Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Dec 9.
Published in final edited form as: Psychol Assess. 2025 Oct;37(10):535–546. doi: 10.1037/pas0001400

Families Being Supportive Together: A Multi-Method and Multi-Informant Intensive Longitudinal Study of Family Protective Mechanisms for Adolescent Depression

Sunhye Bai 1, Nicole M Froidevaux 1, Mengyun Chen 2, Andrew C High 3, Kaitlyn E Ewing 4, Jessica DeFelice 5, Jonathan Weaver 6, Kelly W Ngigi 6, Meghan R Riccio 4, Shou-Chun Chiang 7, Liu Bai 8, Erika S Lunkenheimer 9, Timothy R Brick 6
PMCID: PMC12684419  NIHMSID: NIHMS2115741  PMID: 41129397

Abstract

To advance the design and use of intensive longitudinal methods (ILM) in investigations of adolescent depression, we conducted a multi-method and multi-informant study of daily parent-youth interactions, specifically, supportive communication, consisting of (i) naturalistic video observations of parent-youth interactions; (ii) passive collection of Bluetooth low energy (BLE) signals to approximate parent-youth proximity; and (iii) scheduled, proximity-contingent and self-initiated ecological momentary assessments (EMA). We examined whether these novel and complementary approaches enhanced the assessment of parent-youth interactions, a key source of risk and protection for youth mental health. Specifically, we report participant compliance on the video recording procedures, and describe preliminary results from our observational coding of supportive communication. We also report compliance rates on EMAs and examine the frequencies of parent-youth interactions per self-report and BLE signals. Participants in the two-week long protocol were 12–15-year-old adolescents (N = 138; 63.8% female, 42% CESD≥16) and their parents (95.7% biological mothers, 25% CESD ≥ 16). Dyads completed mean 122.6 minutes (SD = 85.6) of video recordings. In 387 minutes of recordings from three pilot families, we identified 52 supportive communication episodes. The average parent and youth were compliant with EMA procedures, completing the recommended minimum of 40 cumulative surveys each. Parents and youth reported that they interacted with the other member in mean 56% to 83% of the EMAs. The study demonstrates innovative ways to leverage technology to conduct multi-method and multi-informant intensive longitudinal assessments of interpersonal interactions, a key source of risk and protection for adolescent mental health.

Keywords: naturalistic observation methods, ecological momentary assessments, parent-child interaction, depression, social support

Affecting one out of every five adolescents, depression is a common and impairing disorder in youth (Substance Abuse and Mental Health Services Administration, 2022; Thapar et al., 2022). Given its high prevalence rate and sequelae of serious negative outcomes (Copeland et al., 2021; Thapar et al., 2022), identifying specific malleable factors in the proximal environment that can mitigate symptoms in early adolescence is important; parental support, evident in parent-youth interactions, is one such protective factor. Although adolescents become more autonomous, independent and peer-oriented, parents remain an influential source of support in an adolescent’s life (Hadiwijaya et al., 2017; Lam et al., 2012; Tsai et al., 2013). Parents can provide comfort and closeness, assist with practical problem solving, advocate for youth needs, and connect youth to professional services (Gariépy et al., 2016; Rueger et al., 2016). Past research on risk and protective processes for adolescent depression have emphasized parental support yet rarely examined youth support seeking behaviors (Heerde & Hemphill, 2018), despite the latter being one of the strongest predictors of support provision (Barbee & Cunningham, 1995). A nuanced characterization of both youth support seeking and parental support provision in daily parent-youth interactions is needed to inform interventions for youth at risk for depression.

Prior Studies Using Intensive Longitudinal Methods

Intensive longitudinal methods (ILM) refer to a diverse collection of strategies that involve frequent and repeated assessments of individuals’ thoughts, feelings, and behaviors. They are often conducted in the natural environment using ambulatory assessment procedures (Heron et al., 2017; Russell & Gajos, 2020; Trull & Ebner-Priemer, 2013). Numerous ILM studies have been conducted to assess how various aspects of interpersonal interactions confer risk or protection for adolescent depression (Sequeira et al., 2020). For example, one 8-week long daily diary study of 8–13-year-olds found that children who reacted to peer and academic problems with higher levels of negative mood had more depression symptoms three years later when they were 11–16 years old (Bai et al., 2020). In a study of 112 adolescent girls with risk for depressive disorders, EMAs administered 3 to 4 times a day for 16-days assessed interpersonal emotion regulation strategy use (Do et al., 2023). The study found that parental involvement in adaptive emotion regulation strategies predicted 1-year decreases in depressive symptoms (Do et al., 2023). Although primarily relying on self-reports, EMAs have increased our understanding of daily risk and protective processes for depression.

ILM have relied on self-report surveys, and more recently, observations of behaviors detectable on smartphones via passive sensing technology. A 12-month long study of passively collected words that 90 adolescents typed into social communication applications found that on weeks when youth used more first-person singular pronouns (e.g., “I”), they were more likely to meet criteria for a major depressive disorder (Funkhouser et al., 2024). Similarly, a 6-month long study of 141 sexually and gender diverse youth found that more time spent at home was associated with worse mood among sexual and gender minority youth with low family support and for non-sexual and gender minority youth with high family support (Bitran et al., 2024). While these studies have significantly contributed to our understanding of adolescent depression, less accessible have been observations of face-to-face interactions that naturally and spontaneously occur in daily life. Whereas increases in screen time have spurred observations of online interactions (Alexander et al., 2024), observations of face-to-face interactions in daily life, which continue to be highly clinically significant (Hamilton et al., 2021; Twenge et al., 2019), are limited, perhaps due to logistical challenges of collecting this data.

Ecological Behavioral Observations of Interpersonal Interactions in Daily Life

Ecological behavioral observations of interpersonal interactions, via naturalistic video recordings, is a unique form of ILM, that requires minimally burdensome actions on the part of the research participants. Given the significance of interpersonal relationships on depression, conducting ecological behavioral observations of interactions, in naturalistic settings, is crucial to identify specific targets for depression prevention. Past research has relied on laboratory tasks to observe parent-child interactions, standardizing the nature of the interaction, situation, and the physical environment. However, in-laboratory tasks do not represent the real-world contexts in which interactions arise (Repetti et al., 2013).

Of the few ecological behavioral observation strategies in psychological research, one example is the electronically activated recorder (EAR), which intermittently records ambient sound samples at times unknown to the research participant (Mehl, 2017). Researchers have also targeted specific and meaningful moments in the day, such as nighttime, to assess infant sleep arrangements and parental behaviors over one night at each assessment point (i.e., 3M and 6M) in their longitudinal study (Teti et al., 2022). These prior studies demonstrate the feasibility of conducting ecological behavioral observations outside of the laboratory, and each approach offers participants varying degrees of flexibility and control. Notably, variations in one-party consent laws make some approaches to intensive passive video observations infeasible in some geographical regions, necessitating a minimally burdensome approach that retains some participant control. To address these challenges and to naturally assess face-to-face interactions in daily life, the current study conducted event-contingent naturalistic video recording procedures, wherein participants were required to turn on a small and unobtrusive video recording device whenever they and other consented individuals were in the car together over the 14 days of the research study.

EMAs of Dyadic Interactions in Daily Life

Self-report surveys complement behavioral observations, and ambulatory EMAs reduce recall bias in comparison to single-administration questionnaires. Still, these methods often require informants to recall their thoughts, feelings and behaviors during events that had occurred several hours prior to the survey. To reduce possible bias and better capture the immediate context in which these events occur, some studies have conducted event-contingent momentary assessments, wherein participants initiate a survey when an event of interest has occurred. However, this approach also relies on the participant to closely attend to study procedures as they go about their daily lives, which may be a challenge for individuals with elevated depression symptoms (Seidman et al., 2022).

