Table 1 Characteristics of the included studies

Web vs paper-and-pencil administration (n = 9 studies in 10 publications)

 Denniston 2010

Eaton 2010

USA

Cluster RCT

[31, 32]

85 schools in 15 states

5786 students in 9th- or 10th-grade

Mode 1: paper-and-pencil administration (PAPI) (n = 1729)

Mode 2: Web (n = 4057)

2.1 In-class Web without skips (students could skip a question if they did not want to answer it) – used school’s computer lab (n = 1735)

2.2. In-class Web with skips– used school’s stationary computer lab (n = 1763)

2.3. On your own Web – school’s computer lab (n = 559)

Youth Risk Behavior Surveys (YRBS) relevant survey questions:

(a) Sadness—felt sad or hopeless almost every day for 2 or more weeks in a row

(b) Suicide attempt—attempted suicide during the last 12 months

• Response rates

• Mental health variations by survey mode

 Hamann 2016

Switzerland

RCT

[33]

Students in 2 classes in a secondary school

Mode 1: PAPI (n = 28)

Mode 2: Web (n = 28)

German version of Children’s Depression Inventory (CDI) and the Spence Children’s Anxiety Scale (SCAS)

In both tools, higher scores represent worse anxiety/depression

• Mental health variations by survey mode

 Lygidakis 2010

Italy

RCT

[14]

190 adolescents from 3 seniors’ high schools

Mode 1: PAPI (n = 97)

Mode 2: Web (n = 93)

Ad hoc questionnaire based on the European School Survey Project on Alcohol and Other Drugs questionnaire (2003) and the fifth Doxa national survey

• Mental health variations by survey mode

 Mauz 2018

Germany

RCT

[30]

Children and adolescents from 20 municipalities in 5 federal states

11,140 randomized / 1194 adolescents analysed (4662 including parents)

Adolescents (11–17 years old)

Mode 1: PAPI (n = 895)

Mode 2: Web (n = 299)

Survey design:

Single mode design (n = 344 adolescents)

Sequential mixed-mode design (n = 269 adolescents) (Web = 148, 55% of participants)

Concurrent mixed-mode design (n = 290 adolescents) (Web = 43, 15% of participants)

Preselect mixed-mode design (n = 292 adolescents) (Web = 102, 35% of participants)

The German Health Interview and Examination Survey for Children and Adolescents (KiGGS)

The Strengths and Difficulties Questionnaire (SDQ)

• Response rates

• Mental health variations by survey mode

 Miech 2021

USA

RCT

[34]

41,866 students in 8th, 9th, and 12th grades from 397 middle or high schools in 48 states

Mode 1: PAPI (n = 20,039)

Mode 2: Web (n = 21,475)

Monitoring the Future (MTF) school-based national survey

• Response rates

• Mental health variations by survey mode

 Raat 2007

The Netherlands

RCT

[35]

Adolescents (13–17 years old) from 55 classes at various educational levels in seven secondary schools in rural and urban areas

Mode 1: PAPI (n = 458)

Mode 2: Web (n = 475)

The Child Health Questionnaire Child Form (CHQ-CF): has 16 items (higher scores mean better mental health; e.g., children feel peaceful, happy, and calm all the time)

• Mental health variations by survey mode

 Raghupathy 2013

USA

RCT

[36]

Adolescents from a predominantly rural school district in Washington State

Mode 1: PAPI (n = 181)

Mode 2: Web (n = 160)

Youth Risk Behavior Surveys (YRBS) survey: 17 items on alcohol use and risk indicator measures including short and long-term use, binge drinking, and perception of harm

• Mental health variations by survey mode

• Participant variations by survey mode

 Trapl 2007

USA

RCT

[37]

275 seventh grade students from Seven Cleveland Municipal School District (CMSD) K-8 schools

Mode 1: PAPI (n = 90)

Mode 2: Web (n = 185)

