewing, playing video games, etc.) is consistently associated with ensuing attentional problems in middle childhood, as well as in middle and late adolescence (Johnson et al., 2007; Swing et al., 2010). In the clinical domain, a recent meta-analysis of 45 studies indicated that different media measures are positively related to attentional difficulties, impulsivity and combined manifestations of attention deficit hyperactivity disorder (ADHD), with a stronger relation for boys (Nikkelen, Valkenburg, Huizinga, & Bushman, 2014).

Experimental studies that targeted short-term effects of MU on EF show that children have a lower performance on WM tasks immediately after being exposed to fantastical media content, as opposed to more realistic programs (Lillard et al., 2015). Young adults also show a diminished capacity to concentrate after watching high-arousing (vs. slow) media contents (Maass et al., 2011). These results indicate that MU involving complex, dynamic contents induces immediate changes in cognitive functioning not only in young children, who are crossing a time of increased neural development and plasticity (Poole, Nunez, & Warren, 2007) but also in adults, whose neural systems have already reached maturity (Poole et al., 2007). This points toward a more causal effect of MU on EF, at least on the short term.

Neuroscience data from younger samples add to this line of evidence by showing that exceeding MU beyond the recommendations of the American Academy of Pediatrics (Council on Communications and Media, 2016) is associated with “lower microstructural integrity of brain white matter tracts” involved in language, EF, and emergent alphabetization skills in 3 to 5 year-olds (Hutton et al., 2020, pp. E1, E8). Although correlational, these findings might indicate that excessive MU has visible effects on brain structure in early childhood, which may influence cognitive functioning later in development. In adolescent and young adult samples, data coming from fMRI studies indicate more activation in prefrontal areas associated with effort-dependent cognitive control during an attentional task in the presence of distractors (Moisala et al., 2016). This activation is accompanied by a decrease in task performance and is positively related to MM. In other words, it seems that individuals who engage more in MM exert more executive control when confronted with distracters during a task in order to redirect attention from irrelevant information and respond to task demands (Moisala et al., 2016). This lends credence to results showing less efficient interference control for HMMs (e.g., Ophir et al., 2009).

One of the most promising studies to investigate the MM—EF relationship in 11-to-15-year-old adolescents is Baumgartner et al. (2014). Participants in this study completed self-reported and computerized measures of WM, inhibition and task-switching and reported habitual MM during media and non-media activities. Higher levels of general MM were related to more self-reported EF difficulties, but not with performance on computerized tasks. Although nonsignificant, two trends indicated smaller WM capacities for higher levels of MM with non-media activities and better interference control for HMMs. Therefore, results indicated that adolescent HMMs and LMMs do not seem to differ significantly on EF computerized measures. This seems partly inconsistent with results in other adolescent (e.g., van der Schuur et al., 2015) and adult research (e.g., Ophir et al., 2009). However, it is important to consider the fact that self-reported measures and computerized performance tasks tend to correlate weakly (Bodnar, Prahme, Cutting, Denckla, & Mahone, 2007). Also, the former can be subjected to individual bias (Paulhus & Vazire, 2007), while performance on the latter might be dependent on contextual factors (e.g., Elton & Gao, 2014; Simmonds, Pekar, & Mostofsky, 2008). All things considered, these results were found in a large sample of adolescents and using a continuous MM variable. They also held in an extreme group comparison between HMMs and LMMs, as is often done in MM studies (e.g., Alzahabi & Becker, 2013; Ophir et al., 2009).

The Present Study

The findings in adult and adolescent samples that are described above shed light on the complex relationship between MM and cognitive functioning, in particular, EF. However, even if some study results do agree on certain EF difficulties that are associated with MM (e.g., Cardoso-Leite et al., 2016; Ophir et al., 2009), others diverge significantly (e.g., Ophir et al., 2009 vs. Alzahabi & Becker, 2013 vs. Minear et al., 2013). These mixed findings might come as a result of methodological differences, such as using different tasks for measuring the same underlying executive process (e.g., Cardoso-Leite et al., 2016, vs. Minear et al., 2013), or might be a result of changes in MU and MM over time, as reports of MU in both Europe (TNS Opinion and Social, 2018) and the United States (Hitlin, 2018) indicate increases and changes in these variables over the last years. For example, heightened exposure and interaction with media and technology might generate training effects (see Alzahabi & Becker, 2013, for a discussion) and therefore, different relationships between MM and EF might be visible in current studies, as opposed to earlier ones. Also, certain individual variables have been shown to be related to the magnitude of the relationship of interest, such as age and sex (e.g., Baumgartner et al., 2017; Nikkelen et al., 2014). These must be taken into consideration to better explain the nature of the association between MM and EF. In contexts marked by mixed and inconsistent results, replication studies can be a useful method to strengthen findings that describe naturally occurring relationships and distinguish them from methodological artifacts (Makel, Plucker, & Hegarty, 2012). Therefore, the objective of the present research was to replicate the study carried out by Baumgartner et al. (2014) on adolescent MM and EF.

Method

Participants

Participants were middle school children from two urban public schools in Romania. Principals agreed to be a part of the study and informed consent was obtained from parents to enroll students in the research. A total of 321 parents offered their positive consent, out of 690 who were approached. Participants first completed a series of questionnaires in the classroom and were afterwards invited to complete three computerized EF tasks individually. A few children (n = 25) were absent from the first step and were, thus, excluded from the study. The final sample consisted of 296 young adolescents (166 girls, 56.1%; 130 boys, 43.9%) aged between 10 years 5 months and 15 years 2 months ( $\bar{X}$ = 12.9, SD = 1.04), who offered complete questionnaire data. Out of these, 179 adolescents also completed the three computerized tasks (see “Procedure”). Children who only completed the questionnaires were kept in the sample due to the availability of data on MM and self-reported EF.

Instruments

MM behavior and MU

The Romanian translation of the Media Use Questionnaire (MUQ; Baumgartner et al., 2014; Ophir et al., 2009) was used to measure MU and MM behavior. The MUQ targets nine types of media activities: (a) television viewing (tv), (b) reading, (c) listening to music, (d) having a phone conversation, (e) text messaging (e.g., SMS or Messenger etc.), (f) using social media (e.g., Instagram etc.), (g) watching movies/series on the computer, (h) other computer activities (e.g., internet surfing), (i) playing video games. Following Baumgartner et al. (2014), MM while engaged in five non-media activities were also included—(a) doing homework, (b) eating, (c) class activities, (d) using transportation, (d) face to face conversation. In the first section of the questionnaire (MU), participants indicated how much time they engaged in each of the nine media activities on an average day (1 [not at all]–6 [3 hours or more]). In the second section (MM), for each of the nine media activities, adolescents indicated how often they engaged in the eight other media activities at the same time (1 [never] to 4 [very often]; e.g., “While listening to music, how often do you engage in each of the following activities at the same time?—1. Watch TV, 2. Read, . . . ”). For non-media activities, participants indicated how often they engaged in each of these five activities at the same time with each of the nine media activities listed above (1–4 Likert-type scale). Following Baumgartner et al. (2014), the Media Multitasking Index for media activities (MMI-M, Cronbach’s α = .93) was the average of nine sub-scores obtained by averaging the MM responses for each of the nine main media activities separately. The individual sub-scores were moderately correlated (r = .35—.77, p < .000). MMI-M reflects the average MM behavior across media activities. Higher scores indicate a higher frequency of MM behavior across a variety of media activities. The same procedure as above was utilized with the five non-media activities to create a Media Multitasking Index for non-media activities (MMI—NM, Cronbach’s α = .84). This score indicates average MM across non-media activities—the higher the score, the more the participant engages in MM during non-media activities. Table 1 presents means and standard deviations for MM during the nine media and five non-media activities sub-scores and the final MMI-M and MMI-NM indices.

