Grade repetition and bullying victimization in adolescents: A global cross-sectional study of the Program for International Student Assessment (PISA) data from 2018.

BACKGROUND: Grade repetition is practiced worldwide and varies considerably across the globe. Globally, around 32.2 million students repeated a grade at the primary education level in 2010. Although a large body of research has documented grade repetition's academic and non-academic effects, the limited evidence on associations between grade repetition and school bullying is inconsistent and ambiguous. This study aimed to investigate the global association of grade repetition with bullying victimization in a large-scale school-based cross-sectional study. METHODS AND
FINDINGS: We used the latest global data from the Program for International Student Assessment (PISA) 2018. PISA 2018 was conducted between March and August 2018 in 80 countries and economies among students aged 15-16 years attending secondary education. The students reported their experiences of repeating a grade at any time point before the survey and of being bullied in the past 12 months. The outcome measures were 6 types of bullying victimization. We accounted for the complex survey design and used multivariate logistic regression models to estimate the odds ratios (ORs) with 95% confidence intervals (CIs) of grade repetition with bullying victimization after adjusting for potential confounders (sex; age group; migrant status; school type; economic, social, and cultural status; and parental emotional support). This study included 465,146 students (234,218 girls and 230,928 boys) with complete data on grade repetition and bullying victimization in 74 countries and economies. The lifetime prevalence of grade repetition was 12.26%, and 30.32% of students experienced bullying at least a few times a month during the past 12 months. Grade repetition was statistically significantly associated with each type of bullying victimization. The OR (95% CI) of overall bullying victimization for grade repeaters compared with their promoted peers was 1.42 (95% CI 1.32-1.52, p < 0.001). The sex-specific analysis produced similar results in both boys and girls. Furthermore, girls who repeated a grade had higher risks of being made fun of, being threatened, having possessions taken away, and being pushed around than boys. The major limitation is that this study only included students attending schools and therefore may be subject to possible selection bias. In addition, the cross-sectional design hinders us from establishing causality between grade repetition and bullying victimization.
CONCLUSIONS: In this study, we observed that, globally, both boys and girls who repeat a grade are at increased risk of being bullied compared with promoted peers, but girls may experience higher risks than boys of specific types of bullying associated with repeating a grade. These findings provide evidence for the association of grade repetition with bullying victimization. Sex differences in risk of experiencing some types of bullying suggest that tailored interventions for girls who repeat a grade may be warranted.

Introduction

Grade repetition (or grade retention), the practice of having students repeat a grade for an additional school year when they fail to meet required benchmarks, is relatively common in some countries (e.g., Belgium) and is discouraged or banned outright in others (e.g., Norway) [1]. According to data collected by the Program for International Student Assessment (PISA) in 2009, over 10% of 15-year-olds in 30 countries and economies had repeated a grade at least once [2]. Globally, 32.2 million students repeated a grade at the primary education level in 2010 [3]. Grade repetition and its opposite (social promotion) are controversial issues in many countries. The former is viewed as costly [3], with unclear benefits and well-known drawbacks [4-8], and a common misperception is that repeating a grade allows the student to grow academically and socially [9].

Moreover, the practice of grade repetition can have various negative consequences. A US study found that grade repetition was independently associated with increased risk of behavior problems among white children and adolescents [10]. Despite a growing evidence base that links grade repetition to adverse behavioral outcomes [11-13], the literature includes mostly small observational studies of specific behavioral outcomes, such as bullying victimization, and provides inconsistent results [14,15]. One study found no differences in bullying victimization between socially promoted and retained students in public schools [14]. Another found that students who were old for their grade, because of grade retention or delayed school entry, were more likely to be bullied physically, verbally, and socially than their peers who were age appropriate for their grade [15].

Preventing school violence in all its forms is a fundamental human rights issue. Bullying victimization—individuals being repeatedly exposed to intentional harmful behaviors by peers and unable to defend themselves because of a power imbalance [16]—is a severe global public health problem that compromises learners’ right to education. Approximately 130 million 13- to 15-year-olds worldwide experience bullying [17]. The prevalence of bullying victimization varies widely among countries; however, one study of adolescents in 83 countries estimated an overall prevalence of 30.5% in the past 30 days [18]. School bullying has short- and long-term adverse consequences, including physical, cognitive, and mental health issues for the victimized students [19] and longstanding economic impacts for the victimized students, their families, and society [20].

The social–ecological framework is useful for understanding the factors that contribute to grade repetition and related bullying victimization [21]. This framework considers the complex interplay between individual, family, school, and societal factors. One potential pathway that links grade repetition to bullying victimization is poor academic performance [14,16]. Academic problems are the major reason for students to repeat a grade [22] and increase the risk of bullying because struggling with academic work could be viewed as deviant behavior by promoted students [14]. The social stigma attached to repeating a grade may also play an important role in bullying in retained subgroups [23,24], and the degree of stigma varies significantly across cultures.

Addressing school bullying is a global policy priority and essential to achieving the Sustainable Development Goals (SDGs), in particular SDG 4.a.2, to provide safe, non-violent, inclusive, and effective learning environments for all, and SDG 16.2, to end all forms of violence against children, including bullying [25]. To date, the association of grade repetition and bullying victimization has not been closely examined in global samples of adolescents, although many studies have estimated the prevalence of grade repetition and bullying victimization separately [2,18]. This study aimed to address the gap by providing a comprehensive overview of the global prevalence of grade repetition and bullying victimization and examining the association of grade repetition with bullying victimization using the latest PISA data. We hypothesized that grade repetition would be associated with bullying victimization in both boys and girls, and that the association would be stronger in girls because grade retention is socially more stigmatizing for girls [26].

Methods

Study design and participants

We used the data from the 2018 survey cycle of PISA. The assessment is conducted by the Organisation for Economic Co-operation and Development (OECD) every 3 years to assess the skills and knowledge of 15-year-old students around the world in reading, mathematics, and science. The PISA target population is all students attending educational institutions in grade 7 or higher, aged 15 years 3 months to 16 years 2 months, with a 1-month variation of this age window at the beginning of each assessment period. Given that 15-year-olds make up the majority of the population, we follow the convention of using “15-year-old students” as shorthand for the PISA target population.

PISA 2018 was conducted between 1 March 2018 and 31 August 2018 in 80 countries and economies worldwide. A 2-stage sampling procedure was adopted within most countries/economies. In the first stage, schools were sampled systematically with probabilities proportional to the estimated size of their 15-year-old student population from a comprehensive national list of all PISA-eligible schools. At least 150 schools were selected within each country/economy. In the second stage, 42 15-year-old students were selected with equal probability from each school. If fewer than 42 students were available at a school, all 15-year-old students in the school were sampled. The number of sampled students could not fall below 20 [27]. PISA’s international protocol requires a minimum weighted response rate of 85% of sampled schools within participating countries/economies and 80% of students within selected schools. A detailed description of sampling procedures, data collection methodology, quality assurance, and data access is available on the official website [28]. Each participating student had 2 hours to complete the tests and about 35 minutes to answer a background questionnaire. In total, 612,002 students completed the PISA 2018 assessments, representing about 32 million 15-year-old students in the participating 80 countries/economies [27].

We used the PISA 2018 dataset in the present study because it is the most recent publicly accessible PISA dataset that measured both grade repetition and bullying victimization. We merged the PISA 2018 school dataset and student dataset and included all students (n = 612,002) who participated in PISA 2018, as shown in Fig 1. All participants in Israel, Lebanon, and North Macedonia were excluded because the 3 countries did not measure school bullying (n = 17,806). All participants in Japan, Malaysia, and Norway were excluded because these 3 countries do not allow grade repetition, and no students reported being retained (n = 18,033). Lastly, the participants without data on grade repetition or school bullying (n = 111,017) were excluded from the analysis. The final sample consisted of 465,146 students, including 234,218 girls and 230,928 boys, in 74 countries/economies worldwide. The list of participating countries/economies and the corresponding sample sizes are provided in S1 Table.

Fig 1

Participant flow chart.

The present study followed the reporting guidelines of Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) for cross-sectional studies (S2 Table). This study was exempt from institutional review board approval because the data used in this secondary analysis are de-identified and publicly available [28].

Exposure measurement

Given that education systems vary considerably across countries, PISA followed the widely used International Standard Classification of Education (ISCED) [29] to facilitate comparisons of education systems across participating countries and economies.

Standardized questionnaires asked students 3 questions about their experience with grade repetition: (1) “Have you ever repeated a grade at ISCED 1 (the primary education level)?”; (2) “Have you ever repeated a grade at ISCED 2 (the lower secondary education level)?”; and (3) “Have you ever repeated a grade at ISCED 3 (the upper secondary education level)?” The response options were “No, never”, “Yes, once”, or “Yes, twice or more” for each of these 3 questions.

Participants were categorized as grade repeaters if they answered “yes” to any of these 3 questions and were categorized as grade non-repeaters (promoted students) if they did not report the experience of repeating a grade at school before the survey.

Outcome measurement

PISA collected data on students’ exposures to school bullying with the question “During the past 12 months, how often have you had the following experiences in school? (Some experiences can also happen in social media).” The listed experiences were as follows: (1) “Other students left me out of things on purpose”; (2) “Other students made fun of me”; (3) “I was threatened by other students”; (4) “Other students took away or destroyed things that belong to me”; (5) “I got hit or pushed around by other students”; and (6) “Other students spread nasty rumors about me” [30]. The response options were “0 = Never or almost never”; “1 = A few times a year”; “2 = A few times a month”; or “3 = Once a week or more” (Cronbach’s alpha = 0.88 in our sample).