Prior research has used passively collected data such as information about sedentary behaviors to deploy assessments and interventions with more specificity (Fiedler et al., 2023; Yang et al., 2023). However, no study has extended this framework to detect dyadic interactions and signal multiple family members to conduct multi-informant assessments of interactions. To address this limitation, we use Bluetooth Low Energy (BLE) signals emitted and detected by smartphones carried by the parent and youth to assess physical proximity (Schaible et al., 2022) and deploy proximity-contingent EMAs when real-time BLE data indicated a high likelihood of interaction. This highly innovative approach draws on the combined strengths of signal- and event-contingent surveys, and significantly reduces the recall bias in traditional EMAs.

Current Study

The current study tested novel and complementary ILMs to capturing parent-youth interactions in the immediate contexts when and where they occur to enhance clinical assessments: (i) naturalistic video observations of parent-youth interactions in the participant’s car; (ii) passive collection of BLE to identify opportunities for parent-youth interactions; and (iii) scheduled, proximity-contingent, and self-initiated EMAs that assessed parent-youth interactions. First, we examined the feasibility of our naturalistic video recording procedures and described their content: characteristics of supportive communication episodes that spontaneously arise in daily parent-youth interactions. Second, we reported compliance rates of EMA by type (i.e., scheduled, proximity-contingent, and self-initiated) and examined the frequencies of parent-youth interactions per self-report and BLE data. We examined the linkages between depression symptoms and video recording duration, EMA compliance rates, and the frequency of interactions to further assess whether the feasibility of these innovative methods were shaped by depression severity.

Method

Recruitment and Participants

Families were recruited from April 2021 to January 2024 through flyers distributed via middle and high schools, universities, mobile health clinics, community events, youth serving-organizations, direct mailing advertising, community magazines, community websites, and local Facebook groups. Families were also recruited from ResearchMatch.org, a research participant pool maintained by the university Child Study Center, and snowball sampling. Recruitment targeted residents of Pennsylvania. Inclusion criteria for youth were as follows: (a) ages 12 to 15; (b) lived with the participating caregiver at least 5 days/week on a typical week; and (c) primary language spoken at home was English. The participating caregiver had to have a car that they used ≥ 2 times per week, an active driver’s license and legal custody of the youth. Given the clinical focus on depression symptoms, emphasis on family interactions, and the complex study procedures, exclusionary criteria for youth and parents per parent report were: (a) current diagnosis of bipolar disorder, obsessive-compulsive disorder, alcohol/substance use disorder, conduct disorder, eating disorder, and developmental disability such as autism spectrum disorder; (b) past 12-month suicide attempt or preparatory acts; and (c) past 12-month inpatient or partial hospitalization, or intensive outpatient treatment for substance use, suicide, or any other emotional or behavioral problem. Additional exclusionary criteria included current social (e.g., homelessness), medical (e.g., neurocognitive impairment), or psychiatric (e.g., active psychosis) conditions that would interfere with youth or parent participation.

Participants were 138 youth (M age = 13.87, SD = 1.12) and their legal caregivers (M age = 43.88, SD = 5.14). Parents self-identified as biological mothers (95.7%, n = 132), adoptive mothers (1.4%, n = 2), and biological fathers (2.2%, n = 3). Parents reported their child’s gender as female (n = 80), male (n = 52), and other/non-binary (n=6). Parents described their youth as 11.6% Hispanic, 88.4% White, 0.7% Native American, 10.9% Black or African American, 2.9% Asian, and 3.6% Other. Most parents (65.3%, n = 90) reported a household income of $75,000 or more, and 71.0% of the parents reported that they work full time. Mean depression score on the Center for Epidemiologic Studies Depression (CESD) for the youth was 14.17 (SD = 11.07, Range = 0–50; 42% CESD ≥ 16), as the study focused on adolescent depression, and for parents, 9.65 (SD = 8.42, Range = 0–40; 25% CESD ≥ 16).

Procedures

We begin by describing the overall study design, including the baseline appointment, the intensive observational period, and the follow-up assessment. Then, we detail the intensive observational period, which consisted of naturalistic video recordings, BLE signals and EMAs. Next, we report on compensation. Finally, we describe data processing and quality assurance activities, as well as behavioral coding procedures.

Overall Study Design

Screened and eligible participants completed a baseline appointment by video conference, a two-week long intensive observational period and a brief follow-up meeting by video conference. Prior to the baseline appointment, research staff mailed data collection equipment to the participants with prepaid return labels, including: two Android Pixel 2 smartphones1, a Aukey Dash Camera (model DR01 or DRA1) to record interactions in the car, three BLE beacons (small black discs that are approximately 1.5 inches by 1.5 inches) to identify whether the participant phones were located in the car, youth bedroom/private space, or a shared family space (e.g. kitchen or family room), and mounts, chargers and cords as needed.

During the baseline appointment, participants provided informed consent, began single-administration questionnaires on RedCap, set up the study equipment, and received training on study procedures relevant to the intensive observational period. Other family members (i.e., other parent, siblings) who were not the foci of the study also provided informed consent to be video recorded for research purposes during the baseline appointment. The two-week intensive observational period began the day after the baseline appointment and lasted 14 days. This period consisted of: (b) in-car video recordings; (c) passive monitoring of parent-youth proximity via BLE and (c) parent and youth EMAs. Participants were instructed to always carry around their smartphones. The Wear-It application, pre-installed on each device, facilitated the passive monitoring of parent-youth proximity and the collection of EMAs. Research staff monitored compliance and messaged the parents every other day with tailored messages to increase compliance. Follow-up assessments were conducted via video conferencing, after the conclusion of the intensive observational period. Parents provided consent, youth provided assent and all procedures were approved by the institutional review board.

Intensive Observational Period

Naturalistic Video Recordings.

During the baseline appointment, focal parent and youth mounted a dashboard camera on their car’s dashboard or below their rearview mirror with guidance from research staff. The camera pointed to the interior of the car, capturing the driver and passenger’s faces and partial view of passengers in the rear seats. The camera automatically began recording when it was plugged into the car’s auxiliary power outlet and the car engine was on, and stopped recording when it was unplugged or the car engine was turned off. Research staff instructed participants to plug in the camera whenever the youth and the parent were in the car together, with or without other consented family members (e.g., other parent, siblings). Research staff requested that participants do not record when unconsented individuals, such as friends, were in the car. Participants were requested to go about their lives as they normally would, not spending any more or less time in the car than they would have, had they not been participating in the study. The participants also placed a small BLE beacon in the car. Wear-It, installed on the Android smartphones, used real time BLE signals between the two participants’ phones and the stationary BLE beacon to push alerts on the smartphone, reminding participants to turn on the dashboard camera when they were detected to be in the car together.

BLE Signals.

For 112 dyads, the Wear-It application continuously recorded BLE signals emitted from their smartphone carried by their partner (i.e., parent or youth) during the intensive observational period. Signal strength was then converted to meters between the youth and parent to assess physical proximity.

Ecological Momentary Assessments.

During the intensive observational period, participants completed several EMA of parent-youth interactions, which were prompted via three different mechanisms: scheduled, proximity-contingent, and self-initiated. The goal of the design was to maximize the probability of capturing meaningful parent-youth interactions and reduce participant burden. First, participants were prompted to complete four scheduled EMAs per day, at wake, late afternoon (i.e., after school), evening (i.e., after dinner) and bedtime, scheduled at 8AM, 4PM, 7PM and 9PM by default. However, participants were invited to make minor adjustments to schedule as needed during the baseline appointment. Second, participants could self-initiate an EMA if they wanted to report on a particularly meaningful interaction with the other participating member. In practice, this option was used frequently by many participants for a variety of reasons, including making up a narrowly missed scheduled EMAs. Third, for 112 families, real-time BLE signals were used to deploy proximity-contingent EMAs about parent-youth interactions to the parent and youth’s phone. Participants did not receive prompts to complete a proximity-contingent survey between 12AM and 6AM, and participants did not receive any new prompts to complete a survey within 1 hour of completing an EMA. Last, there were no more than 2 proximity EMA prompts spaced at least 1 hour apart made within each of the following four blocks of time: before 10AM; 10AM to 2PM; 2PM to 6PM; and 6PM – 12AM. All EMA surveys asked the same set of questions examining affect, stress, and if endorsing an interaction, characteristics of parent-youth supportive interactions. The scheduled bedtime survey asked additional questions about parent-youth supportive interactions, general parent-youth relationship quality, and academic, peer and/or work problems during the entirety of the day. The morning survey asked additional questions about sleep.