Table 1.

Means and Standard Deviations For Sub-Scores of Media Multitasking With Media and Non-Media Activities and Overall Indices.

Media Multitasking Activities	$\bar{X}$	SD
Media activities
Watching TV	2.0152	0.63167
Reading	1.5382	0.62383
Music	1.7953	0.55419
Phone conversation	1.6788	0.56214
Text messages	1.8130	0.54583
Social media	1.7327	0.56424
Computer videos	1.4170	0.46786
Computer other	1.6227	0.52294
Video games	1.5890	0.58357
MMI-M	1.6453	0.43281
Non-media activities
Homework	1.5745	0.54877
Eating	1.8295	0.62117
School Activities	1.4485	0.58747
Transportation	1.8491	0.53177
Face to face conversation	1.3190	0.46137
MMI-NM	1.5890	0.42084

Note. Higher values indicate higher frequency of media multitasking. MMI-NM = Media Multitasking Index–non-media activities; MMI-M = Media Multitasking Index–media activities.

EF. Self-report

The Romanian version of the Behavior Rating Inventory of Executive Function—Self Report (BRIEF; Guy, Isquith, & Gioia, 2004) was used to measure self-reported EF difficulties. The BRIEF consists of 80 items, grouped into eight subscales. Only three of these subscales were used in the present study: Inhibition, Shifting, and Working Memory. Adolescents indicated how often they encountered specific problems in day-to-day tasks and behaviors that are related to the three EFs (1—never, 2—sometimes, 3—often). The Inhibition subscale consists of 13 items, tapping into difficulties controlling one’s impulses ( $\bar{X}$ = 1.64, SD = 0.35, Cronbach’s α = .79). The WM subscale consists of 12 items, tapping into difficulties keeping and updating information in mind ( $\bar{X}$ = 1.65, SD = 0.35, Cronbach’s α = .79). The Shifting subscale consists of 10 items—including both behavioral and cognitive shifting ( $\bar{X}$ = 1.62, SD = 0.34, Cronbach’s α = .73). Higher scores on these subscales indicate greater difficulties in EF. Descriptive statistics for the scores on all subscales can be found in Table 2. Correlations with other measures can be found in Table 3.

Table 2.

Means and Standard Deviations for Outcomes of Self-Reported and Computerized Measures of Executive Functioning.

EF Measures Outcomes	M	SD	Skewness
BRIEF—working memory	1.65 (1.56)	0.35 (0.36)	0.61
BRIEF—inhibition	1.64 (1.53)	0.34 (0.35)	0.43
BRIEF—shifting	1.62 (1.48)	0.34 (0.35)	0.53
Flanker—congruent trials	548.85 (435.77)	125.20 (62.16)	1.26
Flanker—incongruent trials	607.75 (489.98)	163.39 (68.28)	1.40
Flanker—ratio score	1.10 (1.13)	0.10 (0.06)	0.84
Digit Span—forward	8.08 (7.81)	1.871 (2.08)	0.08
Digit Span—backward	5.97 (6.37)	2.261 (2.59)	−0.34
Digit Span—total score	14.06 (14.18)	3.409 (3.78)	−0.07
Dots-Triangles—repeat trials	1,011.83 (772.66)	231.42 (242.91)	0.93
Dots-Triangles—switch trials	1,294.13 (994.54)	318.98 (337.91)	0.38
Dots-Triangles—ratio score	1.28 (1.29)	0.20 (0.18)	0.29

Note. Values in brackets represent performance indicators in the Baumgartner et al. (2014) study; BRIEF = Behavior Rating Inventory of Executive Function—self-report: higher values indicate more EF difficulties; Flanker congruent/incongruent trials and Dots-Triangles switch/repeat trials: higher values indicate higher RTs; Flanker—ratio score: higher values indicate more cognitive interference, therefore lower performance; Dots-Triangles—ratio score: higher values indicate higher switch costs, therefore lower performance; Digit Span—higher values indicate higher working memory capacity, therefore, better performance. EF = executive functions; RT = response time.

Table 3.

Correlations Between Main Independent, Dependent and Control Variables.

Variables included in analyses	Sex	Age	MU	MMI-M	MMI-NM	BRIEF WM	BRIEF I	BRIEF S	DST	EFT
Age	.02
MU	−.03	.24**
MMI-M	−.05	.24**	.63**
MMI-NM	−.15*	.18**	.50**	.81**
BRIEF working memory (WM)	−.10	.08	.13*	.25**	.18**
BRIEF inhibition (I)	−.11	−.003	.17**	.28**	.18**	.60**
BRIEF shifting (S)	−.09	.09	.01	.12	.09	.59**	.54**
DST	−.08	.12	−.12	−.17	−.16	−.04	−.01	.02
EFT	.17*	−.16	−.01	−.07	.08	−.23**	−.10	−.12	−.17*
Dots-Triangles task (DTT)	−.05	−.10	−.05	.08	.02	−.07	.06	−.06	−.03	.05

Note. Higher scores on the BRIEF subscales indicate more EF difficulties. Higher scores on the Digit Span task indicate better WM capacity. Higher scores on the Eriksen Flankers task and Dots-Triangles task indicate more EF difficulties. MU = media use; MMI-M = Media Multitasking Index for media activities; MMI-NM = Media Multitasking Index for non-media activities; BRIEF = Behavior Rating Inventory of Executive Function—self-report; DST = Digit Span Task; EFT = Eriksen Flankers Task.

p < .05. **p < .01.

EF. Computerized tasks

All tasks were constructed using OpenSesame 3.2.5 and ran on individual laptops for each participant. Descriptive statistics can be found in Table 2.

Inhibition

The Eriksen Flankers Task was used to measure inhibitory control (Ridderinkhof & van der Molen, 1995). Sequential linear displays of five arrows that point either to the left or to the right were presented on the screen. Participants had to indicate whether the middle arrow (i.e., the target) pointed to the right or to the left by using a right or left response button (M—right, Z—left). The target was flanked by four other congruent or incongruent distractor arrows, pointing in the same or opposite direction as the target. The arrays remained on the screen until the participant responded or 2,500 milliseconds had passed. As opposed to Baumgartner et al. (2014), where participants completed 50 practice trials and 50 experimental trials, participants in the present sample only had eight practice trials, followed by 50 experimental trials, as a result of an experimental error. Response time (RT) and accuracy (AC) were recorded for each trial. Following the procedure used by Huizinga et al. (2006) for RT trimming, only RTs for correct trials that followed correct trials were taken into consideration, to avoid bias due to post-error slowing. RTs that were faster than 120 miliseconds or slower than the mean +2.5 SD (for each participant) were also excluded. Finally, if the accuracy for the task was less than 55% for a participant, the scores were removed. The median RT for congruent (C) and incongruent (I) trials was then calculated separately and their ratio (I/C) was used as an indicator of inhibitory control. A higher ratio indicates more interference from irrelevant stimuli and greater inhibitory control difficulty.