Following the cutoff point recommended by previous studies [31,32], for each of the 6 individual bullying experiences mentioned above, we combined the response options for “0 = Never or almost never” and “1 = A few times a year” into 0 (a few times a year or less frequently) and combined the response options for “2 = A few times a month” and “3 = Once a week or more” into 1 (a few times a month or more frequently).

We then combined the 6 individual bullying experiences into a single indicator “any type of bullying victimization” and coded it as 0 (never or rare on all types) if the participants reported “0 = Never or almost never” or “1 = A few times a year” to all of the 6 experiences and 1 (any involvement) if they reported “2 = A few times a month” or “3 = Once a week or more” to any of the 6 experiences.

Covariates

The ecological system framework was used to identify potential confounders of the association between grade repetition or promotion and bullying. The covariates included sex; age group; migrant status; school type; economic, social, and cultural status (ESCS); and parental emotional support.

Students in PISA 2018 were asked the question “Are you female or male?” We generated a dichotomous variable “sex” with 1 = female and 2 = male based on this question.

We calculated student age based on the assessment date and birthday provided by the participating students [33]. As suggested by a previous PISA study [2], we generated a dichotomous variable “age group” with 1 indicating an individual’s age is below the average age of the country’s PISA sample and 2 indicating an individual’s age is at or above this average age.

Students were asked the country of birth of themselves and their parents by the question “In what country were you and your parents born?” We generated a dichotomous variable “migrant status” with 1 = native students and 2 = migrant students. Native students are those students born in the country of assessment with at least 1 parent also born in the same country, while migrant students include those born in the country of assessment but whose parents were born in another country (second-generation migrants) and those born outside the country of assessment and whose parents were also born in another country (first-generation migrants).

The school principals were asked the question “Is your school a public or a private school?” with the responses 1 = a public school and 2 = a private school. Public school means the school is managed directly or indirectly by a public education authority, government agency, or governing board appointed by the government or elected by a public franchise, while private school indicates the school is managed directly or indirectly by a non-government organization (e.g., church, trade union, business, or other private institution).

The ESCS of the students was derived from the variables about their parental highest occupational status, parental educational attainment, and home possessions. The parental occupational status was obtained from responses to 2 open-ended questions: (1) “What’s your mother’s/father’s main job?” and (2) “What does your mother/father do in her/his main job?” The responses were coded and mapped to the International Socio-Economic Index of Occupational Status (ISEI), with higher ISEI scores indicating higher levels of occupational status. The parental educational status was generated by the questions (1) “What is the completed by your mother/father?” and (2) “Does your mother/father have any of the following qualifications?” The index of home possessions was constructed by self-reported availability of 16 household items at home and amount of possessions and books at home. ESCS was computed by attributing equal weight to the 3 standardized components. The value of ESCS has a mean of 0 and a standard deviation of 1 across OECD countries [2]. More details on how ESCS was constructed can be found in the PISA 2018 Technical Report [33].

The variable “parental emotional support” was derived from responses to the following statements: (1) “My parents support my educational efforts and achievements”; (2) “My parents support me when I am facing difficulties at school”; and (3) “My parents encourage me to be confident.” The response options were “strongly disagree,” “disagree,” “agree,” and “strongly agree.” The responses to these 3 statements were combined to create the index of perceived emotional support from parents using the item response theory scaling model; the average of the index is 0, and the standard deviation is 1 across OECD countries [30]. The details of the derived index are documented in the PISA 2018 Technical Report [33] and the PISA 2015 Technical Report [34]. The PISA questions used to obtain data on exposure, outcome, and covariates can be found in S3 Table.

Statistical analysis

There was no documented analytical protocol in PISA-related documents to be followed in addressing our specific research question. An analysis plan was created at the beginning of November 2020, after we identified our research question. The analyses were conducted between 9 November 2020 and 26 July 2021, and no data-driven changes to analyses took place.

We performed all statistical analyses using svy: prefix commands in Stata/SE software (version 15.1; StataCorp), and the tests were considered statistically significant if the 2-sided p-value was less than 0.05. We used the svyset command to account for the complex survey sampling design and sampling weight (w_fstuwt). The primary sampling unit is generated by country ID and international school ID. We included replication weights (wfstr1–wfstr80) and estimated the standard errors using balanced repeated replication with Fay’s adjustment [35,36]. We also specified the mse option in the svyset command to request the mean squared error version of the estimator [35].

We first calculated the percentage of the students’ individual characteristics, then estimated the prevalence of grade repetition and each type of bullying victimization by sex. We then estimated the sex-stratified prevalence of each type of bullying victimization among promoted students and grade repeaters. Then, we used multivariable logistic regression models to estimate the odds ratios (ORs) and corresponding 95% confidence intervals (CIs) of the associations of grade repetition with each type of bullying victimization by sex, adjusting for country differences and the aforementioned control variables. Given the underlying sex differences in grade repetition and bullying victimization, we examined the association of grade repetition and sex interaction with bullying victimization. Considering that the participating students in the study were from diverse countries with different cultures, we further reproduced the analyses of estimation and association at the country level.

To evaluate the direction and magnitude of unmeasured confounding, due to this being an observational study, we used the community-contributed command evalue [37] to report the E-value (a new measure for performing sensitivity analyses) after the multivariable logistic regression models, as suggested by VanderWeele and Ding and colleagues [38,39]. The recently proposed E-value quantifies the minimum strength of association with both the exposure and the outcome that would be required of an unmeasured confounder to fully explain away the observed effect [38,40].

Results

This study included 465,146 students from 74 countries/economies, with a mean age of 15.81 years (range 15.08–16.33 years) and a standard deviation of 0.29 years (the average ages by country are listed in S4 Table); 234,218 (50.31%) participating students were girls, and 230,928 (49.69%) were boys (Table 1). Also, 7.57% were migrants and 18.71% attended private schools.

Table 1

Descriptive statistics of the individual characteristics.

Characteristic	Total sample (n = 465,146)	Girls (n = 234,218)	Boys (n = 230,928)
Age group
At or above country average age	229,302 (49.67)	115,300 (49.60)	114,002 (49.74)
Below country average age	235,844 (50.33)	118,918 (50.40)	116,926 (50.26)
Migrant status
Native	396,370 (92.43)	200,578 (92.49)	195,792 (92.38)
Migrant	58,593 (7.57)	29,380 (7.51)	29,213 (7.62)
School type
Private	82,945 (18.71)	41,606 (18.27)	41,339 (19.16)
Public	354,351 (81.29)	178,777 (81.73)	175,574 (80.84)
Index of economic, social, and cultural status	−0.61 (1.24)	−0.63 (1.26)	−0.65 (1.27)
Index of parental emotional support	−0.10 (1.01)	−0.03 (1.00)	0.02 (0.99)

Data are n (weighted percent) or weighted mean (SD).

On average across participating countries/economies, the prevalence of grade repetition was 12.26% (95% CI 11.80%–12.72%). Overall, 13.72% (95% CI 13.38%–14.07%) of students reported being frequently (a few times a month or more) left out things, 17.92% (95% CI 17.54%–18.32%) reported being frequently made fun of, 10.05% (95% CI 9.73%–10.37%) reported being frequently threatened, 12.03% (95% CI 11.66%–12.42%) reported having possessions taken away frequently, 10.45% (95% CI 10.08%–10.83%) reported being frequently hit or pushed around, 14.01% (95% CI 13.67%–14.37%) reported being a subject of rumors frequently, and 30.32% (95% CI 29.83%–30.81%) reported frequently experiencing any type of bullying. There were statistically significant differences in the prevalence of grade repetition and each bullying type between boys and girls. Boys were more likely to be retained and bullied compared to girls (Table 2).

Table 2

Prevalence of grade repetition and bullying victimization among girls and boys.

Outcome	Prevalence of outcome, n (weighted percent)			OR (95% CI, p-value)
	Total sample (n = 465,146)	Girls (n = 234,218)	Boys (n = 230,928)
Grade repetition	46,773 (12.26)	19,019 (9.53)	27,754 (15.01)	1.68 (1.58–1.78, <0.001)
Type of bullying
Other students left me out of things on purpose	52,410 (13.72)	23,733 (12.61)	28,677 (14.85)	1.21 (1.15–1.27, <0.001)
Other students made fun of me	70,499 (17.92)	28,905 (15.67)	41,594 (20.22)	1.36 (1.29–1.44, <0.001)
I was threatened by other students	39,736 (10.05)	13,304 (7.37)	26,432 (12.77)	1.84 (1.71–1.98, <0.001)
Other students took away or destroyed things that belonged to me	44,327 (12.03)	15,465 (9.37)	28,862 (14.74)	1.67 (1.57–1.78, <0.001)
I got hit or pushed around by other students	41,899 (10.45)	13,175 (7.35)	28,724 (13.60)	1.98 (1.85–2.13, <0.001)
Other students spread nasty rumors about me	58,004 (14.01)	25,201 (12.40)	32,803 (15.66)	1.31 (1.24–1.39, <0.001)
Any type of the above	124,103 (30.32)	53,513 (27.06)	70,59 (33.61)	1.36 (1.31–1.43, <0.001)

OR, odds ratio.

As shown in Table 3, grade repetition was associated with each type and any type of bullying victimization before adjustment for measured confounders. Compared with promoted peers, grade repeaters were more likely to experience being frequently left out (OR = 1.46 [95% CI 1.36–1.57]), being frequently made fun of (OR = 1.44 [95% CI 1.35–1.55]), being frequently threatened (OR = 2.26 [95% CI 2.11–2.42]), having possessions taken away frequently (OR = 2.10 [95% CI 1.95–2.25]), being frequently hit or pushed around (OR = 2.27 [95% CI 2.11–2.45]), being a subject of rumors frequently (OR = 2.02 [95% CI 1.91–2.15), and frequently experiencing any type of bullying (OR = 1.68 [95% CI 1.58–1.78]) (all p-values < 0.001). The sex-specific analyses showed similar results in boys and girls.