Compensation

Each participant was compensated up to $160: $25 for the baseline appointment and the completion of single-administration surveys, $30 for the follow-up appointment, $45 for the first week of intensive observational period, and $60 for the second week of intensive observational period. For the purposes of compensation, we separately assessed compliance on three types of tasks during the intensive observational period: (1) bedtime surveys, (2) daytime surveys inclusive of the morning, afternoon, evening, proximity-contingent and self-initiated surveys, and (3) video recordings. For the bedtime surveys, participant compliance was assessed to be in the highest tier (i.e., Tier 3), if they completed 5–7 surveys during each week of the intensive longitudinal period. For daytime surveys, participant compliance was assessed to be in the highest tier (i.e., Tier 3), if they completed 15–21 surveys in each week of the intensive longitudinal period; no more than 6 daytime surveys within a single day were counted toward the weekly total for the purposes of compensation, although participants could complete more. We shared the compensation structure rubric with participants during the baseline appointment. Throughout the intensive observational period, research staff sent text messages to parent participants approximately every other day. Messages let parents know if they or their youth was on target with the EMA completion, else, gently reminded them to complete surveys. For video recordings, we assessed whether the participants made no recordings, or “some” recordings on each week after we received the equipment.

Data Processing and Quality Assurance

Naturalistic video recordings were stored on microSD cards during the observational period. Upon receipt of the study equipment, staff transferred files to a secure network attached storage server. Two trained research assistants reviewed all video recordings for safety concerns (e.g., disclosures of abuse, suicidality), inclusion of sensitive information (e.g., disclosure of credit card information) and the presence of unconsented people. During these two-stage review process, research assistants also indicated the start and end times of video segments that met inclusionary criteria for further analysis: (1) from day 2 to 14 of the intensive observational period, (2) include the participating youth and parent, and (3) from the camera placed on the front of the car2. Video files that contained no words from any consented member of the family were excluded from further analysis. Finally, we capped the amount of video recordings to 250 minutes for each family to prevent over-representation of select families and to increase feasibility of data processing. Segments of video recordings that contained unconsented individuals (e.g., restaurant employees at drive through windows), and those that contained disclosures of very sensitive information (e.g., disclosure of credit card numbers, parental disclosure of sexual abuse in childhood), were masked or muted using Adobe Premiere Pro. Finally, we exported the audio recording for transcriptions.

We transcribed the audio recordings using one of three approaches to minimize cost and time and maximize accuracy: GoTranscript, a professional human-made transcription service, artificial intelligence driven transcription software (e.g., SONIX), and manual transcription by research assistants. Regardless of the approach, all transcriptions were reviewed for accuracy and stripped of personally identifying information. The research staff performed a third check to ensure transcription quality, and exported the transcriptions to Excel, such that a talk turn was the smallest unit of analysis.

Behavioral Coding of Supportive Communication

Based on the video recordings of three pilot families, the first (SB) and third (MC) author developed an observational coding system to assess supportive communication in parent-youth interactions, with support from co-author MRR. Then, four naïve coders were trained on the coding system, with the codes identified by SB and MC serving as the key. In the first stage of the two-step coding process, coders first identified supportive communication episodes. As described below, the second stage of coding identified youth support seeking behaviors and parent support enactment behaviors in each episode identified in Stage 1. Both stages of coding primarily relied on the transcription and audio recordings, supplemented by non-verbal behaviors observed in video recordings. During training, each coder received weekly feedback on their codes, and met weekly with advanced coders to discuss recurring challenges in coding and to reach group consensus on all codes.

Measures

The description of measures is grouped by modality: single administration surveys, naturalistic video recordings, BLEs and EMAs.

Single Administration Surveys

The Center for Epidemiologic Studies Depression (CESD) is a well validated 20-item measure for assessing depression symptoms in adults and adolescents (Radloff, 1977, 1991). Parents and youth reported on their own past-week feelings and behaviors, with response options ranging from 0 = rarely or none of the time (less than 1 day) to 3 = most or all of the time (5–7 days). Responses were summed and a score of 16 was considered to indicate elevated risk for depression. Participants completed the scale once, at any point between the baseline and follow-up appointments, and the Cronbach’s alpha were .90 for parents and .92 for adolescents.

Naturalistic Video Recordings

Video Recording Compliance was the total minutes of eligible video recordings, with 250 minutes set as the possible maximum.

Supportive Communication Episode was indicated when the youth sought support (see below for details on youth support seeking); a parent did not necessarily have to enact support. When youth support seeking was identified, all parent and youth talk turns related to the topic that was the focus of the youth support seeking was coded as being a part of the supportive communication episode. Mid-episode interruptions (e.g., side conversations) ≤ 30 seconds long were coded as part of the episode. Interruptions > 30 seconds were excluded, and supportive communication before and after the interruption were coded as two distinct episodes. Exclusions were conversations related to the research study (e.g., remarks about EMAs), brief permission seeking behaviors (e.g., asking to play video games), meal planning, snack negotiations, and household management discussions (e.g., discussion of a parent dropping a sibling off at an extracurricular activity).

Youth Support-seeking Behaviors convey the seeker’s negative emotional state and/or that they want some form of help to solve a problem or feel better (Barbee & Cunningham, 1995). In each supportive communication episode, we coded for the presence of three types of support seeking behaviors. Disclosure (“tell”) was defined as telling the parent about a negative event or experience. Brief complaints about a negative situation and long narratives about an unpleasant experience both met criteria for “tell.” Disclosures could be prompted or unprompted by the parent. Emotion expression (“emotion”) was defined as expression or disclosure of current or past negative emotions. Negative emotion could also be conveyed by the tone of the youth’s speech. Requests (“request”) for assistance was defined as direct requests for help, advice or instrumental support. The three support seeking behaviors were not mutually exclusive and all three behaviors could be identified in a supportive communication episode.

Parent Support Enactment refers to statements and behavioral offers of support. As with support seeking, we coded for the presence of four types of support enactment in each supportive communication episode (Crowley & High, 2020). Informational support (“informational”) was coded when parents asked questions to gather or elicit more information or when parents gave advice, suggestions, opinions, perspectives, or share their own experience. Emotional support (“emotional”) was coded when parents provided emotional comfort, in the form of expressing care, concern, assurance or encouragement of an emotional nature. Esteem support (“esteem”) was coded when parents validated and affirmed their child’s identities, abilities and accomplishments, and Tangible/Instrumental support (“instrumental”) was coded when the parent did something active to help the child, including lending an object, or helping with homework. The four support enactment subtypes were not mutually exclusive.