2.1 Personal Digital Assistant (PDA) (n = 93)

2.2 Audio-enhanced personal digital assistant (APDA) (n = 92)

Ad-hoc questionnaire, 178-question survey

• Response rate (assumed 100%)

• Mental health variations by survey mode

 van de Looij-Jansen 2008

The Netherlands

RCT

[16]

551 students in third grade classes from 5 secondary schools

Mode 1: PAPI (n = 261)

Mode 2: Web (n = 270)

The Dutch self-report version of the Strengths and Difficulties Questionnaire (SDQ): sum score 0–10; higher scores mean worse mental health

The Child Health Questionnaire (CHQ). Nine items about feelings and moods

• Response rates

• Mental health variations by survey mode

• Participant variations by survey mode

Telephone interview vs postal questionnaires (n = 2)

 Erhart 2009

Germany

RCT

[29]

42 municipalities

1737 children and their parents

Mode 1: Telephone (n = 825)

Mode 2: Postal/mail survey (n = 912)

Strengths and Difficulties Questionnaire (SDQ):

total score, higher scores indicate worse mental health, emotional symptoms, conduct problems, hyperactivity, peer relationship problems, and prosocial behaviour: scores 1 to 4 added together to generate a total difficulties score (based on 20 items)

• Response rates

• Mental health variations by survey mode

 Wettergren 2011

Sweden

RCT

[28]

585 adolescents from three public healthcare regions (South, Uppsala / Orebro (Middle) and North)

Mode 1: Telephone (n = 300)

Mode 2: Postal/mail survey, PAPI (n = 285)

Short Form 36 (SF-36):

mental health component scores, higher scores indicate better mental health

The Hospital Anxiety and Depression Scale (HADS): a fourteen-item scale with seven items that relate to anxiety and seven that relate to depression; Lower scores indicate better mental health status

• Response rates

• Mental health variations by survey mode

• Population variations by survey mode

Active vs Passive parental consent (n = 1)

 Courser 2009

USA

RCT cluster

[38]

14 school districts in Kentucky

Students in grades 6th through 12^th

Mode 1: Active consent (n = NR)

Mode 2: Passive consent (n = NR)

Kentucky Youth Outcomes (KYOS) Survey. Measure of risk and protective factors, behaviours related to alcohol, tobacco, and other drugs, and school safety issues

• Response rates

• Mental health variations by survey mode

• Population variations by mode

Web first vs in person first interview (n = 1)

 McMorris 2009

USA

RCT

[27]

10 suburban public elementary schools in the Pacific Northwest school district

386 students enrolled in 1^st and 2^nd grade and followed up every spring

Mode 1: In-person first (n = 189)

Mode 2: Web first (n = 197)

Raising Healthy Children (RHC) Project: 14 items on substance use and sexual risk behaviour

• Response rates

• Mental health variations by survey mode

Voucher vs non-voucher (n = 1 reported in 2 documents)

 Pejtersen 2020

Denmark

RCT

[26, 39]

262 children and youth (vulnerable and lonely) from 10 municipalities

Postal questionnaire

Mode 1: Voucher (n = 143)—supermarket voucher worth €15

Mode 2: Non-voucher (n = 119)

Ad-hoc questionnaire on participants’ life situation: family and housing; education and training; sport and leisure time; relation to friends; drug use; and strengths and difficulties

The Danish version of the Strengths and Difficulties Questionnaire (SDQ): emotional symptoms subscale (sum score 0–10; higher scores mean worse mental health)

• Response rates

• Mental health variations by survey mode

Internal supervision vs external supervision (n = 1)

 Walser 2012

Switzerland

Cluster RCT

[40]

80 classes in public high schools from the Canton of St. Gallen

Mode 1. External supervision (n = 40 classes, 598 students)

Mode 2. Internal supervision (n = 40 classes, 599 students)

Computer-assisted-self-interviewing (CASI) using a program from NETQ

• Mental health variations by survey mode