Table 3

Prevalence of bullying victimization among promoted students and grade repeaters.

Type of bullying	Prevalence of outcome, as n (weighted percent), or OR (95% CI, p-value)
	Total sample (n = 465,146)	Girls (n = 234,218)	Boys (n = 230,928)
Other students left me out of things on purpose
Promoted students	45,480 (13.12)	21,189 (12.16)	24,291 (14.16)
Grade repeaters	6,930 (18.07)	2,544 (16.90)	4,386 (18.83)
OR (95% CI, p-value)	1.46 (1.36–1.57, <0.001)	1.47 (1.32–1.64, <0.001)	1.41 (1.30–1.52, <0.001)
Other students made fun of me
Promoted students	61,437 (17.22)	25,654 (15.05)	35,783 (19.56)
Grade repeaters	9,062 (23.11)	3,251 (21.64)	5,811 (24.06)
OR (95% CI, p-value)	1.44 (1.35–1.55, <0.001)	1.56 (1.42–1.71, <0.001)	1.30 (1.20–1.41, <0.001)
I was threatened by other students
Promoted students	25,654 (15.05)	11,276 (6.60)	21,695 (11.48)
Grade repeaters	3,251 (21.64)	2,028 (14.93)	4,737 (20.24)
OR (95% CI, p-value)	2.26 (2.11–2.42, <0.001)	2.48 (2.16–2.86, <0.001)	1.96 (1.80–2.13, <0.001)
Other students took away or destroyed things that belonged to me
Promoted students	36,977 (10.89)	13,228 (8.52)	23,749 (13.45)
Grade repeaters	7,350 (20.40)	2,237 (17.67)	5,113 (22.18)
OR (95% CI, p-value)	2.10 (1.95–2.25, <0.001)	2.31 (2.05–2.60, <0.001)	1.83 (1.68–2.00, <0.001)
I got hit or pushed around by other students
Promoted students	34,922 (9.30)	11,210 (6.61)	23,712 (12.21)
Grade repeaters	6,977 (18.90)	1,965 (14.58)	5,012 (21.69)
OR (95% CI, p-value)	2.27 (2.11–2.45, <0.001)	2.41 (2.11–2.76, <0.001)	1.99 (1.82–2.18, <0.001)
Other students spread nasty rumors about me
Promoted students	48,911 (12.80)	21,936 (11.58)	26,975 (14.12)
Grade repeaters	9,093 (22.90)	3,265 (20.29)	5,828 (24.58)
OR (95% CI, p-value)	2.02 (1.91–2.15, <0.001)	1.94 (1.76–2.15, <0.001)	1.98 (1.84–2.14, <0.001)
Any type of the above
Promoted students	107,383 (28.88)	47,442 (25.99)	59,941 (32.00)
Grade repeaters	16,720 (40.56)	6,071 (37.21)	10,649 (42.72)
OR (95% CI, p-value)	1.68 (1.58–1.78, <0.001)	1.69 (1.55–1.84, <0.001)	1.58 (1.48–1.69, <0.001)

OR, odds ratio.

After adjustment for sex, age group, migrant status, school type, ESCS, and parental emotional support, the OR of each type of bullying victimization for grade repeaters compared with promoted students was significantly greater than 1: 1.27 (95% CI 1.17–1.38) for being frequently left out, 1.20 (95% CI 1.11–1.31) for being frequently made fun of, 1.69 (95% CI 1.56–1.83) for being frequently threatened, 1.61 (95% CI 1.47–1.76) for having possessions taken away frequently, 1.70 (95% CI 1.56–1.85) for being frequently hit or pushed around, 1.65 (95% CI 1.53–1.77) for being a subject of rumors frequently, and 1.42 (95% CI 1.32–1.52) for frequently experiencing any type of bullying (all p-values < 0.001). The sex-specific analyses produced similar results in boys and girls (Table 4). Furthermore, while sex differences were not apparent overall for bullying victimization, there were interactions indicating that sex differences existed for some types of bullying. Specifically, girls who repeated a grade had higher risk than boys of being made fun of, being threatened, having possessions taken away, and being pushed around (S5 Table).

Table 4

Associations of grade repetition with bullying victimization for girls and boys.

Type of bullying victimization	Adjusted OR (95% CI, p-value) or E-value (95% CI lower bound)
	Total sample	Girls	Boys
Other students left me out of things on purpose
Adjusted OR (95% CI, p-value)	1.27 (1.17–1.38, <0.001)	1.28 (1.13–1.45, <0.001)	1.26 (1.15–1.40, <0.001)
E-value (95% CI lower bound)	1.86 (1.62)	1.88 (1.51)	1.83 (1.57)
Other students made fun of me
Adjusted OR (95% CI, p-value)	1.20 (1.11–1.31, <0.001)	1.32 (1.18–1.47, <0.001)	1.15 (1.05–1.27, <0.001)
E-value (95% CI lower bound)	1.69 (1.46)	1.97 (1.64)	1.57 (1.28)
I was threatened by other students
Adjusted OR (95% CI, p-value)	1.69 (1.56–1.83, <0.001)	1.93 (1.63–2.30, <0.001)	1.59 (1.44–1.75, <0.001)
E-value (95% CI lower bound)	2.77 (2.50)	3.27 (2.64)	2.56 (2.24)
Other students took away or destroyed things that belonged to me
Adjusted OR (95% CI, p-value)	1.61 (1.47–1.76, <0.001)	1.93 (1.67–2.23, <0.001)	1.48 (1.33–1.65, <0.001)
E-value (95% CI lower bound)	2.60 (2.30)	3.27 (2.73)	2.32 (1.99)
I got hit or pushed around by other students
Adjusted OR (95% CI, p-value)	1.70 (1.56–1.85, <0.001)	1.91 (1.63–2.24, <0.001)	1.62 (1.47–1.80, <0.001)
E-value (95% CI lower bound)	2.79 (2.50)	3.23 (2.64)	2.62 (2.30)
Other students spread nasty rumors about me
Adjusted OR (95% CI, p-value)	1.65 (1.53–1.77, <0.001)	1.63 (1.45–1.83, <0.001)	1.64 (1.49–1.79, <0.001)
E-value (95% CI lower bound)	2.69 (2.43)	2.64 (2.26)	2.66 (2.34)
Any type of the above
Adjusted OR (95% CI, p-value)	1.42 (1.32–1.52, <0.001)	1.49 (1.33–1.66, <0.001)	1.37 (1.26–1.48, <0.001)
E-value (95% CI lower bound)	2.19 (1.97)	2.34 (1.99)	2.08 (1.83)

All models adjusted for country; sex; age group; migrant status; school type; economic, social, and cultural status; and parental emotional support. OR, odds ratio.

In the sensitivity analysis, the observed associations between grade repetition and bullying victimization seemed moderately robust to potential unmeasured confounding, according to the calculated E-values and corresponding 95% CIs (Table 4). For instance, to explain away an OR of 1.42 for any type of bullying victimization, an unmeasured confounder associated with both grade repetition and bullying victimization of any type with an OR of 2.19 each could suffice, but weaker confounding could not. To shift the CI lower bound to contain the null value, an unmeasured confounder associated with both grade repetition and bullying victimization of any type with an OR of 1.97 could suffice, but weaker confounding could not.

In the country-specific analysis, the prevalence estimates for grade repetition and any type of bullying victimization varied widely across the participating countries/economies (grade repetition range 0.81%–42.53%; bullying victimization range 9.4%–64.81%) (S6 Table). Also, grade repeaters reported a higher prevalence of any type of bullying victimization than promoted peers. These findings were true for all countries/economies (S7 Table). The multivariable logistic regression showed that in 46 countries/economies (62.16%), grade repetition was statistically associated with any type of bullying victimization (S8 Table).

Discussion

To our knowledge, this global analysis of bullying victimization among adolescents is the first to examine the association of grade repetition with bullying using data that were collected with a common international protocol. Our study contributes 3 key findings to the literature. First, there was a high prevalence of grade repetition and bullying victimization observed among the 465,146 participating students, and prevalence varied widely across the 74 countries/economies. Second, grade repeaters were significantly more likely to experience bullying victimization than promoted students. This association was highly consistent between boys and girls. Third, girls who repeated a grade were more likely to report having experienced specific types of bullying than boys.

The prevalence of bullying victimization observed in PISA was consistent with previous studies. A study conducted between 2009 and 2015 reported that the prevalence of being bullied at least once in the past 30 days was 34.4% in 68 low- and middle-income countries [41]. Another similar study conducted between 2003 and 2015 in 83 low-, middle-, and high-income countries reported that the prevalence of bullying victimization on 1 or more days in the past 30 days was 30.5% [18]. However, these 2 studies used a generic measure of school bullying. Using multiple specific items with similar cutoffs [42,43], our reported prevalence of any type of frequent bullying of 30.32% is much higher than the value reported by a large study conducted in 2005–2006 in 40 European and North American countries, which revealed a prevalence of bullying of 16.2% in the past 2 months [44]. However, consistent with previous studies [18,44], there was wide variation in the estimated prevalence of bullying victimization between countries.

According to the 2012 Global Education Digest of the United Nations Educational, Scientific and Cultural Organization [3], 32.2 million pupils in 2010 repeated a grade in ISCED 1 and 14.1 million repeated a grade in ISCED 2 globally. An earlier study based on PISA 2009 reported that about 13% of students in OECD countries repeated at least 1 grade, although the prevalence by country ranged from 0% to over 40% [2]. Our reported overall prevalence of 12.26% is therefore consistent with previous studies [2], and we also found wide variation in prevalence across countries.