BLE Signals

BLE Estimated Interaction. The Wear-It application on the study smartphones tracked the signal strength of every BLE device within its range, including the partner’s smartphones, approximately every 60 seconds. For a subset of participants (N=112, 81.1%), the Wear-It app on the study smartphone was programmed to recognize its partner’s signals, and start a timer when the partner’s smartphone was detected to be within 4 meters per the AltBeacon code library’s estimate (Young & Radius Networks, 2023). If that device stayed within the 4-meter range for more than 15 minutes, the Wear-IT app considered that a potential interaction between the two participants, which lasted until the participants separated. A short grace period (up to 5 minutes) of greater distance or missed detection was permitted during the interaction to allow for minor signal issues. If the paired device came back into the 4-meter interaction range, the potential interaction continued; otherwise it was considered ended, and the 15-minute timer restarted when the devices next came into proximity. Five minutes after the conclusion of any interaction lasting more than 15 minutes, the participant received an alert to complete a proximity-contingent EMA about the potential interaction. The 5 minutes of separation was intended to ensure the conclusion of any interaction that may have taken place during the 15-minute epoch of proximity, and to avoid social pressure from completing a survey in close proximity to the other participant. We report the number of BLE signal estimated interactions from 39 families.

Ecological Momentary Assessments

EMA Compliance was computed as the number of surveys started during the 14 day long intensive observational period. Due to participants being able to self-initiate a survey at any time, there was no upper cap on the number of surveys the participants could take.

Self-Reported Interaction was assessed with a single item on every EMA: “since the last survey, did you see or communicate with/talk to your teen/parent?” Our operationalization of an interaction was intentionally inclusive, as participants could seek or provide support using verbal or non-verbal signals, in-person or via smartphone devices. Participants could select “yes” if they had only seen their partner without necessarily talking, seen and talked to their partner, or not seen in-person but communicated with the partner via text messaging. An affirmative response indicated that the participant did interact with their partner.

Analytic Plan

First, we report descriptive statistics of the minutes of eligible video recordings completed by study participants to demonstrate the feasibility of these procedures. Then we describe their content - descriptive statistics from initial coding of supportive communication episodes for three pilot families. Second, we report compliance rates of EMA by type (i.e., scheduled, proximity and self-initiated). Then, we describe frequencies of parent-youth interactions per self-report, Spearman’s correlations representing their split-half reliability, and BLE signals, and conduct repeated measures ANOVAs to assess when and on which type of EMA participants most frequently report having interacted with their parent or youth. In each aim, we describe Pearson’s correlations with parent and youth depression symptoms.

Transparency and Openness

All data, analysis code, and research materials are available upon request from the first author. Data from individuals who consented to public data sharing are available on the NIMH Data Archive (doi: 10.15154/5z08-gq64; Bai, 2025). Data were analyzed using STATA v.18 and SPSS v.29 and R v.4. This study’s design and its analysis were not preregistered.

Results

Descriptives of Naturalistic Video Observations

Of the 138 families enrolled in the study, 133 families provided some video recordings, and 5 families did not provide any content: 3, due to technical issues, and 2 due to noncompliance. Among the 133 families who provided some video recordings, mean cumulative duration was 122.6 minutes (SD = 85.6, Median = 109.8 minutes), with a range of 0 minutes to 250 minutes. Seven of the 133 families had 0 minutes of eligible content, 4 due to non-compliance, and 3 due to having recordings only on the first day of the intensive observational period. As shown in Table 1, parent depression symptoms was negatively correlated with video duration, such that higher levels of depression symptoms was correlated with shorter video duration (r = −0.17). Youth depression symptoms was not correlated with duration (r = −0.11).

Table 1.

Video recording duration, EMA counts and their correlations to parent and youth depression symptoms
Video recording duration & EMA counts Parent Depression Correlation Youth Depression Correlation
N M (SD) Median Range r r
Video recording (min) 133 122.6 (85.6) 109.8 0–250 −.17* −.11
Parent EMA
Totala 138 46.93 (19.01) 49 9–102 −.09
Daytime Totala 138 37.17 (16.53) 38 8 – 93 −.06
Scheduled
 Wake 138 7.78 (4.12) 8 0–14 −.17*
 Afternoon 138 6.96 (4.18) 7 0–14 −.14
 Evening 138 6.63 (3.99) 6 0–14 −.13
 Bedtime 138 9.75 (4.11) 11 0–14 −.18*
Self-initiated 138 10.49 (9.61) 8.5 0–58 .10
Proximity-contingent 78 9.94 (9.59) 7 1–39 −.06
Youth EMA
Totala 138 40.62 (18.66) 41.5 7–90 −.30**
Daytime Totala 138 31.97 (15.60) 32 6–79 −.30**
Scheduled
 Wake 138 6.43 (3.86) 6 0–13 −.13
 Afternoon 138 6.33 (3.57) 7 0–14 −.25**
 Evening 138 5.91 (3.80) 6 0–13 −.30**
 Bedtime 138 8.65 (4.05) 10 0–14 −.24**
Self-initiated 138 7.75 (7.21) 6 0–39 −.06
Proximity-contingent 84 9.11 (8.78) 6 1–50 −.25*
Open in a new tab
a

Cumulative counts, including proximity contingent surveys; not all participants received proximity contingent surveys

*

p < .05

**

p < .01

Supportive Communication in Video Recordings

We used video recordings, inclusive of recordings completed after the baseline appointment and Day 1 of the intensive observational period, from three pilot families to develop and pilot our approach to coding youth support seeking and parent support enactment behaviors. The focal parents in all three dyads identified as mothers with a CESD score of 2 (Dyad 1), 10 (Dyad 2) and 9 (Dyad 3). The focal youth were a 12-year-old male with a CESD score of 7 (Dyad 1), a 14-year-old female with a CESD score of 18 (Dyad 2), and a 14-year-old female with a CESD score of 25 (Dyad 3). All three youth were non-Hispanic White, per parent report. We coded 93, 136 and 158 minutes of parent-youth interactions for Dyads 1, 2 and 3, respectively. See Table 2 for details.

Table 2.

Coding supportive communication episodes
Dyad 1 Dyad 2 Dyad 3
Count (%) Count (%) Count (%)
Total minutes of recording 93 min 136 min 158 min
Talk turns - total 1186 2030 1971
Talk turns - supportive communication 132 (11.1%) 232 (11.4%) 802 (40.7%)
Episodes - supportive communication 12 13 27
Youth support seeking behaviors
Tell 12 (100%) 13 (100%) 26 (96.3%)
Emotion 8 (66.7%) 5 (38.5%) 15 (55.6%)
Request 2 (16.7%) 2 (15.4%) 2 (7.4%)
Parent support enactment behaviors
Information 10 (83.3%) 11 (84.6%) 27 (100%)
Emotional 7 (58.3%) 1 (7.7%) 11 (40.7%)
Esteem 2 (16.7%) 2 (15.4%) 2 (7.4%)
Tangible 0 (0%) 3 (23.1%) 4 (14.8%)
Open in a new tab

Interrater reliability for the identification of supportive communication episodes was κ = 0.72. For Dyad 1, 132 out of 1186 (11.1%) talk turns were coded as being part of a supportive communication episode. The total number of supportive communication episodes was 12, the mean episode spanning over 11 talk turns. Dyad 2 showed a similar proportion of talk turns spent in supportive communication: 232 out of 2023 (11.4%) talk turns for 13 total talk turns. The mean number of talk turns in a supportive communication episode was 17.8. Dyad 3 spent 40.7% of their talk turns (i.e., 802 out of 1971) in 27 supportive communication episodes. The mean number of talk turns in a supportive communication episode was 29.7.

We relied on consensus coding procedures to continuously improve the reliability of our coding of youth support-seeking and parent support enactment behaviors. Each coder coded supportive communication episodes independently and then met weekly to reach group consensus. Disclosure of negative events and experiences was the most observed type of youth support seeking behavior, observed in all but one supportive communication episodes across the three dyads. The second most common form of support seeking was the expression of negative emotion, identified in 38.5% - 66.7% of supportive communication episodes. Requests for support were infrequent: 2 times (16.7%) in Dyad 1, 2 times (15.4%) in Dyad 2, 2 times (7.4%) in Dyad 3. Parents in all dyads frequently provided informational support, wherein parents asked questions to gather or elicit more information and/or gave advice, suggestions, opinions and perspectives: 10 times (83.3%) in Dyad 1, 11 times (84.6%) in Dyad 2, 27 times (100%) in Dyad 3. The proportion of episodes wherein the parent provided emotional support ranged from 7.7% to 58.3%. Esteem and tangible support were infrequently observed. Each parent showed esteem support in 2 of their supportive communication episodes. Tangible support was observed 0 (0%), 3 (23.1%) and 4 (14.8%) times in Dyads 1, 2 and 3 respectively.