WM

The Digit Span Task (Wechsler, 2003) was used to measure WM. Series of random digits of different lengths were presented on the screen. Each participant began the task with two-digit series. The number of digits in a trial increased by one digit after two consecutive series of the same length were presented. Participants used the keyboard to reproduce the digits in a sequence either in the same order they appeared on the screen (forward condition) or in the reverse order (backward condition). Participants first completed the forward condition, followed by the backward condition. Each of them was preceded by four training series (2 × 2-digit, 2 × 3-digit). Each condition ended after two consecutive incorrect responses. The total number of correctly reproduced series for the two conditions was computed separately. The sum of the two scores represented the final score for each participant—higher values reflect better WM.

Task-shifting

The Dots-Triangles Task (Huizinga et al., 2006) was used to measure shifting ability. A series of 4 × 4 grids containing either red dots or green triangles appeared on the screen. There were three experimental blocks, each preceded by a training block. In the first block (“Dots”), only grids that contained three to eight red dots in each of the left and right halves were presented (30 training, 50 experimental trials). Participants had to indicate which half contained more red dots using left or right response keys (Z—left, M—right). In the second block (“Triangles”), only grids that contained three to eight green triangles in each of the upper and lower halves were presented (30 training, 50 experimental trials). Participants had to indicate which of the halves contained more triangles (Z—up, M—down). In the third block (“Mixed”), four red dots trials and four green triangles trials were presented alternatively (90 training, 152 experimental). Participants had to respond to each of them according to their corresponding rule. The stimulus remained on the screen until the participant responded or until 3,500 milliseconds had passed. Interstimulus interval varied randomly between 900 and 1,000 milliseconds. The same RT trimming procedure as for the Flanker Task was used. The median RTs for repeat (R) and switch (S) trials were computed separately and their ratio (S/R) was used as an indicator of shifting ability. Higher ratios reflect more switch costs when changing task-sets, therefore, more difficulties in shifting ability.

Control variables

Age, sex, and MU (time spent with media on an average day) were used as control variables. MU was calculated by averaging participants’ scores (1–6) for each of the nine media activities (MUQ, Section 1; $\bar{X}$ = 2.95, SD = 0.84). The correlations between these variables and the predictor and outcome variables are presented in Table 3.

Procedure

Schools were contacted and permission was obtained from the principal to conduct the research on school premises. Parental consent forms were distributed and retrieved by class masters from all the classes in the two schools. The study had two steps: (a) Completion of the BRIEF and MUQ and (b) Completion of the computerized tasks. First, children completed the two questionnaires in the classroom (20–40 minutes), in counterbalanced order. As opposed to Baumgartner et al. (2014), the questionnaires were in paper-pencil format (vs. online). One of the authors or a research assistant supervised the correct completion of the questionnaires and answered questions for students. All were rewarded with a sticker after finishing this step. A number of 293 children completed both questionnaires. Second, children completed all three computerized tasks in groups of one to five children, in a quiet room, each on an individual laptop. This was also different from the initial study where children were in groups of maximum two children. The testing location differed between the two schools. Children from one of the institutions completed the tasks at school, in one of two available rooms, during regular school program. Due to lack of space on school premises, children from the other institution were invited to the laboratory. They were scheduled at a convenient time and completed the tasks in one of two rooms. A number of 179 participants from both schools completed all three computerized tasks. Those who did not do so (n = 117) were mainly from the school where it was not possible to conduct the testing sessions on school premises, as participants were unable to attend the laboratory meetings. However, apart from this logistical variable, no systematic differences in terms of age, sex, MU, MM, or self-reported EF difficulties were found between completers and non-completers.¹

Results

Outlier Analysis

The initial data exploration identified five participants with significantly higher ratio scores on the Eriksen Flankers Task, which led to a highly skewed distribution. Therefore, their Flanker scores were eliminated. The Digit Span scores for one participant were also eliminated due to not understanding task instructions. No other scores were eliminated due to outlying values.

General Analyses and Results

Closely mirroring the results of Baumgartner et al. (2014), MM while watching TV was highest in the present sample, followed by MM while sending text messages and MM while listening to music. The lowest levels of MM were observed while watching movies/series on the computer. MM while performing non-media activities was lower than MM with other media activities (see Table 1). Regarding the control variables, as opposed to Baumgartner et al. (2014), age was significantly correlated with both indices of MM (MMI-M: r = .24, p = .001; MMI-NM: r = .18, p = .006) and with general MU (r = .24, p < .000). General MU was significantly correlated with both MM indices (MMI-M: r = .63, p < .000; MMI-NM—r = .50, p < .000). As in the original study, girls reported significantly higher scores on the MMI-NM ( $\bar{X}$ = 1.64, SD = 0.40) than boys ( $\bar{X}$ = 1.51, SD = 0.44), t(227) = −2.281, p = .02, Cohen’s d = 0.30, but not on the MMI-M. Sex differences were also found in the performance on the Eriksen Flankers Task, with girls ( $\bar{X}$ = 1.09, SD = 0.10) having lower ratio scores than boys ( $\bar{X}$ = 1.12, SD = 0.10), t(147) = 2.071, p = .04, Cohen’s d = 0.30, indicating slightly better interference control for girls. The two MMIs correlated strongly (r = .81, p < .000). Mirroring results in the initial study, self-reported EF difficulties were moderately correlated with each other (r = .54–.60, p < .000), but not with scores on the computerized tasks (see Table 3). All other correlations between predictor, outcome and control measures can be found in Table 3. Finally, it is important to notice that, in the present sample, participants had slower RTs for both switch and repeat trials on the Dots-Triangles Task and for both congruent and incongruent trials on the Eriksen Flankers Task than in the Baumgartner et al. (2014) study (see Table 2). However, ratio score means and standard deviations that were calculated based on the RTs were similar to the initial study for both tasks.

Main Analyses

Continuous variable approach

Separate multiple linear regression analyses were carried out to test the relationship between MM and EF (self-reported and performance tasks). For all outcome variables, the two MMIs were separately introduced as predictor variables. Scores on the BRIEF subscales and performance indicators on the computerized tasks were used as six separate outcome variables. All analyses were controlled for age, sex and MU.

MM and WM

Self-reported WM difficulties

MMI-M scores significantly predicted self-reported WM difficulties as measured by the BRIEF, F_change(4, 161) = 3.230, p = .01; B = 0.22, SE = 0.08, β = .276, p = .006. In other words, participants with higher scores on the MMI-M reported more difficulties keeping information in mind and processing it. Contrary to Baumgartner et al. (2014), MMI-NM did not significantly predict self-reported WM problems, F_change(4, 203) = 2.268, p = .06; B = .115, SE = 0.067, β = .138, p = .09.

Digit Span performance

No significant relationship was identified between MMI-M scores and the performance on the Digit Span, F_change(4, 102) = 1.748, p = .15; B = −1.49, SE = 0.98, β = −.189, p = .13, nor between MMI-NM scores and Digit Span scores, F_change(4, 130) = 2.161, p = .08; B = −1.327, SE = 0.81, β = −.16, p = .10. Thus, general self-reported MM does not significantly predict participants’ performance on WM computerized tests in the present sample.