Research into the consequences of grade repetition has so far focused on its academic outcomes, including academic achievement [2], educational attainment [45], and achievement motivation [5]. However, its non-academic consequences on adolescents’ self-esteem [46-48], peer relationships [46], and psychological adjustment [15] have received less attention. Studies on the association between grade repetition and bullying victimization are especially scarce, and so far have produced inconsistent findings. One study of 378 students in grades 5 to 9 in southern Brazil observed no differences in bullying victimization, including self-reported verbal, physical, and social bullying, between grade repeaters and promoted students [14]. The authors attributed the null findings to the relatively greater physical size of grade repeaters and the normality of being retained in Brazilian public schools [14]. However, a US study with 276 students found that students who were old for their grade were rated by teachers as more likely to be bullied physically, verbally, or socially than their peers who were age appropriate for their grade [15]. The authors speculated that needing to be retained might be considered a red flag for behavioral and learning disabilities, although the intervention of repeating a grade has the negative effect on their social and emotional adjustments [15]. The US study was limited by a lack of distinguishing grade repeaters from their delayed-entry peers [14]. Our study therefore adds to a sparse body of evidence that links grade repetition to bullying victimization. Globally, grade repetition is associated with bullying victimization. Consistent with our hypothesis, we observed that grade repeaters, either boys or girls, were more likely to report being bullied for each type of bullying and any type of bullying.

Although the pathways that underly this association remain unclear, given the study design and coverage of PISA questionnaires, according to the social–ecological framework we posit that an increased risk of victimization is partly attributed to social marginalization and perceived power imbalance between grade repeaters and promoted peers. The stigma associated with “failing” a grade may be a sign of disgrace that alienates the student from the dominant “in group” [24]. Because being retained is often perceived as an academic failure in the school community or society, grade repeaters could be labeled as deviant and stigmatized as failing [49] and could therefore be more vulnerable to bullying than promoted peers [23]. It is important to differentiate the experience of verbal and social bullying from that of physical violence because students who have been retained are more likely to be bullied verbally or socially due to their relatively greater physical size [14]. The counterintuitive finding on physical victimization in our study could be explained by the perceived power imbalance of grade repeaters. According to Olweus, the perceived imbalance is associated with not only physical attributes (e.g., strength, size, and numbers) but also social status (e.g., popularity and preference) in the peer group [16]. However, the increased risks of bullying victimization among grade repeaters could also be partly attributed to the fact they have more difficulty adjusting to their new environment than promoted students who may have lower academic achievement [50]. Also, the significant negative coefficient found for the interaction of grade repetition and sex indicates that girls have a higher risk than boys of some types of bullying victimization associated with repeating a grade. The higher risk of experiencing some types of bullying among girls who repeat a grade suggests the possibility that targeted anti-bullying prevention interventions may be especially beneficial for girls.

The country-specific analyses showed some inconsistency in the association of grade repetition with any type of bullying victimization, although the association was statistically significant in most countries. There are a few potential reasons for these country differences that correspond with the societal level of the social–ecological framework. First, it is possible that the prevalence of bullying victimization is lower where there are bullying prevention policies and programs [44]. Second, cultural differences in what constitutes bullying and tacit social controls over the behavior could also contribute to the variation [51]. Third, the cultural acceptability of violence, violent crime rates, and the degree of stigma for repeating a grade at the country level would also contribute [24,52]. Unfortunately, we did not have data on these indicators to examine these social and cultural mechanisms.

Strengths and limitations

The main strength of this study is the large representative samples of 74 countries/economies, which enabled us to thoroughly analyze the sex-stratified association of grade repetition with each type of bullying victimization at the global level. In addition, the PISA methodology represents a collaborative standardized procedure on sampling, questionnaire design, and data collection, which generates comparative results across participating countries and economies.

There are also potential limitations that should be considered. First, this study is susceptible to possible selection bias because only adolescents attending schools participate in PISA. Students who are absent or drop out of school are more likely to be bullied and retained [6,19]. Second, the cross-sectional design used in this study hinders us from establishing causality. Although grade repetition may increase the risk of being bullied, bullying victimization could increase the likelihood of repeating a grade because of reduced academic performance [53,54]. Hence our results should be viewed as exploratory and in need of replication using longitudinal data. Third, our data were derived solely from PISA questionnaires, and bias due to unmeasured confounders, including body mass index, disability, ethnicity, and school location, cannot be ruled out. To address this limitation, we calculated E-values to quantify the potential unmeasured confounding necessary to nullify our findings [38]. The sensitivity analyses with the E-values suggest that our findings are robust against unmeasured confounding. Fourth, the PISA measurement is retrospective self-report and therefore subject to recall bias, although self-report is an accepted method of measuring experiences of bullying and grade repetition [2,16]. Fifth, this study only included 15-year-old adolescents in school, making our findings difficult to generalize to adolescents of other age groups [41].

Implications

To achieve the global SDG targets for children and adolescents, urgent action is needed to tackle grade repetition and school bullying [25]. If generalizable to broader age groups, our findings may have significant implications for the well-being of 1.3 billion school-age children and adolescents worldwide in an era that increasingly embraces evidence-based interventions [55]. Our study provides evidence for the association of grade repetition with bullying victimization. The observed link raises the possibility that the widespread educational policy of grade repetition may partly contribute to differences in bullying victimization and adds to the evidence against the policy of grade repetition; however, our study cannot establish a causal relationship. Also, sex differences may require particular interventions for girls, although all students who repeat a grade need attention. These results are of great concern for parents, teachers, principals, and policymakers at different levels, especially in countries where grade repetition is particularly prevalent.

Conclusions

This global analysis of school-age children found that grade repetition and bullying victimization are prevalent, vary widely across countries, and are significantly related. Grade repetition is associated with increased likelihood of bullying victimization, and girls are more likely than boys to experience specific types of bullying associated with repeating a grade.

Sample size in this study, by country/economy.

(DOCX)

Click here for additional data file.

STROBE checklist for cross-sectional studies.

(DOCX)

Click here for additional data file.

Measure-related questions in PISA 2018.

(DOCX)

Click here for additional data file.

Country-specific description of age.

(DOCX)

Click here for additional data file.

Associations of grade repetition and sex interaction with bullying victimization.

(DOCX)

Click here for additional data file.

Country-specific prevalence of grade repetition and any type of bullying victimization (weighted percent).

(DOCX)

Click here for additional data file.

Country-specific prevalence of any type of bullying victimization among promoted students and grade repeaters (weighted percent).

(DOCX)

Click here for additional data file.

Country-specific association between grade repetition and any type of bullying victimization.

(DOCX)

Click here for additional data file.

29 Jan 2021

Dear Dr Zuo,

Thank you for submitting your manuscript entitled "Grade repetition and bullying victimization in teenagers: a global cross-sectional study" for consideration by PLOS Medicine.

Your manuscript has now been evaluated by the PLOS Medicine editorial staff as well as by the Special Issue guest editors and I am writing to let you know that we would like to send your submission out for external peer review.

However, before we can send your manuscript to reviewers, we need you to complete your submission by providing the metadata that is required for full assessment. To this end, please login to Editorial Manager where you will find the paper in the 'Submissions Needing Revisions' folder on your homepage. Please click 'Revise Submission' from the Action Links and complete all additional questions in the submission questionnaire.

Please re-submit your manuscript within two working days, i.e. by .

Login to Editorial Manager here: https://www.editorialmanager.com/pmedicine

Once your full submission is complete, your paper will undergo a series of checks in preparation for peer review. Once your manuscript has passed all checks it will be sent out for review.

Feel free to email us at plosmedicine@plos.org if you have any queries relating to your submission.

Kind regards,

Artur A. Arikainen

Associate Editor

PLOS Medicine

14 May 2021

Dear Dr. Zuo,

Thank you very much for submitting your manuscript "Grade repetition and bullying victimization in teenagers: a global cross-sectional study" (PMEDICINE-D-21-00390R1) for consideration in PLOS Medicine’s Special Issue: Global Child Health: From Birth to Adolescence and Beyond.

Your paper was evaluated by a senior editor and discussed among all the editors here. It was also discussed with the special issue guest editors, and sent to independent reviewers, including a statistical reviewer. The reviews are appended at the bottom of this email and any accompanying reviewer attachments can be seen via the link below:

[LINK]

In light of these reviews, I am afraid that we will not be able to accept the manuscript for publication in the journal in its current form, but we would like to consider a revised version that addresses the reviewers' and editors' comments. Obviously we cannot make any decision about publication until we have seen the revised manuscript and your response, and we plan to seek re-review by one or more of the reviewers.

In revising the manuscript for further consideration, your revisions should address the specific points made by each reviewer and the editors. Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments, the changes you have made in the manuscript, and include either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. Please submit a clean version of the paper as the main article file; a version with changes marked should be uploaded as a marked up manuscript.

In addition, we request that you upload any figures associated with your paper as individual TIF or EPS files with 300dpi resolution at resubmission; please read our figure guidelines for more information on our requirements: http://journals.plos.org/plosmedicine/s/figures. While revising your submission, please upload your figure files to the PACE digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at PLOSMedicine@plos.org.

We expect to receive your revised manuscript by Jun 04 2021 11:59PM. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

We ask every co-author listed on the manuscript to fill in a contributing author statement, making sure to declare all competing interests. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. If new competing interests are declared later in the revision process, this may also hold up the submission. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT. You can see our competing interests policy here: http://journals.plos.org/plosmedicine/s/competing-interests.

Please use the following link to submit the revised manuscript:

https://www.editorialmanager.com/pmedicine/

Your article can be found in the "Submissions Needing Revision" folder.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

We look forward to receiving your revised manuscript.

Sincerely,

Caitlin Moyer, Ph.D.

Associate Editor

PLOS Medicine

plosmedicine.org

-----------------------------------------------------------

Requests from the editors:

1. Abstract: Background: Line 24 (and throughout): Please refer to low or middle income countries rather than "developing countries" and high income countries rather than "developed" countries.