Descriptives of EMA Compliance

As there were three different mechanisms by which participants could complete an EMA survey, we considered a participant to be fully compliant if they completed 40–56 total surveys throughout the two-week-long intensive longitudinal period. Specifically, during the baseline appointment, we encouraged the participants to complete 30 to 42 surveys during the day inclusive of the wake, afternoon and evening scheduled surveys, any proximity-contingent surveys and any self-initiated surveys, and 10 to 14 scheduled bedtime surveys. See Table 1.

To review, the average participant was prompted to complete 42 scheduled day time (wake, afternoon and evening) and 14 bedtime surveys during the two-week long observational period. In addition, participants could self-initiate an unlimited number of surveys, and receive up to 112 proximity-contingent survey alerts. The mean total number of EMA surveys was 46.9 (SD = 19.0) for parents, and 40.6 (SD = 18.7) for youth. Considering the lower threshold for full compliance (i.e., 40 total surveys submitted of any type), the average parent and youth were compliant. During the day, parents and youth started mean 31.6 (SD = 12.7) and 26.4 (SD = 11.7) surveys, respectively, inclusive of the wake, afternoon and evening scheduled surveys and self-initiated surveys but excluding proximity-contingent surveys. The mean number of bedtime surveys started was 9.8 (SD = 4.11) and 8.7 (SD = 4.05) for parents and youth, respectively.3 For the 112 dyads who could have received prompts to complete proximity-contingent surveys, only 78 parents and 84 youth received and started any (parents, M = 9.9, SD = 9.6; youth, M = 9.1, SD = 8.8). Many participants did not receive any proximity-contingent prompts due to a lack of qualifying interactions, technological issues, or procedure problems.

As shown in Table 1, youth depression symptoms were significantly correlated with 6 out of 8 indices of youth EMA compliance, such that youth with higher symptoms of depression completed fewer surveys. Specifically, depression symptoms were negatively correlated with the afternoon, evening and bedtime surveys, and the proximity-contingent surveys. Parent depression symptoms were correlated with only two out of eight indices of EMA compliance.

Parent-Youth Interactions Reported in EMAs and Estimated by BLE

As reported on Table 3, the average parent reported that they saw, communicated with or talked to their teen since the last survey they completed on 60–83% of their EMAs. The average youth reported that they saw, communicated with or talked to their parent on 56–71% of their EMAs. Split-half reliability scores at the between-person level were moderate. Repeated measures ANOVA showed that parents were least likely to report interacting with their youth on their morning survey in comparison to all other surveys, and more likely to report interacting with their youth on the bedtime survey in comparison to the self-initiated surveys F(5, 596) = 16.1, p < .001). Similarly, repeated measures ANOVA conducted on youth proportions showed that youth were least likely to interact with their parent on the morning surveys (F (5, 593) = 8.22, p < .001). Together, parents reported that they interacted with their teen on 4,665 out of 6,426 responses (72.6%), and youth, on 3,613 out of 5,557 responses (65.0%).

Table 3.

Frequency of parent and youth-reported interactions

Proportion of Interactions Split-half reliability Parent Depression Correlation Youth Depression Correlation
N M (SD) Median Range rho r r
Parent EMA
Scheduled
 Wake 132 .60 (.30) 0.67 0–1.00 .53*** −.20*
 Afternoon 133 .77 (.25) 0.80 0–1.00 .30*** −.08
 Evening 132 .76 (.29) 0.86 0–1.00 .59*** −.08
 Bedtime 136 .83 (.18) 0.90 0.11–1.00 .16 −.10
Self-initiated 128 .72 (.24) 0.75 0–1.00 .47*** −.12
Proximity-contingent 78 .79 (.23) 0.86 0–1.00 .16 .11
Youth EMA
Scheduled
 Wake 131 .56 (.34) 0.57 0–1.00 .53*** −.07
 Afternoon 132 .70 (.31) 0.75 0–1.00 .68*** −.05
 Evening 126 .68 (.31) 0.75 0–1.00 .45*** −.12
 Bedtime 136 .69 (.28) 0.75 0–1.00 .39*** −.02
Self-initiated 127 .67 (.30) 0.71 0–1.00 .45*** −.06
Proximity-contingent 84 .71 (.30) 0.76 0–1.00 .46*** −.21
Open in a new tab
*

p < .05

Parent depression symptoms were correlated with the frequency of parent-reported interactions such that parents with more depression symptoms endorsed parent-youth interactions less frequently on the morning surveys (r = −.20).

Last, we examined direct proximity data from the phone from a subset of families (N=39) to objectively examine the frequency of interactions. Across the 39 families examined, we observed on average 38.2 qualifying interactions (SD=23.9) between 6AM and 12 AM during the 14 days of the study. Most qualifying interactions were not prompted because they followed closely after one another, suggesting that many of these interactions may have been longer times spent together with short breaks, although others did not prompt due to proximity to the car sensor or recency of other EMA questionnaires.

Discussion

Intensive longitudinal methods (ILMs) are essential for assessing risk and protective processes in the development of psychopathology. Family and peer relationships are highly influential to the developmental trajectory of adolescent depression, and advancements in ILMs can help to capture protective or stressful interactions in the immediate contexts when and where they occur. The current study advances ILMs by describing procedures for conducting ecological behavioral observations of parent-youth interactions in naturalistic settings and using Bluetooth low energy signals to identify parent-youth interactions exactly when they occur. We found that naturalistic video recording procedures were feasible, with 126 out of 138 families having at least some codable video recordings. Across three pilot families, we identified a total of 52 distinct supportive communication episodes, wherein the youth was seeking, and the parent was enacting support. Complex EMA procedures, that integrate scheduled, self-initiated and proximity-contingent surveys, were also feasible, with the average parent and youth showing compliance with the minimum required EMAs. Parents and youth reported that they interacted with each other in the majority of the EMAs; at the survey level, our EMA procedures captured 4,665 distinct parent-youth interactions per parent report, and 3,613 interactions, per youth report. The study demonstrates the feasibility of implementing the described innovative strategies to better contextualize assessments of interpersonal interactions in daily life.

Ecological Behavioral Observations of Parent-Youth Interactions in Daily Life

Perhaps the most novel task requested of the study participants was to turn on a video recording device that would record family interactions whenever the participants were in the car together. Preliminary and unpublished qualitative data from focus groups and our community advisory board indicated that the car was an ideal location for the video recordings, as recordings in this setting are less invasive than in the home and youth tend to disclose information about their days in the car. We consented 138 families into the study, 126 provided eligible and codable video recordings. Parents who had more depression symptoms completed fewer minutes of video recordings. To further reduce participant burden, which may be higher among individuals with mental health problems, future studies should consider leveraging BLE or other inputs to automate video recording procedures. Nonetheless, supporting the feasibility of obtaining ecological behavioral observations of parent-child interactions in daily life, this approach adds to existing passive sensing procedures (Brick et al., 2020; Langener et al., 2023; Lind et al., 2023; Nelson & Allen, 2018), which to date primarily capture individual behaviors, and electronically activated recordings of ambient sound (Mehl, 2017).