MM and Inhibition

Self-reported inhibition difficulties

MMI-M scores significantly predicted self-reported inhibition difficulties as measured by the BRIEF, F_change(4, 161) = 4.151, p = .003; B = .240, SE = 0.078, β = .30, p = .003. Thus, more frequent MM with media activities was related to more difficulty inhibiting inappropriate or unwanted thoughts or behaviors. For the MMI-NM, the overall model yielded a significant relationship with BRIEF inhibition scores, F_change(4, 203) = 2.699, p = .03. When controlling for age, sex, and MU, however, the relationship became nonsignificant (B = .097, SE = 0.066, β = .117, p = .15). These results contrast with the results of Baumgartner et al. (2014), who found a significant positive relation between MMI-NM and self-reported inhibition.

Eriksen Flankers Task performance

Regression analyses yielded no significant relationship between MMI-M and performance on the Flankers Task, F_change(4, 88) = 1.349, p = .26; B = −.015, SE = .032, β = −.064, p = .64. Also, no significant relationship between MMI-NM and scores on the Flankers Task scores was found, F_change(4, 110) = 2.243, p = .07; B = .040, SE = 0.026, β = .166, p = .13. These results indicate the lack of a relationship between general MM and interference control as measured by the Flankers Task.

MM and Shifting

Self-reported shifting difficulties

Analyses yielded no significant relationships between MMI-M and BRIEF shifting scores, F_change(4, 161) = 1.406, p = .23; B = .134, SE = 0.079, β = .17, p = .09, nor between MMI-NM and BRIEF shifting scores, F_change(4, 203) = 1.122, p = .35; B = .067, SE = 0.066, β = .083, p = .31. These results oppose those of Baumgartner et al. (2014).

Dots-Triangles Task performance

Analyses yielded no significant relationship between MMI-M and performance on the Dots-Triangles Task, F_change(4, 102) = 0.944, p = .44; B = .093, SE = 0.059, β = .198, p = .12, and no significant relationship between MMI-NM and performance on the Dots-Triangles Task, F_change(4, 130) = 0.472, p = .76; B = .024, SE = 0.049, β = .051, p = .62. These results support the findings of Baumgartner et al. (2014).

Extreme groups analysis approach

Following Baumgartner et al. (2014) and the method frequently used in MM research (e.g., Ophir et al., 2009; Ralph & Smilek, 2017), data were also analyzed using an extreme-group comparison procedure. To obtain extreme groups with sample sizes and MMI-M means comparable with those in the initial study, participants were split in quartiles based on their MMI-M index. Only those in the upper (i.e., HMMs) and lower (i.e., LMMs) quartiles were included in subsequent analyses. Thus, the descriptive data for the two extreme groups are: LMM (n = 46, $\bar{X}$ = 1.18, SD = 0.098); HMM (n = 43, $\bar{X}$ = 2.23, SD = 0.37).

MM and self-reported EF difficulties

As per Baumgartner et al. (2014), a repeated measures ANCOVA (RM-ANCOVA) was carried out with scores on the three self-reported BRIEF subscales as within-subject factors, MM extreme group as between-subject factor and sex, age and MU, and covariates. Levene’s test indicated unequal variances in the two groups for the BRIEF shifting, F(1,79) = 5.429, p = .02, and WM subscales, F(1,79) = 7.601, p = .007. Therefore, a natural log (log10) transformation of the three BRIEF subscale scores was employed. The RM-ANCOVA was carried out using these transformed variables.

Results indicated a significant main effect of extreme group, F(1, 76) = 10.105, p = .002, $η^{2}$ = .12, d = 0.72. HMM had significantly higher scores than LMM on all three subscales (see Table 4 for adjusted group means). No significant main effect of self-reported EF problems, F(2, 152) = 2.166, p = .12, $η^{2}$ = .028, d = 0.33, and no significant interaction, F(2, 152) = .339, p = .71, $η^{2}$ = .004, d = 0.17, were identified. The model that included the non-transformed variables yielded similar results: a significant main effect of extreme group, F(1, 76) = 11.425, p = .001, $η^{2}$ = .13, d = 0.77, no significant main effect of self-reported EF difficulties, F(2, 152) = 2.339, p = .10, $η^{2}$ = .030, d = 0.35, and no interaction, F(2, 152) = 0.239, p = .79, $η^{2}$ = .003, d = 0.10.

Table 4.

Adjusted Group Means for Computerized Executive Functioning Performance Tasks in Extreme Group Analysis.

	LMM		HMM
	M	SE	M	SE
BRIEF subscales
Working memory	1.50	0.07	1.84	0.07
Inhibition	1.49	0.07	1.86	0.07
Shifting	1.51	0.08	1.81	0.08
Digit Span Task
Forward	8.43	0.33	8.04	0.39
Backward	7.13	0.41	6.10	0.48
Eriksen Flankers Task
Congruent	549.45	18.44	485.24	24.32
Incongruent	606.46	25.36	522.61	33.45
Dots-Triangles Task
Repeat	1,153.31	43.52	911.64	54.44
Switch	1,487.82	73.56	1,209.17	92.02

Note. LMM = light media multitaskers; HMM = heavy media multitaskers; BRIEF = Behavioral Rating Inventory of Executive Function–self-report.

MM and performance on computerized EF tasks

Separate RM-ANCOVAs were also carried out with performance indicators for each performance task as within-subject factors, MM extreme group as between-subject factor and sex, age and MU and covariates. Adjusted means for all computerized tasks are presented in Table 4.

Digit Span—WM

The repeated measures ANCOVA for the Digit Span Task included condition (forward vs. backward) as within-subject factor. Results indicated no significant main effect of extreme group, F(1, 45) = 1.497, p = .23, $η^{2}$ = .032, d = 0.36, no significant main effect of condition, F(1, 45) = 3.224, p = .08, $η^{2}$ = .067, d = 0.53, nor a significant condition by extreme group interaction, F(1, 45) = .847, p = .36, $η^{2}$ = .018, d = 0.27; these results follow the data in the Baumgartner et al. (2014) study.

Eriksen Flankers Task—Inhibition

For the Flankers Task, Levene’s Test indicated unequal variances in the two groups for both congruent, F(1, 43) = 5.793, p = .02, and incongruent trial types, F(1, 43) = 4.313, p = .04. The same natural log10 transformation was computed for the median RTs on both types of trials. The RM-ANCOVA included these transformed trial type variables as within-subject factor. Results indicated a significant effect of extreme group, F(1, 40) = 4.331, p = .04, $η^{2}$ = .098, d = 0.65. No significant main effect of trial type, F(1, 40) = 0.116, p = .74, $η^{2}$ = .003, d = 0.10, and no significant Trial Type × Extreme Group interaction, F(1, 40) = 0.594, p = .45, $η^{2}$ = .015, d = 0.24, were identified. When disentangling group effects, the analyses revealed that the difference between HMMs and LMMs for the incongruent trials is significant (B = .067, SE = 0.033, p = .046, partial $η^{2}$ = .096, d = 0.65), with the HMM group having faster RTs (see Table 4). The difference between groups for the congruent trials is marginally significant (B = .056, SE = 0.028, p = .054, partial $η^{2}$ = .089, d = 0.62), also with faster RTs for HMMs (see Table 4). No speed-accuracy trade-off was identified in the two groups.

Dots-Triangles Task—Task-shifting

Finally, the RM-ANCOVA for the Dots-Triangles Task included trial type (switch vs. repeat) as within-subject factor. A significant main effect of extreme group was found, F(1, 45) = 6.457, p = .02, $η^{2}$ = .125, d = 0.75. HMMs were faster to respond to the task, as compared with LMMs, irrespective of trial type (see Table 4). No main effect of trial type, F(1, 45) = 0.672, p = .41, $η^{2}$ = .015, d = 0.24, or Extreme Group × Trial Type interaction was found, F(1, 45) = 0.141, p = .71, $η^{2}$ = .003, d = 0.10. No speed-accuracy trade-off was identified in the two groups. These results contrast with the findings in the regression analyses and those of Baumgartner et al. (2014).