2. Abstract: Methods and findings: Please clarify the age of the students (at line 32 this is indicated as 15 years, but at line 41 a range of 15.08-16.33 years is given). number of participants, years during which the study took place, length of follow up, and main outcome measures.

3. Abstract: Conclusions: Line 57-58: Here, and throughout the manuscript, please avoid statements implying causality of the results. “These findings provide robust evidence for the negative effect of grade repetition on bullying victimization.”

4. Author summary: Please format the author summary with bullet points rather than paragraph form (2-3 points per section). Please revise the final statement to avoid overstating what can be concluded from your study: “Globally the policy of grade repetition is harmful to students’ health and wellbeing in some way regarding to school bullying.”

5. Methods: Line 139-142: Please at least briefly mention how schools and students were selected for inclusion: “A two stage sampling procedure was adopted within each country/economy. In the first stage, a representative sample of at least 150 schools was selected, then roughly 42 students were selected from each school [22]. At line 149-150, please clarify this statement further: “The threshold for an acceptable participation rate after replacement was 85% [22].”

6. Methods: LIne 165-166: make sure that the Methods section transparently describes when analyses were planned, and when/why any data-driven changes to analyses took place (with rationale).

7. Methods: Line 198: Please revise if this should be “Are you female or male?” in this question.

8. Methods: Line 217-219: Please clarify the measures used to derive ESCS.

9. Methods: Line 220-225: Please provide more detail on how the parental support measure was derived.

10. Methods: If possible, please include the PISA questions used to obtain the data on measures and covariates as a supporting information file.

11. Results: Line 258-259: Please clarify this sentence, as the footnote of Table 2 suggests that there are differences: “There were no statistically significant differences in the distribution of individual characteristics between boys and girls.”

12. Results: Line 260-261: It may be useful to provide a range here for prevalence of grade repetition.

13. Results: Line 290-292: Please provide P values in addition to 95% CIs for sensitivity analyses.

14. Discussion: Line 302-304: Please avoid the use of causal language, here and throughout the Discussion.

15. Discussion: Line 338: “The effect of grade repetition on bullying victimization is negative in our study.” Please avoid implying causality in this sentence.

16. Discussion: Line 381: Please remove the phrase “if causal” from this sentence, as your study cannot inform on the causal nature of this relationship.

17. Discussion: Line 383-384: Please revise to avoid causal implications: “Our study provides robust evidence for the negative consequence of grade repetition on bullying victimization.”

18. Discussion: Line 393-396: Please revise, interpreting the study based on the results presented in the abstract, emphasizing what is new without overstating your conclusions. “Our findings confirm that grade repetition is associated with bullying victimization and highlight that social promotion is a better policy practice than grade retention regarding to bullying victimization although both practices are inefficient 396 to improve educational outcomes of students.”

19. Please provide titles and legends for each individual table and figure in the Supporting Information.

20. Table 2 and Table 3: Please provide the statistical results (95% CIs and P values ) that are missing from the Table.

21. S3 Table: Please provide individual P values for each result.

Comments from the reviewers:

Reviewer #1: This study aims to investigate the global association of grade repetition with bullying victimization in a large-scale school-based cross-sectional study.

Comments:

"All participants in Japan, Malaysia and Norway were excluded because no students in the three countries have ever repeated a grade (n=18,033). "

Can the authors please assess the bullying in these three countries for comparison of prevalence/frequency, given no grade repetition?

"Last, the participants without the data on grade repetition or school bullying (n=111,017) were excluded from the analysis"

Can the authors please discuss whether these excluded participants can be considered to be missing at random (MAR)?

The authors have followed STROBE and suitably included the associated checklist with the supplementary material.

"Participants were categorized as retained students if they answered "Yes" to any of these three questions, and promoted students if they didn't have the experience with repeating a grade at school before the survey."

Did the authors consider keeping "Yes, once" or "Yes, twice or more" as separate groups for the analysis, to explore the possibility of a dose-response relationship?

"The response options were "never or almost never", "a few times a year", "a few times a month", or "once a week or more". For each of the six experiences mentioned above, we identified bullying victimization by the responses of "a few times a month" or "once a week or more" [25, 26]. Besides, we combined the six experiences into a single indicator, "any type of bullying victimization", if the participants reported being bullied frequently (at least a few times a month) in any type".

Did the authors complete any sensitivity analyses looking at different ways to dichotomise bullying victimisation? Furthermore, did they consider keeping all four categorical outcomes within the analysis?

"Demographic characteristics assessed included sex, age group, immigration status, school type, economic, social and culture status (ESCS), and perceived parents' emotional support".

Did the authors consider including BMI, disability, ethnicity, or urban/rural school location as other potential confounders in their analysis?

"As suggested by a study of PISA 2009 [2], we generated a dichotomous variable "age group" with 1 indicating the age is below the country average age, and 2 indicating the age is at or above the country average age".

Did the authors consider completing additional analyses treating age as a continuous variable?

"Finally, we adopted multivariable logistic regression models to estimate the strength, direction, and statistical significance of the relations linking experience with grade repetition to each type of bullying victimization by sex, adjusting the participating country ID and the measured control variables aforementioned in covariates assessment".

Did the authors consider applying hierarchical or multilevel modelling to better allow for clustering by country differences? Furthermore, there may be school differences within countries?

"The recently proposed E-value quantitates the minimum strength of association with both the exposure and the outcome which would be required of an unmeasured confounder to fully explain away the observed effect [30, 32]."

This is a novel and useful approach for estimating the possible strength of impact of unknown confounding on the study inferences. Can the authors please clarify within the manuscript if this analysis considers one confounder at a time, or multiple confounders working together?

The authors are correct to state that the study design means that causality cannot be inferred.

Therefore, the evidence provided from this study cannot determine whether grade repetition increases the chance of bullying, or bullying victimisation increases chance of grade repetition.

To introduce analyses that take into account the timing of events is a sensible idea, that might help uncover the direction of association that is observed in the study findings: "The reference period of being bullied (past 12 months before PIAS 2018) and repeating a grade at ISCED3 may overlap. To establish the temporal relationship between grade repetition and bullying victimization, we further repeated the analyses of the associations of repeating a grade at ISCED1 or ISCED2 with the risk of being bullied."

However, this does not allow for occurrences of bullying prior to the last 12 months, which is not captured by the data.

The authors state in the results that: "As a sensitivity analysis we also evaluated associations of repeating a grade at ISCED1 or ISCED2 with the risk of being bullied to make sure exposure (grade repetition) preceded outcomes (bullying victimization), and found similar results (S3 Table). For instance, the multivariable adjusted OR (95% 291 CI) for bullying victimization in any type comparing promoted students to retained students in all students was 1.44(1.34-1.55); in girls: 1.51(1.34-1.70); and in boys: 1.40(1.29-1.52). "

This is misleading as in fact this analysis does not 'make sure exposure (grade repetition) preceded outcomes (bullying victimization)'. Occurrences of bullying victimisation preceding ISCED1 or ISCED2 have not been captured in the data included within this analysis. Even more entwined, the relationship between grade repetition and bullying victimisation could be iterative in nature.

Furthermore, the conclusions that ""Globally the policy of grade repetition is harmful to students' health and wellbeing in some way regarding to school bullying" cannot and should not be inferred from the study outcomes found from this research.

The authors acknowledge that: "The major limitation is this study only included students attended the school, which may be subject to possible selection bias. In addition, the cross-sectional design hinders us from establishing causality between grade repetition and bullying victimization", but later state that "however, we also used a finer category of repeating a grade before ISCED3 to ensure grade reputation preceded bullying victimization, and further examined the temporal direction of the relationship between grade repetition and bullying victimization", which is inaccurate as temporal direction is not ensured by the analysis due to the outcome question asking about the last 12 months only.

Similarly, the authors go on to state that: "Our findings, if causal and generalizable to broader age groups, may have significant implications for the wellbeing of 1.3 billion school-age children and adolescents around the world in an era embracing evidence-based interventions [47]. Our study provides robust evidence for the negative consequence of grade repetition on bullying victimization. The observed link with grade repetition and bullying victimization implies that the widespread educational policy of grade repetition may partly contribute to the differences in bullying victimization, and adds to the evidence against the policy of grade repetition."

This needs to be completely reworded, with the authors applying far more caution and care with their stated study implications and conclusions. Causality cannot be inferred by this analysis, and therefore causality should not be implied in the interpretation of the outcomes.

And again, "Our findings... highlight that social promotion is a better policy practice than grade retention regarding to bullying victimization although both practices are inefficient to improve educational outcomes of students."

This statement is not reflective of the evidence presented here.

Reviewer #2: I like the idea of looking at grade retention and bullying victimization using the PISA data. I have a few concerns to consider.

1. I would not look at bullying victimization by type. We have shown as have others that what matters for outcomes is severity and not form (see Haltigan & Vaillancourt, 2018). You also improve measurement precision by creating a composite score as suggested by Vaillancourt et al., 2010.

2. I would concentrate my efforts on looking at grade retention by bullying victimization by country. You have this in your S2 table; it should replace your table on type.

3. The implied direction is that being retained impacts bullying but it is possible that being bullied impact grades which impacts retention. There are longitudinal studies supporting both pathways bullied to poor grades and poor grades to being bullied. This needs to be addressed in the manuscript.

4. Boys are more likely than girls to be retained than girls and to be bullied. I would look at this as a 3-way interaction. Maybe even a 4-way as per suggestion #2.

5. There are a number of typos that need to be addressed (example lines 107-108). Avoid using contractions like isn't or didn't.