The analysis for the current study emphasized verbal and acoustic information, including tone of voice, over information that was only available via video (e.g., facial expression), because the quality of the video varied significantly between families. Timing of the video recordings, inaccurate placement of the camera, and seated location of the target child interfered led to inconsistent video quality. Nonetheless, video was critical for identifying speakers, and remains a rich source of data for future research studies, as barriers to obtaining consistently high video recordings could be easily remediated with staff-led installation of the equipment and technological improvements in the field of view and nighttime recording quality. At the same time, a benefit of recording audio only may be that participants find audio recording less conspicuous and less burdensome. Some participants may also find audio recording procedures less intrusive than video, although in the current study, the restriction of video recordings to a particular setting (e.g., car) helped to reduce the sense of intrusion, irrespective of recording modality. A consideration of such trade-offs is critical to ensuring that research procedures are equitable, and that research data are representative of diverse families.

The intent of these study procedures was to capture spontaneous episodes of supportive communication between youth and parents, wherein youth are seeking, and parents are enacting social support. Observational methods that enable the objective assessment of behaviors are critical because of the importance of social support on preventing adolescent depression (Heerde & Hemphill, 2018; Pössel et al., 2018; Rueger et al., 2016), and respondent bias inherent to many self-report assessments of social support in individuals with elevated depression symptoms. In our three pilot families, we identified 52 unique episodes of supportive communication with an acceptable level of reliability. Youth disclosure behaviors, wherein they told the parent about a negative situation or an experience, were the most common. Request behaviors, when youth directly ask for help, advice or assistance, were the least common strategies adolescents used to seek support. In response, parents most often enacted informational support, defined by eliciting more information or sharing relevant experiences, information and/or advice. Existing research documents that informational support, perhaps especially in the form of advice, is a particularly common type of support (MacGeorge, 2009). In fact, some research describes advice as being nearly ubiquitous in supportive interactions (Feng, 2014; Goldsmith & Fitch, 1997). Beyond informational support, emotional support is also common in both in-person and online exchanges (Rains et al., 2015), though emotional support was notably less common in these data. It remains to be seen whether emotional support is more common in other parent-adolescent dyads or if something about the context diminished the communication of emotional support. Although findings are preliminary, this study shows how this naturalistic approach can reveal the specific ways in which youth and parents spontaneously seek and provide social support. Further, the video data also provide opportunities for future theory- and data-driven exploration of various family related risk and protective processes for adolescent depression. Importantly, the naturalistic approach to data collection enables examinations of competing theories in future research. For example, research on capitalization suggests that partner responses to disclosures of positive events may benefit the discloser in a manner similar to social support (Shorey & Lakey, 2011). The naturalistic observational data allows a future examination of disclosures of positive and negative experiences, to inform competing theories about social support.

EMAs of Parent-Youth Interactions in Daily Life

To better capture parent-youth interactions when and where they occur, we used BLE signals emitted and detected by smartphones carried by the parent and youth to assess physical proximity and deploy proximity-contingent EMAs. These EMAs complemented scheduled and self-initiated surveys, all of which assessed parent-youth interactions since the last survey. Although cumulative EMA compliance was high, we found that youth with more depression symptoms completed fewer prompted surveys, including proximity-contingent surveys, but not self-initiated surveys. Past studies have explored differences in EMA compliance rates by clinical characteristics of the sample without specificity (Wen et al., 2017). Depression symptoms are of particular interest because their symptoms may specifically reduce the motivation or ability to complete EMAs when prompted. Our findings suggest that missingness in EMAs conducted in some clinical samples is likely not random (Li et al., 2024), and that youth with depression may benefit from more concrete and immediate reinforcement structures or parental support to support youth complying with daily responsibilities, including study procedures.

Parents and youth reported that they saw or interacted with their parent on the majority of their EMAs, suggesting that there are ample opportunities for mutual dyadic influence despite the emphasis on growing levels of autonomy and increasing salience of peers during adolescence (Andrews et al., 2021). However, the likelihood of interacting may vary significantly within individuals, contributing to moderate levels of between-person reliability. We observed few correlations between depression symptoms and frequencies of youth- and parent-reported interactions, likely due to parents and youth living in the same house. The notion that parents and adolescents are seeing or talking to each other frequently was further confirmed by our BLE data. Of note, the precise number of interactions was difficult to mark for certain, with parent and child phones reporting large differences. Most likely, these differences arise because the phone’s operating system slowed (“throttled”) or stopped BLE monitoring to save battery power or differences in radio sensitivity across different environments. Although a thorough discussion of these issues is outside the scope of this paper, this type of technological challenge is important to note for others engaging in this research. Despite the need for more research, this study is among the first to demonstrate that BLE signals can be applied to estimate dyadic interactions. BLE monitoring can be improved by first testing and fine-tuning algorithms in a controlled and observed setting wherein the number of actual interactions between research participants are fixed. Alongside this testing, recording additional metadata on BLEs and EMAs, including instances when participants were invited to but did not start a BLE-activated EMA, is essential to improving this methodology. Upon further refinement, this innovation can improve the assessment of protective or risky interpersonal interactions in daily life and facilitate the harmonization of multi-informant perspectives in studies of interpersonal interactions.

Constraints on Generalizability and Limitations

This study has several additional limitations. First, the sample is not culturally or socioeconomically diverse, limiting the generalizability of the study findings. Study procedures may be less acceptable to members of groups typically under-represented in research, or individuals who may not have time to complete the requested tasks. However, increasing participant control and minimizing participant burden – two priorities in the study design – are key strategies to make research more inclusive of diverse samples (Chapman-Davis et al., 2023; La Scala et al., 2023; Yancey et al., 2006). Relatedly, study procedures required that participants have access to a car that they drive at least two times a week, excluding families with minimal economic resources or without a need for a car. These inclusionary criteria and the exclusion of families with specific mental illness additionally reduces the generalizability of the findings. These limitations can be addressed in future studies by further leaning on community-engaged practices to increase participation of members from groups under-represented in ILM research. Second, naturalistic video observations were limited to times when the participants were in cars and relied on participants to turn on and off the recording device. The study procedures likely missed important supportive communication episodes that occurred elsewhere, such as the home, or out and about, and thus are not inclusive of all contexts. While the study procedures offered participants control over when recordings took place, this increased participant control means that the recordings we obtained are subject to self-censorship. For example, participants may strive to present their best selves, or not conduct recordings when they are experiencing strong negative emotions. However, such censorship is likely less extreme than is possible on self-report surveys. Third, participants were instructed to carry around study specific phones, and BLE signals emitted from these devices approximated participant’s locations relative to that of their partners. Participants were likely not always carrying around the study phones, which could have contributed to error in the estimation of their proximity, and delays in their receipt of EMA prompts; we do know that participants carried their phones frequently as the rates of responding and interaction tracking were high. Future studies can utilize more lightweight BLE devices (e.g., AirTag-like trackers, smartwatches) that participants can more easily carry to precisely approximate location with minimal burden.

ILM studies have relied on self-reports through EMAs, and only recently incorporated objective assessments of individual behaviors. However, to improve the rigor of ILM research, future studies should consider multi-method and multi-informant designs: essential aspects of rigorous traditional longitudinal studies of developmental psychopathology. Despite some limitations, the study demonstrates the feasibility of improving intensive longitudinal assessments of interpersonal interactions in two ways. First, we demonstrated the feasibility of obtaining ecological behavioral observations of interactions in daily life, alongside self-report EMAs. Second, we used BLE signals to estimate when an interaction occurred. These advancements are essential for improving the assessment of interpersonal interactions, which are key sources of protection and risk for adolescent depression and other psychopathology.

Supplementary Material

1

Public Significance Statement.

This study describes methodological advancements that are important for improving how we assess interpersonal interactions – key sources of protection and risk for adolescent depression. First, findings suggest that we can supplement information from self-reports with objective observations conducted using unobtrusive recording devices. Second, findings show that we can use Bluetooth Low Energy signals to estimate when an interaction occurs and immediately gather multiple perspectives.

Funding:

This work was supported by National Institute of Mental Health R21MH125046; Penn State University Social Science Research Institute Level 2 Grant; and Society for Research on Child Development Small Grants for Early Career Scholars.