Discussion

The present study aimed to be a direct replication of the study conducted by Baumgartner et al. (2014) concerning the relationship between MM and EF in adolescents. Results support part of the findings of the initial endeavor, but also highlight some diverging patterns of relationships. Our study replicated the following results of the research conducted by Baumgartner et al. (2014): (a) Higher engagement in MM with media activities (MMI-M) predicted significantly more self-reported difficulties in WM and inhibition; (b) neither MMI-M, nor MMI-NM significantly predicted performance on the three EF computerized performance tests; (c) in the extreme group analysis, HMMs reported more EF difficulties in WM, inhibition and shifting than LMMs; (d) no significant difference between HMMs and LMMs was found in performance on the WM computerized test. In contrast, the following results were not in concordance with the findings in the study of Baumgartner et al. (2014): (a) MMI-M did not predict significantly more self-reported shifting difficulties; (b) MMI-NM did not predict self-reported EF difficulties; (c) The extreme group analysis indicated that HMMs responded faster than LMMs in the Dots-Triangles task on both repeat and switch trials; (d) HMMs responded significantly faster than LMMs on the Flankers Task, irrespective of trial type.

Therefore, when considering the continuous variable approach, the present study mainly supports the findings of Baumgartner et al. (2014). Results indicate that the more adolescents engage in MM with media activities, the more they report having trouble inhibiting inappropriate responses or irrelevant stimuli and keeping information in WM. These results are also backed up by the extreme group approach in both studies. In contrast, MM with media activities did not predict more self-reported problems in shifting ability in this sample. However, when comparing HMMs with LMMs in the extreme group approach, HMM status was associated with more self-reported shifting difficulties than LMM status. This indicates that adolescents at the extreme ends of the MM continuum did differ in significant ways regarding shifting. However, these differences may have been diluted when including adolescents with moderate MM. This pattern of results opens the possibility of a nonlinear relationship between MM and shifting or other aspects of EF.

Some of the differences in results between the two studies might be explained by cultural or societal differences in media behavior between Dutch and Romanian adolescents. As reported by Baumgartner et al. (2014), Dutch adolescents tend to engage highly with technology and media content early in development, leading to higher levels of media exposure than Romanian adolescents. However, recent data about Romanian youth show steep increases in internet access and use (Velicu et al., 2019). Therefore, the Dutch participants in the initial sample and those in the present Romanian sample seem to have been exposed to online media to a similar degree at the time of data collection (86% Dutch in 2013 vs. 84% Romanian in 2018). This is partly supported by the data regarding self-reported MM in the present and initial samples: Dutch and Romanian participants report similar levels of MM while watching TV, using social media, other computer activities and doing homework. However, an interesting difference is that Dutch adolescents reported higher levels of MM with other media activities than Romanians, while Romanian participants reported more MM with non-media activities than Dutch adolescents. These results might reflect different levels of technology-related mastery and self-efficacy. As Dutch adolescents had been exposed to media and technology for a longer time, they might have been more adept at managing multiple media activities simultaneously and thus, engaged more in MM with media activities. In contrast, the relatively recent increase in internet exposure for Romanian teens along with the concomitant increases in the sophistication of mobile gadgets might mean that Romanian participants are still in a process of developing the necessary skills to manage media activities effectively. Therefore, they might be more comfortable doing MM during simpler, non-media activities (e.g., using transportation), as opposed to complex, media activities.

The fact that Romanian adolescents had the highest levels of engagement in non-media activities that are simple, passive, and relatively automatic (using transportation and eating) might also explain the failure of our study to replicate the relationship between MMI-NM and self-reported EF difficulties. More specifically, engaging with media content during these simpler, less cognitively demanding non-media activities does not require fast switching between different contents or the need to control highly competing information. However, these capacities are highly needed when multitasking with more complex and dynamic media activities (Brasel & Gips, 2011). Thus, MM with media activities might subject the cognitive system to more strain than MM with non-media activities, resulting in more perceived EF difficulties. Also, media activities might be used by individuals with higher impulsivity and more difficulty delaying reward during non-media activities to create stimulation in tasks that require sustained attention and cognitive effort (e.g., homework; Nikkelen et al., 2014; Wilmer et al., 2017). The extra stimulation might not be associated with perceived difficulties in completing the task itself or reduced performance, but with more motivation to engage with the task (Ralph & Smilek, 2017).

Results concerning standardized computerized tasks indicate that general MM was unrelated to performance on the WM capacity, interference control and task-shifting tests in the continuous variable approach. However, when taking an extreme group perspective, HMMs responded faster than LMMs in the Dots-Triangles (shifting) and the Flankers (inhibition) tasks irrespective of trial type. At first sight, this pattern of results might be interpreted as reflecting better switching performance for HMMs than LMMs, supporting the findings of Alzahabi and Becker (2013) on adult populations. HMMs also appear to have better interference control, supporting the trend that Baumgartner et al. (2014) identified in their study. However, these differences point toward a different hypothesis for both tasks. For the Dots-Triangles Task, the results did not indicate a significant Trial Type × Group interaction, but only a significant between-group effect. This means that HMMs and LMMs were not faster in their RTs only on switch trials—which would indicate better switching ability (Alzahabi & Becker, 2013). They did so on both types of trials, indicating that HMMs generally responded faster on the task than LMMs, without accuracy costs. These results mirror findings in video-gaming studies, which show that action video-game players generally have faster RTs on computerized tasks without accuracy costs, as opposed to non-players (Dye, Green, & Bavelier, 2009). This appears to be the result of a higher processing speed of video-game players (Dye et al., 2009). Therefore, it is possible that HMMs in the present sample processed visual stimuli faster than LMMs and responded better to task instructions. Given that higher MM involves more engagement with various devices, we can speculate that the RT of the HMM group might have also been aided by a higher familiarity with technology. However, it is important to note that the participants in the present sample had significantly slower RTs on the task as compared with the initial sample. Keeping this in mind, it could be that HMMs here had an average performance and that the LMMs were more cautious in their approach to the task. Given the lower levels of MU also reported by this latter group, it may be that reduced MM came as a result of less experience and familiarity with technology, which also made them wary of engaging fully with the computerized tasks. The pattern of results regarding the performance on the Flankers Task mirror those on the Dots-Triangles Task. Even though HMMs were faster to respond than LMMs, they did so irrespective of trial type. Similar RTs to congruent trials for both groups and faster RTs for incongruent trials only for HMMs would indicate better interference control for the latter group. However, this is not the case here, even though the group effect on congruent trials was only marginally significant. Therefore, HMMs appear to have responded generally faster on the task than LMM, without accuracy costs. Like the Dots-Triangles tasks, this might indicate faster information processing or better compliance to task demands for HMMs, mirroring the results of Dye et al. (2009).