Reviewer #3: The manuscript, Grade repetition and bullying victimization in teenagers: A global cross-sectional study, examines the association between a history of grade retention and experiences of bullying victimization in the past year for adolescents with an average age of about 15.8 years in 74 countries around the world. The study addresses a relatively neglected area of grade retention research by examining its association with bullying victimization. Strengths and weaknesses include the large, global sample. Strengths include the inclusion of so many countries around the world with a large sample size. Weaknesses include the averaging of the experiences of adolescents from such diverse countries with different cultures related to school experiences, peer relationships, and gender. The paper also would benefit from significant copy editing due to numerous typos, run-on sentences, and errors in word selection and grammar.

The authors build their study rationale based on a lack of empirical studies relating grade retention and bullying victimization. They touch on some reasons why other researchers have explained associations or no associations between grade retention and bullying victimization. However, the study is not guided by a theoretical orientation or explanation of the mechanisms why grade retention would be related to bullying victimization. I recommend authors add an orientation and then also attend to how this orientation would relate to various cultural contexts around the globe. Would associations differ in different cultural contexts? Is an analysis by country warranted? Why or why not?

At the bottom of page 10, the authors wrote, "Finally, we adopted multivariable logistic regression models to estimate the strength, direction, and statistical significance of the relations linking experience with grade repetition to each type of bullying victimization by sex, adjusting the participating country ID and the measured control variables aforementioned in covariates assessment." I was unclear how the authors used the participating country ID in the analysis. I am also curious why authors did not report the full results of the logistic regression in terms of the association of control variables to bullying victimization.

Authors state that their research goal is to understand the association between grade repetition and bullying victimization, but they also touch upon gender differences. It would be helpful for authors to move from a theoretical orientation to the status of empirical evidence for their purpose to a final, complete set of research questions and hypotheses. Authors examine overall bullying, single bullying experiences (measured and not latent constructs), and gender differences. They should explain also why a global average is best and why examining individual country contexts is not needed.

In their discussion, authors wrote, "The effect of grade repetition on bullying victimization is negative in our study" but, in fact, grade repetition was positively related to bullying victimization (students who had been retained were more likely to experience bullying victimization). Also, authors should avoid using terms like "effect" that imply a causal association.

Any attachments provided with reviews can be seen via the following link:

[LINK]

2 Jun 2021

Submitted filename: Rebuttal letter.docx

Click here for additional data file.

19 Jul 2021

Dear Dr. Zuo,

Thank you very much for re-submitting your manuscript "Grade repetition and bullying victimization in teenagers: a global cross-sectional study" (PMEDICINE-D-21-00390R2) for consideration in PLOS Medicine’s Special Issue: Global Child Health: From Birth to Adolescence and Beyond.

Your revised paper was evaluated by a senior editor and discussed among all the editors here. It was also discussed with the Special Issue Guest Editors, and evaluated by two of the original reviewers, including a statistical reviewer. The reviews are appended at the bottom of this email and any accompanying reviewer attachments can be seen via the link below:

[LINK]

Given the remaining requests of Reviewer 3, we would like to consider a further revised version that addresses the reviewers' and editors' comments. Obviously we cannot make any decision about publication until we have seen the revised manuscript and your response, and we may to seek re-review by one or more of the reviewers.

In revising the manuscript for further consideration, your revisions should address the specific points made by each reviewer and the editors. Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments, the changes you have made in the manuscript, and include either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. Please submit a clean version of the paper as the main article file; a version with changes marked should be uploaded as a marked up manuscript.

In addition, we request that you upload any figures associated with your paper as individual TIF or EPS files with 300dpi resolution at resubmission; please read our figure guidelines for more information on our requirements: http://journals.plos.org/plosmedicine/s/figures. While revising your submission, please upload your figure files to the PACE digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at PLOSMedicine@plos.org.

We expect to receive your revised manuscript by Jul 30 2021 11:59PM. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

We ask every co-author listed on the manuscript to fill in a contributing author statement, making sure to declare all competing interests. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. If new competing interests are declared later in the revision process, this may also hold up the submission. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT. You can see our competing interests policy here: http://journals.plos.org/plosmedicine/s/competing-interests.

Please use the following link to submit the revised manuscript:

https://www.editorialmanager.com/pmedicine/

Your article can be found in the "Submissions Needing Revision" folder.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

We look forward to receiving your revised manuscript.

Sincerely,

Caitlin Moyer, Ph.D.

Associate Editor

PLOS Medicine

plosmedicine.org

-----------------------------------------------------------

Requests from the editors:

1. Title: Please capitalize the first word of the subtitle: “Grade repetition and bullying victimization in teenagers: A global cross-sectional study”

2. Abstract: Methods and Findings: Please note the PISA assessment is administered to all students attending secondary education, with ages ranging from 15-16 years. Please clarify what is meant by “deprived victimization score” by briefly describing this is a composite score.

3. Abstract: Methods and Findings: Line 38: Please note the potential confounders adjusted for in the analysis.

4. Abstract: Methods and findings: Line 48: Please briefly elaborate on the sex differences observed.

5. Abstract: Conclusions: Please clarify the final sentence, we suggest: “Sex differences in risk of experiencing bullying suggest that girls who repeat a grade may benefit from tailored interventions.” or similar.

6. Author summary: Please revise the authors summary. Please provide 2-3 single sentence bullet points under each of the following headings. Bullet points should be objective, brief, succinct, specific, accurate, and avoid technical language.

-Why Was This Study Done? Authors should reflect on what was known about the topic before the research was published and why the research was needed.

-What Did the Researchers Do and Find? Authors should briefly describe the study design that was used and the study’s major findings. Do include the headline numbers from the study, such as the sample size and key findings.

-What Do These Findings Mean? Authors should reflect on the new knowledge generated by the research and the implications for practice, research, policy, or public health. Authors should also consider how the interpretation of the study’s findings may be affected by the study limitations.

7. In text references: Please include a space between the word and the reference bracket. For multiple references listed within brackets, please do not include a space (for example [13,15]).

8. Introduction: Line 76: Please change “Belgian” to “Belgium” and please revise to read “...while grade repetition has been discouraged strongly…”

9. Methods: Line 154-155: From looking at the Figure 1 flowchart, it is not clear whether in these countries grade repetition does not occur, or whether the data were not collected.

10. Methods: Line 188: Please provide more information on the construction of the composite victimization score.

11. Methods: Line 216: Please define the abbreviation ESCS at the first point of use in the text.

12. Methods: Line 240: Please indicate any changes made to the analysis in response to peer reviewer comments (such as the implementation of the victimization composite score).

13. Methods: Is it possible to include a supporting information file describing the average ages by country?

14. Results: Line 289: Please provide the exact p values for the sex-specific analyses describing victimization score differences for students who were retained compared to those who were promoted.

15. Results: Line 314: Please change “regress” to “regression” in this sentence. Please quantify the results from this analysis with 95% CIs and p values.

16. Discussion: Line 364-365: We suggest revising to “...may be a contributing factor to the relationship between grade repetition and bullying…” or similar to avoid causal implications.

17. Discussion: Line 378: We suggest revising to “These findings suggest that girls may benefit from targeted attention in anti-bullying programs.” or similar.

18. Discussion: Line 384-385: Please clarify what is meant by “...because the actual cross-national comparable data on bullying can only be collected by pictures.”

19. Discussion: Line 400: Please replace “improve” with “increase” or similar.

20. Page 20: Please remove the Data Availability, Funding, Competing Interests sections from the main text, and ensure all information is accurately entered in the Data Availability, Financial Disclosure, and Competing Interests sections of the manuscript submission metadata.

21. References: Please use the "Vancouver" style for reference formatting, and see our website for other reference guidelines https://journals.plos.org/plosmedicine/s/submission-guidelines#loc-references

Please check the abbreviations for Journal titles: For example, reference 36 should be Lancet Child Adolesc Health.

22. Figure 1: Please clarify whether Japan, Malaysia and Norway did not measure/report data on grade repetition, or whether grade repetition does not occur in these countries (Methods: Line 154-155: “All participants in Japan, Malaysia, and Norway were excluded because no students in the three countries have ever repeated a grade (n=18,033).”

23. Table 2 and Table 3: Please do include the versions of the table, with 95% CIs and p values reported for each comparison for bullying victimization and retention between boys and girls.

24. Table 4: Please also include the results from the unadjusted models. Please provide the 95% CIs as upper and lower bounds, as well as the p values for each comparison.

25. S3 Table: Please remove all trademark/copyright symbols from the table (e.g. ,)

26. S4 Table, S7 Table, and S10 Table: Please include the results of the unadjusted analyses as well.

Comments from the reviewers:

Reviewer #1: The authors have responded to each comment in turn, substantially rewording the manuscript where required.

Reviewer #3: The manuscript, Grade repetition and bullying victimization in teenagers: A global cross-sectional study, examines the association between a history of grade retention and experiences of bullying victimization in the past year for adolescents with an average age of about 15.8 years in 74 countries around the world. The study addresses a relatively neglected area of grade retention research by examining its association with bullying victimization. Overall, the paper is stronger but still has some areas that could be improved with further work including strengthening the theoretical rationale and clarifying the measurement and bullying victimization composites. The paper would still benefit from additional significant copy editing due to numerous typos and errors in word selection and grammar.

Authors mention that several theoretical orientations have been proposed to explain why retained students are at higher risk of being bullied but doesn't really describe a theory - authors could go more in depth in detailing a theory that drives their selection of variables. What is missing for me is thought put into how cultural differences between countries may impact the association. For example, authors mention that the degree of stigma of repeating a grade varies by culture - can the author provide some key examples? The theory should be connected to the selection of variables and the interpretation of results. I want to know why authors feel there was an association in some but not all countries. Are there cultural or policy or prevalence differences?