Footnotes

1

At the time of study enrollment, the passive monitoring of proximity via BLE signals was only compatible on the Android smartphones, which was carried by less than 10% of our sample. Given the low rate of Android ownership in our sample and for consistency, all participants were loaned Android smartphones (Pixel 2) for the study.

2

We retained video recordings from the 2 families who only had available recordings from the back camera that recorded only the rear seat passengers

3

On average, parents completed 55% (SD=25%), and youth completed 49% (SD = 25%) of all possible scheduled surveys. This is an underestimate of compliance, as individuals did not get alerted for a scheduled surveys if they had completed a proximity survey within the past 1 hour. Further, some individuals likely compensated for missed scheduled surveys by self-initiating a survey.

References

  1. Alexander JD, Linkersdörfer J, Toda-Thorne K, Sullivan RM, Cummins KM, Tomko RL, Allen NB, Bagot KS, Baker FC, Fuemmeler BF, Hoffman EA, Kiss O, Mason MJ, Nguyen-Louie TT, Tapert SF, Smith CJ, Squeglia LM, & Wade NE (2024). Passively sensing smartphone use in teens with rates of use by sex and across operating systems. Scientific Reports, 14(1), 17982. 10.1038/s41598-024-68467-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Andrews JL, Ahmed SP, & Blakemore S-J (2021). Navigating the Social Environment in Adolescence: The Role of Social Brain Development. Biological Psychiatry, 89(2), 109–118. 10.1016/j.biopsych.2020.09.012 [DOI] [PubMed] [Google Scholar]
  3. Bai S (2025). Families Being Supportive Together: A Multi-Method and Multi-Informant Intensive Longitudinal Study of Family Protective Mechanisms for Adolescent Depression. (ID 2964). NIMH Data Archive. 10.15154/5z08-gq64 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bai S, Robles TF, Reynolds BM, & Repetti RL (2020). Daily Mood Reactivity to Stress during Childhood Predicts Internalizing Problems Three Years Later. Journal of Abnormal Child Psychology, 48(8), 1063–1075. 10.1007/s10802-020-00650-7 [DOI] [PubMed] [Google Scholar]
  5. Barbee AP, & Cunningham MR (1995). An Experimental Approach to Social Support Communications: Interactive Coping in Close Relationships. Annals of the International Communication Association, 18(1), 381–413. 10.1080/23808985.1995.11678921 [DOI] [Google Scholar]
  6. Bitran AM, Sritharan A, Trivedi E, Helgren F, Buchanan SN, Durham K, Li LY, Funkhouser CJ, Allen NB, Shankman SA, Auerbach RP, & Pagliaccio D (2024). The effects of family support and smartphone-derived homestay on daily mood and depression among sexual and gender minority adolescents. Journal of Psychopathology and Clinical Science, 133(5), 358–367. 10.1037/abn0000917 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Brick TR, Mundie J, Weaver J, Fraleigh R, & Oravecz Z (2020). Low-Burden Mobile Monitoring, Intervention, and Real-Time Analysis Using the Wear-IT Framework: Example and Usability Study. JMIR Formative Research, 4(6), e16072. 10.2196/16072 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Chapman-Davis E, Webster EM, & Hines JF (2023). Achieving diversity in clinical trial enrollment by reducing burden and increasing value: A patient-centered approach. Gynecologic Oncology, 172, A1–A2. 10.1016/j.ygyno.2023.04.020 [DOI] [PubMed] [Google Scholar]
  9. Copeland WE, Alaie I, Jonsson U, & Shanahan L (2021). Associations of Childhood and Adolescent Depression With Adult Psychiatric and Functional Outcomes. Journal of the American Academy of Child & Adolescent Psychiatry, 60(5), 604–611. 10.1016/j.jaac.2020.07.895 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Crowley JL, & High AC (2020). Validating the support gaps framework: Longitudinal effects and moderators of experiencing deficits and surpluses during supportive interactions. Communication Quarterly, 68(1), 29–53. 10.1080/01463373.2019.1668447 [DOI] [Google Scholar]
  11. Do QB, McKone KMP, Hamilton JL, Stone LB, Ladouceur CD, & Silk JS (2023). The link between adolescent girls’ interpersonal emotion regulation with parents and peers and depressive symptoms: A real-time investigation. Development and Psychopathology, 1–15. 10.1017/S0954579423001359 [DOI] [PubMed] [Google Scholar]
  12. Feng B (2014). When Should Advice Be Given? Assessing the Role of Sequential Placement of Advice in Supportive Interactions in Two Cultures. Communication Research, 41(7), 913–934. 10.1177/0093650212456203 [DOI] [Google Scholar]
  13. Fiedler J, Seiferth C, Eckert T, Woll A, & Wunsch K (2023). A just-in-time adaptive intervention to enhance physical activity in the SMARTFAMILY2.0 trial. Sport, Exercise, and Performance Psychology, 12(1), 43–57. 10.1037/spy0000311 [DOI] [Google Scholar]
  14. Funkhouser CJ, Trivedi E, Li LY, Helgren F, Zhang E, Sritharan A, Cherner RA, Pagliaccio D, Durham K, Kyler M, Tse TC, Buchanan SN, Allen NB, Shankman SA, & Auerbach RP (2024). Detecting adolescent depression through passive monitoring of linguistic markers in smartphone communication. Journal of Child Psychology and Psychiatry, 65(7), 932–941. 10.1111/jcpp.13931 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Gariépy G, Honkaniemi H, & Quesnel-Vallée A (2016). Social support and protection from depression: Systematic review of current findings in Western countries. British Journal of Psychiatry, 209(4), 284–293. 10.1192/bjp.bp.115.169094 [DOI] [PubMed] [Google Scholar]
  16. Goldsmith DJ, & Fitch K (1997). The Normative Context of Advice as Social Support. Human Communication Research, 23(4), 454–476. 10.1111/j.1468-2958.1997.tb00406.x [DOI] [Google Scholar]
  17. Hadiwijaya H, Klimstra TA, Vermunt JK, Branje SJT, & Meeus WHJ (2017). On the Development of Harmony, Turbulence, and Independence in Parent–Adolescent Relationships: A Five-Wave Longitudinal Study. Journal of Youth and Adolescence, 46(8), 1772–1788. 10.1007/s10964-016-0627-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Hamilton JL, Do QB, Choukas-Bradley S, Ladouceur CD, & Silk JS (2021). Where it Hurts the Most: Peer Interactions on Social Media and in Person are Differentially Associated with Emotional Reactivity and Sustained Affect Among Adolescent Girls. Research on Child and Adolescent Psychopathology, 49(2), 155–167. 10.1007/s10802-020-00725-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Heerde JA, & Hemphill SA (2018). Examination of associations between informal help-seeking behavior, social support, and adolescent psychosocial outcomes: A meta-analysis. Developmental Review, 47, 44–62. 10.1016/j.dr.2017.10.001 [DOI] [Google Scholar]
  20. Heron KE, Everhart RS, McHale SM, & Smyth JM (2017). Using Mobile-Technology-Based Ecological Momentary Assessment (EMA) Methods With Youth: A Systematic Review and Recommendations. Journal of Pediatric Psychology, 42(10), 1087–1107. 10.1093/jpepsy/jsx078 [DOI] [PubMed] [Google Scholar]