When approaching the differences in results between self-reported EF difficulties and performance on computerized tasks, two explanations might be relevant. First, questionnaires tap into self-perceived individual functioning across contexts, tasks, and social interactions (Paulhus & Vazire, 2007). Thus, they offer a more general evaluation of functioning. In contrast, EFs are time-sensitive and context-dependent (e.g., Albert, Chein, & Steinberg, 2013; Simmonds et al., 2008). They are recruited and interact in a way that is most suitable and needed for the context and task at hand (Elton & Gao, 2014). As life situations demands and contextual task demands are different, it is likely that EF engagement in more elaborate, ecological situations is qualitatively different than in short, specific, and contextual ones. In other words, daily functioning is characterized by complex situations and dynamic interactions with tasks and other individuals. In contrast, performance tasks are completed in a highly controlled, non-emotional, and less demanding context, different from that in which the person usually functions. Therefore, it is likely that the results of performance tasks reflect individual EF performance in this specific laboratory context (Elton & Gao, 2014) and cannot be readily extrapolated to the ecological MM context. This means that, even if an individual has difficulties in daily life situations, they might still be able to voluntarily exercise effort and perform well on a short-term, specific task, completed in a supportive context (Ralph & Smilek, 2017). In contrast, when reporting EF difficulties in questionnaires, respondents refer specifically to their performance in their daily lives. Here, tasks are more diffuse, more spread out and more subjected to interference from unpredictable dynamic factors (e.g., in the school setting). Thus, effort and control must be exerted for a longer time across situations with changing demands, which is more challenging. Putting this into a MM context, HMMs might have difficulties recruiting the necessary resources to focus and respond effectively in an ecological context, where tasks are complex and motivation to resist distraction is lower, especially if media devices are readily available and can be used at will. However, they might be able to do so in a specific, supportive context, with short-term demands, thus yielding a good performance. In this case, this is likely due to contextual characteristics, not necessarily because of better cognitive abilities. Therefore, it is the results on the questionnaires that could be more informative for assessing the relationship between daily EF and MM.

We also propose a second possible explanation for the differences in results between self-reported EF difficulties and performance on computerized tasks, albeit more speculative. Several authors have proposed that EFs rely on a limited pool of cognitive resources (e.g., Barkley, 2012; Baumeister, Bratslavsky, Muraven, & Tice, 1998). These resources are shared among tasks of varying complexity, those that require more self-regulation and cognitive control being more resource-consuming than simpler or more automatic ones. Sustained self-regulatory efforts can lead to a depletion of these resources, which can further result in self-regulatory problems and lapses in performance (Baumeister et al., 1998; Schmeichel, 2007). Relevant data supporting this view come from dual-task interference studies, that show slower reaction times, longer times needed for task completion and more errors when performing tasks with common demands simultaneously versus independently (see Courage et al., 2015, for a review regarding multitasking). As the same participant performs better or worse on the same task depending on cognitive load, it is likely that the difference is due to resource availability, rather than dysfunctions in underlying cognitive process. The MM context is somewhat similar to that in dual-task interference studies: MM involves attending to media activities simultaneously with other media or non-media tasks, the need to coordinate rapid shifts between similar contents of varying complexities and simultaneous information processing (Brasel & Gips, 2011). This requires considerable self-regulatory efforts and is likely to recruit more cognitive resources than unitasking. As such, it is likely that similar decrements in performance as those in dual-task studies appear when performing a task in a MM versus a unitasking context. This can be seen, for example, in studies showing lower learning performance when engaging in media activities during instructive activities (Wood et al., 2012) or lower performance on various computer tasks when multitasking with computer activities that are perceived as difficult (Adler & Benbunan-Fich, 2015). Because HMMs engage in MM very frequently, it is possible that they also face the above-mentioned lapses in performance more often than LMMs, who mostly engage in unitasking. Therefore, HMMs could perceive themselves as having more difficulties over time in tasks known to rely on EFs and, thus, report more problems on EF rating scales than LMMs. However, this perception might result from repeated experiences of contextual resource depletion and not from greater general dysfunctions at the level of underlying EF processes. If this is the case, then HMMs’ and LMMs’ performances should not be significantly different when assessing EFs with computerized tasks, which target underlying EF processes and are performed in a similarly (non)demanding context for both groups. However, as our study did not target this specific mechanism, this explanation remains speculative. Further research is necessary to clarify if the observed differences are a matter of context and resource management or a matter of underlying cognitive deficit.

Because results relating MM to self-reported EF difficulties and computerized tasks performance are in contrast, further research is needed to clarify what mechanisms drive the responses elicited using these two methods of assessing EFs. These will offer important data that may indicate which instrument or combination of instruments is most suitable to extricate the complexity of media effects (self-reports, computerized tasks, or a combination of the two?).

Limitations

The present study is not without limitations. First, as our results are correlational, we cannot derive causal explanations for the relationship between EF and MM. Because we did not include an experimental manipulation that could clarify the direction of influence between these two variables, we cannot rule out the possibility that poorer EFs are responsible for higher levels of MM and not the other way around. Thus, it is possible that adolescents who have less efficient EFs engage more often in MM than those who have more efficient EFs. Experimental studies that manipulate exposure to MM and measure post hoc changes in EFs are needed in order to clarify this matter. Second, the performance of this sample on the Dots-Triangles Task was significantly slower than that reported in Baumgartner et al. (2014) and the literature in general (e.g., Huizinga et al., 2006; i.e., the present sample had significantly higher RTs than those in other studies). This might reflect a failure of our participants to comply to task instructions or the action of other contextual factors that were not considered and controlled for. The present study also differed in some methodological respects from the initial study, which might have led to differences in results. Sample size was smaller (296 vs. 523) and there was a higher percentage of girls (56.1% vs. 48.2%). Questionnaires were completed here in paper-pencil format (vs. online) and computerized tasks were completed in groups of three to five persons (vs. maximum two). Also, in the Eriksen Flankers Task, children only had eight practice trials, as opposed to 50 in the replicated study, due to an experimental error. These aspects must be taken into consideration when interpreting the results.

Conclusion

The present study aimed to replicate the research conducted by Baumgartner et al. (2014). Results replicated the positive relationship between MM with media activities and self-reported EF difficulties. MM with non-media activities was not related to self-reported EF problems in this sample. Results from computerized tasks are mixed. However, HMMs responded faster to task-switching and interference control tasks than LMMs. There were differences between the results of the continuous variable approach and the extreme group approach. This indicates a need for more robust methodology in the research of MM effects and for exploring non-linear relationships.

Footnotes

Acknowledgements

The authors would like to thank the two schools that participated in the research (students, parents, and staff) and Liviu Crișan, Flavia Medrea, Alexandra Colda, and Miruna Cotârlan for their involvement in the data gathering and data analysis process. They would also like to thank the reviewers and editor that offered important feedback that led to the better form of this article.

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.

ORCID iD

Doris Antonia Rogobete

Notes

Author Biographies

Doris Antonia Rogobete is a PhD student at the Applied Cognitive Psychology Doctoral School, Department of Psychology, Babeș-Bolyai University, Romania. Her research interests include child and adolescent development, technology and media use in adolescence and their relationship with cognitive development.

Thea Ionescu, PhD, is an associate professor and researcher in the field of Developmental Psychology at the Department of Psychology, Babeș-Bolyai University, Romania. Her research interests include the development of cognitive flexibility—the variability-stability-flexibility pattern—and the embodied cognition approach—implications for development

Mircea Miclea, PhD, is a professor and researcher at the Department of Psychology, Babeș-Bolyai University, Romania. His interests and activity center on applied research in e-mental health solutions, e-learning, educational software, and computerized psychological evaluation, but also fundamental research in anxiety and cognition.