- Authors describe two previous studies - one in Brazil and one in the U.S. - how did the results differ from the Brazil and U.S. results in this study?

- Are higher retention rates related to less association between retention and victimization?

- I wonder if the limitation of the sample only being students attending school differs by country as well. Are there differences in school attendance rates for 15year-olds around the globe?

Regarding the measurement of victimization, I am a bit unclear on why all three types of measures are needed. Looking at each type and any type of victimization makes sense. How does the composite mean score add value? Would this be a "diversity" of bullying victimization experience? It seems hard to validate such a composite b/c someone could be physically attacked every day but never experienced any of the other types of victimization and they would have a lower score than someone who has been teased a few times a month and excluded a few times a month. The terms "victimization prevalence" and "victimization score" don't help with understanding the differences. I would appreciate a rationale provided for each measure along with a citation with validity evidence to support using this measure from extant literature.

When describing the participants, it doesn't seem helpful to write that half the participants are lower than the average age (delete) since that is self-evident.

In their interpretation of results, authors should discuss the possibility that bullying victimization caused the grade repetition, unless their theoretical orientation or existing empirical literature strongly support a one directional relation.

There were too many types and wording errors to document but here are a couple of them:

- What is "deprived victimization score"? Perhaps authors mean, "derived"

- In third author summary bullet point, I recommend changing "and vary" for "yet prevalence varies"

- I don't understand the sentence on page 18 lines 383-385 "Second, the cultural differences in the bullying definition also partly responsible for the variation because the actual cross-national comparable data on bullying can only be collected by pictures."

Additional minor wording recommendations:

- Delete, "as we mentioned in the background section."

Any attachments provided with reviews can be seen via the following link:

[LINK]

3 Aug 2021

Submitted filename: Rebuttal letter2.docx

Click here for additional data file.

26 Aug 2021

Dear Dr. Zuo,

Thank you very much for submitting your manuscript "Grade repetition and bullying victimization in teenagers: A global cross-sectional study" (PMEDICINE-D-21-00390R3) for consideration in PLOS Medicine’s Special Issue: Global Child Health: From Birth to Adolescence and Beyond.

Your revised paper was evaluated by a senior editor and discussed among all the editors here. It was also discussed with an academic editor with relevant expertise, and sent to one of the original reviewers. The reviews are appended at the bottom of this email and any accompanying reviewer attachments can be seen via the link below:

[LINK]

In light of the remaining comments from the reviewer and academic editor, we will not be able to accept the manuscript for publication in the journal in its current form, but we would like to consider a further revised version that addresses the reviewers' and editors' comments. Obviously we cannot make any decision about publication until we have seen the revised manuscript and your response, and we plan to seek re-review by one or more of the reviewers.

In revising the manuscript for further consideration, your revisions should address the specific points made by each reviewer and the editors. Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments, the changes you have made in the manuscript, and include either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. Please submit a clean version of the paper as the main article file; a version with changes marked should be uploaded as a marked up manuscript.

In addition, we request that you upload any figures associated with your paper as individual TIF or EPS files with 300dpi resolution at resubmission; please read our figure guidelines for more information on our requirements: http://journals.plos.org/plosmedicine/s/figures. While revising your submission, please upload your figure files to the PACE digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at PLOSMedicine@plos.org.

We expect to receive your revised manuscript by Sep 02 2021 11:59PM. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

We ask every co-author listed on the manuscript to fill in a contributing author statement, making sure to declare all competing interests. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. If new competing interests are declared later in the revision process, this may also hold up the submission. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT. You can see our competing interests policy here: http://journals.plos.org/plosmedicine/s/competing-interests.

Please use the following link to submit the revised manuscript:

https://www.editorialmanager.com/pmedicine/

Your article can be found in the "Submissions Needing Revision" folder.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

We look forward to receiving your revised manuscript.

Sincerely,

Caitlin Moyer, Ph.D.

Associate Editor

PLOS Medicine

plosmedicine.org

-----------------------------------------------------------

Requests from the editors:

1. From the Academic Editor: Please cite the source of the bullying composite score, indicate whether or not the score has been validated in other study populations, the alpha level of reliability of the composite score in this study population to ensure the composite score is reflective of a single, underlying bullying construct. This information is necessary to support that the scale is sufficiently reliable to be used as a single composite.

2. Abstract: Methods and Findings Line 26-27: Please revise to: “...the limited evidence of associations between grade repetition on school bullying is inconsistent…”

3. Abstract: Line 51: Please revise to “...having possessions taken away…” or similar.

4. Abstract: Conclusions: We suggest opening the first sentence with “In this study, we observed that globally, both boys and girls who ever repeat…”

5. Author summary: Please format these as bullet points.

6. Author summary: Line 79: Please revise to “Grade repetition was associated with bullying victimization in both boys and girls.”

7. Author summary: Line 87: Please revise to: “Girls who repeat a grade may be at higher risk of experiencing bullying than boys, and may benefit from targeted interventions.” or similar.

8. Introduction: Line 105: Please reference this statement: “...a common misperception is that repeating a grade allows the student to grow academically and socially.”

9. Introduction: Line 144: This is the first instance where sex differences are mentioned. It would be helpful to have some rationale for your hypothesis that the association between bullying and repetition would be stronger among girls.

10. Methods: Line 166: Please cite the PISA website as a reference.

11. Methods: Please provide more detail here or in the supporting information files (such as S3 Table) describing how the ESCS and parental emotional support scores were calculated from the survey responses.

12. Methods: Line 210: Please address the comments of the Academic Editor and Reviewer 3. We require that you please provide necessary information to support the use of the composite victimization score, including information pertaining to validity and rationale for its use, and please discuss the limitations of the score.

13. Results: Line 310- : Please provide the p values in addition to the 95% CIs for the adjusted ORs for the associations of each type of bullying with school promotion.

14. Results: Line 333-334: Please clarify this sentence: “Across all countries, retained students reported a higher level and higher prevalence of victimization than promoted students.” or similar, depending on your meaning.

15. Conclusions: Line 452-453: Please avoid the causal language in the final sentence, and please revise to: “Grade repetition was associated with increased likelihood of bullying victimization…”

16. References: Please check that the citation for reference 1 is complete.

17. S8 Table: Please remove the extra parentheses at the top of the “Adjusted” column under Odds ratio.

Comments from the reviewers:

Reviewer #3: Overall the paper is stronger in this revision, however, the authors did not adequately address my concerns about the measurement of victimization. The construct of bullying cannot be measured by creating a sum score of frequency responses to a diversity of types of victimization experiences and using it as a continuous item. I do not believe a reviewer was justified in recommending this practice, it does not make conceptual sense as bullying victimization experiences (e.g., verbal, physical, relational, cyber) do not all need to occur for someone to be severely bullied. The authors cite Vaillancourt et al. (2010) as recommending this process but, from my reading, their research recommended using a measure of behavioral based bullying with a cut-off score for a binary determination of yes/no bullied with no=never and yes=any bullying. Thus, if the authors and editor feel that the reviewer feedback is important to comply with and that a composite measure of bullying victimization is critical to answering their research questions, they should provide adequate explanation for and documentation of why this approach is valid and what this approach is measuring. I recommend Green et al. (2013) and a couple of additional articles about measuring bullying in informing the authors' decision.

Green, J. G., Felix, E. D., Sharkey, J. D., Furlong, M. J., & Kras, J. E. (2013). Identifying bully victims: Definitional versus behavioral approaches. Psychological Assessment, 25, 651-657.

Sharkey, J. D., Ruderman, M. A., Mayworm, A. M., Green, J. G., Furlong, M. J., Rivera, N., & Purisch, L. (2015). Psychosocial functioning of bullied youth who adopt versus deny the bullied victim label. School Psychology Quarterly, 30, 91-104.

Felix, E. D., Sharkey, J. D., Green, J. G., Furlong, M., F., & Tanigawa, D. (2011). Getting precise and pragmatic about the assessment of bullying: The development of the California Bullying Victimization Scale. Aggressive Behavior, 37, 234-247.

The rest of my feedback was adequately addressed.

Any attachments provided with reviews can be seen via the following link:

[LINK]

29 Aug 2021

Submitted filename: Rebuttal letter3.docx

Click here for additional data file.

20 Sep 2021

Dear Dr. Zuo,

Thank you very much for submitting your manuscript "Grade repetition and bullying victimization in teenagers: A global cross-sectional study" (PMEDICINE-D-21-00390R4) for consideration in PLOS Medicine’s Special Issue: Global Child Health: From Birth to Adolescence and Beyond.

Your paper was evaluated by a senior editor and discussed among all the editors here. It was also discussed with an academic editor with relevant expertise, and re-evaluated by one of the reviewers. The reviews are appended at the bottom of this email and any accompanying reviewer attachments can be seen via the link below:

[LINK]

In light of the comments from the academic editor, we will not be able to accept the manuscript for publication in the journal in its current form, but we would like to consider a revised version that addresses the editors' comments. Obviously we cannot make any decision about publication until we have seen the revised manuscript and your response.

In revising the manuscript for further consideration, your revisions should address the specific points made by each reviewer and the editors. Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments, the changes you have made in the manuscript, and include either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. Please submit a clean version of the paper as the main article file; a version with changes marked should be uploaded as a marked up manuscript.

In addition, we request that you upload any figures associated with your paper as individual TIF or EPS files with 300dpi resolution at resubmission; please read our figure guidelines for more information on our requirements: http://journals.plos.org/plosmedicine/s/figures. While revising your submission, please upload your figure files to the PACE digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at PLOSMedicine@plos.org.

We expect to receive your revised manuscript by Sep 27 2021 11:59PM. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

We ask every co-author listed on the manuscript to fill in a contributing author statement, making sure to declare all competing interests. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. If new competing interests are declared later in the revision process, this may also hold up the submission. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT. You can see our competing interests policy here: http://journals.plos.org/plosmedicine/s/competing-interests.