  21. La Scala S, Mullins JL, Firat RB, Emotional Learning Research Community Advisory Board, & Michalska, K. J. (2023). Equity, diversity, and inclusion in developmental neuroscience: Practical lessons from community-based participatory research. Frontiers in Integrative Neuroscience, 16, 1007249. 10.3389/fnint.2022.1007249 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Lam CB, McHale SM, & Crouter AC (2012). Parent-child shared time from middle childhood to late adolescence: Developmental course and adjustment correlates. Child Development, 83(6), 2089–2103. 10.1111/j.1467-8624.2012.01826.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Langener AM, Stulp G, Kas MJ, & Bringmann LF (2023). Capturing the Dynamics of the Social Environment Through Experience Sampling Methods, Passive Sensing, and Egocentric Networks: Scoping Review. JMIR Mental Health, 10, e42646. 10.2196/42646 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Li Y, Oravecz Z, Ji L, & Chow S-M (2024). Multiple Imputation with Factor Scores: A Practical Approach for Handling Simultaneous Missingness Across Items in Longitudinal Designs. Multivariate Behavioral Research, 1–29. 10.1080/00273171.2024.2371816 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lind MN, Kahn LE, Crowley R, Reed W, Wicks G, & Allen NB (2023). Reintroducing the Effortless Assessment Research System (EARS). JMIR Mental Health, 10, e38920. 10.2196/38920 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. MacGeorge EL (2009). Social Support. In Eadie W (Ed.), 21st Century Communication: A Reference Handbook (pp. 283–294). SAGE Publications, Inc. 10.4135/9781412964005.n32 [DOI] [Google Scholar]
  27. Mehl MR (2017). The Electronically Activated Recorder (EAR): A Method for the Naturalistic Observation of Daily Social Behavior. Current Directions in Psychological Science, 26(2), 184–190. 10.1177/0963721416680611 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Nelson BW, & Allen NB (2018). Extending the Passive-Sensing Toolbox: Using Smart-Home Technology in Psychological Science. Perspectives on Psychological Science, 13(6), 718–733. 10.1177/1745691618776008 [DOI] [PubMed] [Google Scholar]
  29. Pössel P, Burton SM, Cauley B, Sawyer MG, Spence SH, & Sheffield J (2018). Associations between Social Support from Family, Friends, and Teachers and depressive Symptoms in Adolescents. Journal of Youth and Adolescence, 47(2), 398–412. 10.1007/s10964-017-0712-6 [DOI] [PubMed] [Google Scholar]
  30. Radloff LS (1977). The CES-D Scale: A Self-Report Depression Scale for Research in the General Population. Applied Psychological Measurement, 1(3), 385–401. 10.1177/014662167700100306 [DOI] [Google Scholar]
  31. Radloff LS (1991). The use of the Center for Epidemiologic Studies Depression Scale in adolescents and young adults. Journal of Youth and Adolescence, 20(2), 149–166. 10.1007/BF01537606 [DOI] [PubMed] [Google Scholar]
  32. Rains SA, Peterson EB, & Wright KB (2015). Communicating Social Support in Computer-mediated Contexts: A Meta-analytic Review of Content Analyses Examining Support Messages Shared Online among Individuals Coping with Illness. Communication Monographs, 82(4), 403–430. 10.1080/03637751.2015.1019530 [DOI] [Google Scholar]
  33. Repetti R, Wang S, & Sears M (2013). Using direct observational methods to study the real lives of families: Advantages, complexities, and conceptual and practical considerations. In Gryzwacz J & Demerouti E (Eds.), New Frontiers in Work and Family Research (pp. 172–189). Psychology Press & Routledge. [Google Scholar]
  34. Rueger SY, Malecki CK, Pyun Y, Aycock C, & Coyle S (2016). A meta-analytic review of the association between perceived social support and depression in childhood and adolescence. Psychological Bulletin, 142(10), 1017–1067. 10.1037/bul0000058 [DOI] [PubMed] [Google Scholar]
  35. Russell MA, & Gajos JM (2020). Annual Research Review: Ecological momentary assessment studies in child psychology and psychiatry. Journal of Child Psychology and Psychiatry, 61(3), 376–394. 10.1111/jcpp.13204 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Schaible J, Oliverira M, Zens M, & Génois M (2022). Sensing close-range proximity for studying face-to-face interaction. In Quan-Haase A, Engel U, Liu SX, & Lyberg L (Eds.), Handbook of computational social science: Theory, case studies and ethics. Routledge, Taylor & Francis Group. [Google Scholar]
  37. Seidman AJ, George CJ, & Kovacs M (2022). Ecological momentary assessment of affect in depression-prone and control samples: Survey compliance and affective yield. Journal of Affective Disorders, 311, 63–68. 10.1016/j.jad.2022.05.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Sequeira L, Perrotta S, LaGrassa J, Merikangas K, Kreindler D, Kundur D, Courtney D, Szatmari P, Battaglia M, & Strauss J (2020). Mobile and wearable technology for monitoring depressive symptoms in children and adolescents: A scoping review. Journal of Affective Disorders, 265, 314–324. 10.1016/j.jad.2019.11.156 [DOI] [PubMed] [Google Scholar]
  39. Shorey RC, & Lakey B (2011). Perceived and Capitalization Support Are Substantially Similar: Implications for Social Support Theory. Personality and Social Psychology Bulletin, 37(8), 1068–1079. 10.1177/0146167211406507 [DOI] [PubMed] [Google Scholar]
  40. Substance Abuse and Mental Health Services Administration. (2022). Key substance use and mental health indicators in the United States: Results from the 2021 National Survey on Drug Use and Health (No. HHS Publication No. PEP22–07-01–005, NSDUH Series H-57; Rockville, MD: ). Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration. https://www.samhsa.gov/data/report/2021-nsduh-annual-national-report . [Google Scholar]
  41. Teti DM, Fronberg KM, Fanton H, & Crosby B (2022). Infant sleep arrangements, infant-parent sleep, and parenting during the first six months post-partum. Infant Behavior and Development, 69, 101756. 10.1016/j.infbeh.2022.101756 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Thapar A, Eyre O, Patel V, & Brent D (2022). Depression in young people. The Lancet, 400(10352), 617–631. 10.1016/S0140-6736(22)01012-1 [DOI] [PubMed] [Google Scholar]
  43. Trull TJ, & Ebner-Priemer U (2013). Ambulatory Assessment. Annual Review of Clinical Psychology, 9(1), 151–176. 10.1146/annurev-clinpsy-050212-185510 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Tsai KM, Telzer EH, & Fuligni AJ (2013). Continuity and Discontinuity in Perceptions of Family Relationships From Adolescence to Young Adulthood. Child Development, 84(2), 471–484. 10.1111/j.1467-8624.2012.01858.x [DOI] [PubMed] [Google Scholar]
  45. Twenge JM, Spitzberg BH, & Campbell WK (2019). Less in-person social interaction with peers among U.S. adolescents in the 21st century and links to loneliness. Journal of Social and Personal Relationships, 36(6), 1892–1913. 10.1177/0265407519836170 [DOI] [Google Scholar]
  46. Wen CKF, Schneider S, Stone AA, & Spruijt-Metz D (2017). Compliance With Mobile Ecological Momentary Assessment Protocols in Children and Adolescents: A Systematic Review and Meta-Analysis. Journal of Medical Internet Research, 19(4), e132. 10.2196/jmir.6641 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Yancey AK, Ortega AN, & Kumanyika SK (2006). EFFECTIVE RECRUITMENT AND RETENTION OF MINORITY RESEARCH PARTICIPANTS. Annual Review of Public Health, 27(1), 1–28. 10.1146/annurev.publhealth.27.021405.102113 [DOI] [PubMed] [Google Scholar]
  48. Yang M-J, Sutton SK, Hernandez LM, Jones SR, Wetter DW, Kumar S, & Vinci C (2023). A Just-In-Time Adaptive intervention (JITAI) for smoking cessation: Feasibility and acceptability findings. Addictive Behaviors, 136, 107467. 10.1016/j.addbeh.2022.107467 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Young D, & Radius Networks. (2023). Android Beacon Library. https://altbeacon.github.io/android-beacon-library/ [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS2115741-supplement-1.pdf (308.5KB, pdf)

RESOURCES