References

Adler

R. F.

Benbunan-Fich

(2015). The effects of task difficulty and multitasking on performance. Interacting With Computers, 27, 430-439.

Albert

Chein

Steinberg

(2013). The teenage brain: Peer influences on adolescent decision making. Current Directions in Psychological Science, 22, 114-120.

Alzahabi

Becker

M. W.

(2013). The association between media multitasking, task-switching, and dual-task performance. Journal of Experimental Psychology: Human Perception and Performance, 39, 1485-1495.

Barkley

R. A.

(2012). Executive functions: What they are, how they work, and why they evolved. New York, NY: The Guilford Press.

Baumeister

R. F.

Bratslavsky

Muraven

Tice

D. M.

(1998). Ego depletion: Is the active self a limited resource? Journal of Personality and Social Psychology, 74, 1252-1365.

Baumgartner

S. E.

van der Schuur

W. A.

Lemmens

J. S.

te Poel

(2017). The relationship between media multitasking and attention problems in adolescents: Results of two longitudinal studies. Human Communication Research, 44, 3-30.

Baumgartner

S. E.

Weeda

W. D.

van der Heijden

L. L.

Huizinga

(2014). The relationship between media multitasking and executive function in early adolescents. Journal of Early Adolescence, 34, 1120-1144.

Best

J. R.

Miller

P. H.

Jones

L. L.

(2009). Executive functions after age 5: Changes and correlates. Developmental Review, 29, 180-200.

Bodnar

L. E.

Prahme

M. C.

Cutting

L. E.

Denckla

M. B.

Mahone

E. M.

(2007). Construct validity of parent ratings of inhibitory control. Child Neuropsychology, 13, 345-362.

10.

Brasel

S. A.

Gips

(2011). Media multitasking behavior: Concurrent television and computer usage. Cyberpsychology, Behavior, and Social Networking, 14, 527-534.

11.

Cain

M. S.

Mitroff

S. R.

(2011). Distractor filtering in media multitaskers. Perception, 40, 1183-1192.

12.

Cardoso-Leite

Kludt

Vignola

W. J.

Green

C. S.

Bavelier

(2016). Technology consumption and cognitive control: Contrasting action video game experience with media multitasking. Attention, Perception, & Psychophysics, 78, 218-241.

13.

Carr

(2010). The shallows: How the internet is changing the way we think, read, and remember. London, England: Atlantic Books.

14.

Casey

B. J.

Getz

Galván

(2008). The adolescent brain. Developmental Review, 28, 62-77.

15.

Council on Communications and Media. (2016). Media and young minds. Pediatrics, 138(5), e20162591.

16.

Courage

M. L.

Bakhtiar

Fitzpatrick

Kenny

Brandeau

(2015). Growing up multitasking: The costs and benefits for cognitive development. Developmental Review, 35, 5-41.

17.

Crone

E. A.

Steinbeis

(2017). Neural perspectives on cognitive control development during childhood and adolescence. Trends in Cognitive Sciences, 21, 205-215.

18.

Del Missier

Mäntylä

Bruine de Bruin

. (2010). Executive functions in decision making: An individual differences approach. Thinking and Reasoning, 16, 69-97.

19.

Dye

M. W.

Green

C. S.

Bavelier

(2009). Increasing speed of processing with action video games. Current Directions in Psychological Science, 18, 321-326.

20.

Elton

Gao

(2014). Divergent task-dependent functional connectivity of executive control and salience networks. Cortex, 51, 56-66.

21.

Friedman

N. P.

Miyake

Young

S. E.

DeFries

J. C.

Corley

R. P.

Hewitt

J. K.

(2008). Individual differences in executive functions are almost entirely genetic in origin. Journal of Experimental Psychology: General, 137, 201-225.

22.

Galván

(2014). Insights about adolescent behavior, plasticity, and policy from neuroscience research. Neuron, 83, 262-265.

23.

Galván

(2017). Adolescence, brain maturation and mental health. Nature Neuroscience, 20, 503-504.

24.

Gardner

Davis

(2013). The app generation: How today’s youth navigate identity, intimacy, and imagination in a digital world. New Haven, CT: Yale University Press.

25.

Gestsdottir

Lerner

R. M.

(2008). Positive development in adolescence: The development and role of intentional self-regulation. Human Development, 51, 202-224.

26.

Guy

S. C.

Isquith

P. K.

Gioia

G. A.

(2004). Behavior Rating Inventory of Executive Function-Self-Report version. Lutz, FL: Psychological Assessment Resources.

27.

Himi

S. A.

Bühner

Schwaighofer

Klapetek

Hilbert

(2019). Multitasking behavior and its related constructs: Executive functions, working memory capacity, relational integration, and divided attention. Cognition, 189, 275-298.

28.

Hitlin

(2018). Internet, social media use and device ownership in US have plateaued after years of growth. Pew Research Center. Retrieved from https://www.pewresearch.org/fact-tank/2018/09/28/internet-social-media-use-and-device-ownership-in-u-s-have-plateaued-after-years-of-growth/

29.

Huizinga

Dolan

C. V.

van der Molen

M. W.

(2006). Age-related change in executive function: Developmental trends and a latent variable analysis. Neuropsychologia, 44, 2017-2036.

30.

Hutton

J. S.

Dudley

Horowitz-Kraus

DeWitt

Holland

S. K.

(2020). Associations between screen-based media use and brain white matter integrity in preschool-aged children. JAMA Pediatrics, 174(1), e193869.

31.

İmren

Tekman

H. G.

(2019). The relationship between media multitasking, working memory and sustained attention. Uludağ Üniversitesi Fen-Edebiyat Fakültesi Sosyal Bilimler Dergisi, 20, 1075-1100.

32.

Institutul Național de Statistică [The National Institute of Statistics]. (2018). Accesul populaţiei la tehnologia informaţiilor şi comunicaţiilor în anul 2018. Editura Institutului de Statistică.

33.

Johnson

J. G.

Cohen

Kasen

Brook

J. S.

(2007). Extensive television viewing and the development of attention and learning difficulties during adolescence. Archives of Pediatrics & Adolescent Medicine, 161, 480-486.

34.

Kaufmann

Alnæs

Doan

N. T.

Brandt

C. L.

Andreassen

O. A.

Westlye

L. T.

(2017). Delayed stabilization and individualization in connectome development are related to psychiatric disorders. Nature Neuroscience, 20, 513-515.

35.

Lillard

A. S.

Drell

M. B.

Richey

E. M.

Boguszewski

Smith

E. D.

(2015). Further examination of the immediate impact of television on children’s executive function. Developmental Psychology, 51, 792-805.

36.

Lui

K. F.

Wong

A. C. N.

(2012). Does media multitasking always hurt? A positive correlation between multitasking and multisensory integration. Psychonomic Bulletin & Review, 19(4), 647-653.

37.

Maass

Klöpper

K. M.

Michel

Lohaus

(2011). Does media use have a short-term impact on cognitive performance? Journal of Media Psychology, 23, 65-67. doi:10.1027/1864-1105/a000038

38.

Makel

M. C.

Plucker

J. A.

Hegarty

(2012). Replications in psychology research: How often do they really occur? Perspectives on Psychological Science, 7, 537-542.

39.

Miller

E. K.

Cohen

J. D.

(2001). An integrative theory of prefrontal cortex function. Annual Review of Neuroscience, 24, 167-202.