Please use the following link to submit the revised manuscript:

https://www.editorialmanager.com/pmedicine/

Your article can be found in the "Submissions Needing Revision" folder.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

We look forward to receiving your revised manuscript.

Sincerely,

Caitlin Moyer, Ph.D.

Associate Editor

PLOS Medicine

plosmedicine.org

-----------------------------------------------------------

Requests from the editors:

1. From the academic editor: Given that the authors report an alpha reliability for the scale of 0.88, it is unclear why they removed the composite scale. It also is unclear whether the alpha that is reported is from the original article describing the scale or whether it is from the sample under study. If the former, this should be stated, and the alpha for the sample should be provided. If the latter, it is unclear why the authors do not use a composite score, as an alpha of 0.88 is adequate evidence justifying the use of a composite score. In sum, the descriptions of the outcome measure needs to be cleaned up and the justification for using whatever measurement scale for the outcome(s) needs to be more clearly justified.

2. Title: We suggest revising the title to “Grade repetition and bullying victimization in adolescents: A global cross-sectional study of the Program for International Student Assessment (PISA) data from 2018.

3. Throughout: If possible, please replace “retained students” or “retained girls” etc. with “students who repeat a grade” or similar.

4. Abstract: Lines 24-25: Please clarify what is meant by primary in the sentence “Globally around 32.2 million students repeated a grade in primary in 2010.” Please revise if this should read “in primary school” or “in the primary education level” or similar.

5. Abstract: Line 49: Please revise “were associated with” to “retained girls had higher risks of” or similar.

6. Abstract: Line 58: Please revise “from repeating a grade than boys” to “...risks of specific types of bullying associated with repeating a grade than boys.”

7. Author summary: Line 84: We suggest revising to “The experience of repeating a grade may suggest a need for bullying interventions among both boys and girls”

8. Author summary: Line 86-87: We suggest revising to “Girls who repeat a grade may be at higher risk of experiencing some forms of bullying than boys, and may benefit from targeted interventions.”

9. Methods: Lines 198-209: As noted by the academic editor, please clarify whether the alpha here represents the original study, and provide the alpha for the sample if so. Please provide a clear explanation and rationale for the bullying victimization outcome used.

10. Methods: Lines 215-218, 245-246; 253-254: Reference 33 cited here is listed as “forthcoming” in the reference list. There are a number of details of the methods that reference this technical report, please update the reference in the event the 2018 report is available now, or otherwise clarify how the PISA 2018 technical report may be accessed.

11. Results Lines 314-317: “The sex-specific analyses produced similar results in boys and girls (Table 4). Furthermore, interaction analyses showed that retained girls were at higher risk of being made fun of, being threatened, having possessions taken away, and being pushed around compared with retained boys (S5 Table).” We suggest making it more clear that while sex differences were not apparent overall for bullying victimization, there were interactions indicating that for subtypes of bullying, there were differences between girls and boys who repeated a grade.

12. Discussion: Line 376: Please remove the word “undoubtedly” from the sentence, and please revise “As predicted” to “Consistent with previous work” or “Consistent with our hypothesis” or similar.

13. Discussion: Line 394: Please revise “than the promoted low achievers” to “than students who are promoted but who may have lower academic achievement” or similar.

14. Conclusion: Line 444: Please revise to avoid causal implications, we suggest: “girls are more likely than boys to experience specific types of bullying associated with repeating a grade.”

Comments from the reviewers:

Reviewer #3: I appreciate your careful consideration of the composite of bullying and ultimate removal of the composite from the study.

Any attachments provided with reviews can be seen via the following link:

[LINK]

25 Sep 2021

Submitted filename: Rebuttal letter4.docx

Click here for additional data file.

6 Oct 2021

Dear Dr. Zuo,

Thank you very much for re-submitting your manuscript "Grade repetition and bullying victimization in adolescents: A global cross-sectional study of the Program for International Student Assessment (PISA) data from 2018" (PMEDICINE-D-21-00390R5) for consideration in PLOS Medicine’s Special Issue: Global Child Health: From Birth to Adolescence and Beyond.

I have discussed the paper with my colleagues and the academic editor. I am pleased to say that provided the remaining editorial and production issues are dealt with we are planning to accept the paper for publication in the journal.

The remaining issues that need to be addressed are listed at the end of this email. Please take these into account before resubmitting your manuscript:

[LINK]

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

In revising the manuscript for further consideration here, please ensure you address the specific points made by the editors. In your rebuttal letter you should indicate your response to the editors' comments and the changes you have made in the manuscript. Please submit a clean version of the paper as the main article file. A version with changes marked must also be uploaded as a marked up manuscript file.

Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. If you haven't already, we ask that you provide a short, non-technical Author Summary of your research to make findings accessible to a wide audience that includes both scientists and non-scientists. The Author Summary should immediately follow the Abstract in your revised manuscript. This text is subject to editorial change and should be distinct from the scientific abstract.

We expect to receive your revised manuscript within 1 week. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

We ask every co-author listed on the manuscript to fill in a contributing author statement. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT.

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript.

Please note, when your manuscript is accepted, an uncorrected proof of your manuscript will be published online ahead of the final version, unless you've already opted out via the online submission form. If, for any reason, you do not want an earlier version of your manuscript published online or are unsure if you have already indicated as such, please let the journal staff know immediately at plosmedicine@plos.org.

If you have any questions in the meantime, please contact me or the journal staff on plosmedicine@plos.org.

We look forward to receiving the revised manuscript by Oct 13 2021 11:59PM.

Sincerely,

Caitlin Moyer, Ph.D.

Associate Editor

PLOS Medicine

plosmedicine.org

------------------------------------------------------------

Requests from Editors:

1. Abstract: Conclusions: Line 59-61: Please remove the word “robust” at line 59. We suggest revising the final sentence slightly to: “Sex differences in risk of experiencing some subtypes of bullying suggest that tailored interventions for girls who repeat a grade may be warranted.” or similar, as the study findings do not seem to exactly imply that girls could benefit more than boys from a future intervention.

2. Methods: Line 207-211: Please clarify the description slightly that for the six individual bullying experiences the response options for 0= never or almost never and 1=a few times a year were combined into 0 (a few times a year or less frequently) and the response options for 2= a few times a month and 3= once a week or more were combined into 1 (a few times a month or more frequently). For the combined measure, please make it clear if “0 (never on all types)” corresponds to the “never or almost never” original response to the question, and 1 (any involvement) corresponds to the combination of responses 1, 2, or 3.

3. Methods: Line 225-229: We suggest replacing “immigrant” with “migrant” as this term may imply a more neutral point of view. Please replace immigrant/immigration status throughout, including Table 1.

4. Results: Line 291: We suggest avoiding the term “immigrant” and revising to “...reported a non-native status” or similar.

5. Discussion: Line 344-345: We suggest revising slightly to: “Third, girls who repeated a

grade are more likely to report having experienced specific types of bullying than boys.”

6. Discussion: Line 363: Please revise this sentence if this should read “...have so far focused on its academic outcomes…”

7. Discussion: Line 400-401: We suggest revising this sentence slightly: “The higher risk of experiencing some types of bullying among girls who repeated a grade suggests the possibility that targeted anti-bullying prevention interventions may be especially beneficial for girls.” or similar.

8. Discussion: Lines 438-441: We suggest revising the sentence to read: “The observed link raises the possibility that the widespread educational policy of grade repetition may partly contribute to the differences in bullying victimization and adds to the evidence against the policy of grade repetition; however, our study cannot establish a causal relationship.” or similar.

9. Social Media: To help us extend the reach of your research in the event that your article is accepted for publication, please provide any Twitter handle(s) that would be appropriate to tag, including your own, your coauthors’, your institution, funder, or lab.

Any attachments provided with reviews can be seen via the following link:

[LINK]

8 Oct 2021

Submitted filename: Rebuttal letter5.docx

Click here for additional data file.

11 Oct 2021

Dear Dr Zuo,

On behalf of my colleagues and the Academic Editor, Kathryn Yount, I am pleased to inform you that we have agreed to publish your manuscript "Grade repetition and bullying victimization in adolescents: A global cross-sectional study of the Program for International Student Assessment (PISA) data from 2018" (PMEDICINE-D-21-00390R6) in PLOS Medicine’s Special Issue: Global Child Health: From Birth to Adolescence and Beyond.

Before your manuscript can be formally accepted you will need to complete some formatting changes, which you will receive in a follow up email. Please be aware that it may take several days for you to receive this email; during this time no action is required by you. Once you have received these formatting requests, please note that your manuscript will not be scheduled for publication until you have made the required changes.

In the meantime, please log into Editorial Manager at http://www.editorialmanager.com/pmedicine/, click the "Update My Information" link at the top of the page, and update your user information to ensure an efficient production process.

PRESS

We frequently collaborate with press offices. If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximise its impact. If the press office is planning to promote your findings, we would be grateful if they could coordinate with medicinepress@plos.org. If you have not yet opted out of the early version process, we ask that you notify us immediately of any press plans so that we may do so on your behalf.

We also ask that you take this opportunity to read our Embargo Policy regarding the discussion, promotion and media coverage of work that is yet to be published by PLOS. As your manuscript is not yet published, it is bound by the conditions of our Embargo Policy. Please be aware that this policy is in place both to ensure that any press coverage of your article is fully substantiated and to provide a direct link between such coverage and the published work. For full details of our Embargo Policy, please visit http://www.plos.org/about/media-inquiries/embargo-policy/.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Thank you again for submitting to PLOS Medicine. We look forward to publishing your paper.

Sincerely,

Caitlin Moyer, Ph.D.

Associate Editor

PLOS Medicine