40.

Miller

Nevado-Montenegro

A. J.

Hinshaw

S. P.

(2012). Childhood executive function continues to predict outcomes in young adult females with and without childhood-diagnosed ADHD. Journal of Abnormal Child Psychology, 40, 657-668.

41.

Minear

Brasher

McCurdy

Lewis

Younggren

(2013). Working memory, fluid intelligence, and impulsiveness in heavy media multitaskers. Psychonomic Bulletin and Review, 20, 1274-1281.

42.

Miyake

Friedman

N. P.

(2012). The nature and organization of individual differences in executive functions: Four general conclusions. Current Directions in Psychological Science, 21, 8-14.

43.

Moisala

Salmela

Hietajärvi

Salo

Carlson

Salonen

… Alho

(2016). Media multitasking is associated with distractibility and increased prefrontal activity in adolescents and young adults. NeuroImage, 134, 113-121.

44.

Nikkelen

S. W.

Valkenburg

P. M.

Huizinga

Bushman

B. J.

(2014). Media use and ADHD-related behaviors in children and adolescents: A meta-analysis. Developmental Psychology, 50, 2228-2241.

45.

Ophir

Nass

Wagner

A. D.

(2009). Cognitive control in media multitaskers. Proceedings of the National Academy of Sciences, 106, 15583-15587.

46.

Paulhus

D. L.

Vazire

(2007). The self-report method. Handbook of Research Methods in Personality Psychology, 1, 224-239.

47.

Pollard

M. A.

Courage

M. L.

(2017). Working memory capacity predicts effective multitasking. Computers in Human Behavior, 76, 450-462.

48.

Poole

D. A.

Nunez

Warren

(2007). The story of human development. Pearson.

49.

Ralph

B. C.

Smilek

(2017). Individual differences in media multitasking and performance on the n-back. Attention, Perception, & Psychophysics, 79, 582-592.

50.

Ralph

B. C.

Thomson

D. R.

Cheyne

J. A.

Smilek

(2014). Media multitasking and failures of attention in everyday life. Psychological Research, 78, 661-669.

51.

Ralph

B. C.

Thomson

D. R.

Seli

Carriere

J. S.

Smilek

(2015). Media multitasking and behavioral measures of sustained attention. Attention, Perception, & Psychophysics, 77, 390-401.

52.

Redick

T. S.

Shipstead

Meier

M. E.

Montroy

J. J.

Hicks

K. L.

Unsworth

… Engle

R. W.

(2016). Cognitive predictors of a common multitasking ability: Contributions from working memory, attention control, and fluid intelligence. Journal of Experimental Psychology: General, 145, 1473-1492.

53.

Ridderinkhof

K. R.

van der Molen

M. W.

(1995). A psychophysiological analysis of developmental differences in the ability to resist interference. Child Development, 66, 1040-1056.

54.

Rideout

V. J.

Foehr

U. G.

Roberts

D. F.

(2010). Generation M²: Media in the lives of 8-to 18-year-olds. Henry J. Kaiser Family Foundation, Menlo Park, California.

55.

Rosen

L. D.

Carrier

L. M.

Cheever

N. A.

(2013). Facebook and texting made me do it: Media-induced task-switching while studying. Computers in Human Behavior, 29, 948-958.

56.

St. Clair-Thompson

H. L.

Gathercole

S. E

. (2006). Executive functions and achievements in school: Shifting, updating, inhibition, and working memory. Quarterly Journal of Experimental Psychology, 59, 745-759.

57.

Sanbonmatsu

D. M.

Strayer

D. L.

Medeiros-Ward

Watson

J. M.

(2013). Who multi-tasks and why? Multitasking ability, perceived multi-tasking ability, impulsivity, and sensation seeking. PLoS ONE, 8(1), e54402.

58.

Schmeichel

B. J.

(2007). Attention control, memory updating, and emotion regulation temporarily reduce the capacity for executive control. Journal of Experimental Psychology: General, 136, 241-255. doi:10.1037/0096-3445.136.2.241

59.

Simmonds

D. J.

Pekar

J. J.

Mostofsky

S. H.

(2008). Meta-analysis of Go/No-go tasks demonstrating that fMRI activation associated with response inhibition is task dependent. Neuropsychologia, 46, 224-232.

60.

Swing

E. L.

Gentile

D. A.

Anderson

C. A.

Walsh

D. A.

(2010). Television and video game exposure and the development of attention problems. Pediatrics, 126, 214-221.

61.

Szumowska

Kossowska

(2016). Need for closure and multitasking performance: The role of shifting ability. Personality and Individual Differences, 96, 12-17.

62.

TNS Opinion and Social. (2018). “Media use in the European Union” report [Standard Eurobarometer, 88]. European Commission, Directorate-General for Communication. doi:10.2775/116707

63.

Uncapher

M. R.

Thieu

M. K.

Wagner

A. D.

(2016). Media multitasking and memory: Differences in working memory and long-term memory. Psychonomic Bulletin and Review, 23, 483-490.

64.

van der Schuur

W. A.

Baumgartner

S. E.

Sumter

S. R.

Valkenburg

P. M

. (2015). The consequences of media multitasking for youth: A review. Computers in Human Behavior, 53, 204-215.

65.

Velicu

Balea

Barbovschi

(2019). Acces, utilizări, riscuri și oportunități ale internetului pentru copiii din România. Rezultate ale proiectului EU Kids Online 2018. EU Kids Online și DigiLiv-REI. [Access, utilizations, risks and opportunities of the internet for children in Romania. Results of the EU Kids Online 2018 Project] Retrieved from http://rokidsonline.net/wp/wp-content/uploads/2019/01/EU-Kids-Online-RO-report-15012019_DL.pdf

66.

Wallis

(2010, January). The impacts of media multitasking on children’s learning and development: Report from a research seminar. New York, NY: The Joan Ganz Cooney Center at Sesame Workshop.

67.

Wechsler

(2003). Wechsler Intelligence Scale for Children–Fourth Edition. San Antonio, TX: The Psychological Corporation.

68.

Wiley

Jarosz

A. F.

(2012). How working memory capacity affects problem solving. In Ross

B.H.

(Ed.), Psychology of learning and motivation (Vol. 56, pp. 185-227). Academic Press. Retrieved from https://doi.org/10.1016/B978-0-12-394393-4.00006-6

69.

Williams

P. G.

Thayer

J. F.

(2009). Executive functioning and health: Introduction to the special series. Annals of Behavioral Medicine, 37, 101-105.

70.

Wilmer

H. H.

Sherman

L. E.

Chein

J. M.

(2017). Smartphones and Cognition: A review of research exploring the links between mobile technology habits and cognitive functioning. Frontiers in Psychology, 8, Article 605.

71.

Wiradhany

Nieuwenstein

M. R.

(2017). Cognitive control in media multitaskers: Two replication studies and a meta-analysis. Attention, Perception, & Psychophysics, 79, 2620-2641.

72.

Wood

Zivcakova

Gentile

Archer

De Pasquale

Nosko

(2012). Examining the impact of off-task multi-tasking with technology on real-time classroom learning. Computers and Education, 58, 365-374.

73.

Zelazo

P. D.

Cunningham

W. A.

(2007). Executive function: Mechanisms underlying emotion regulation. In Gross

J. J.

(Ed.), Handbook of emotion regulation (pp. 135-158). New York, NY: The Guilford Press.