A content analysis of school anti-bullying policies in England: signs of progress

ABSTRACT

Schools in England are required to have an anti-bullying policy. A revised 42-item scoring scheme was used to report a content analysis of 200 anti-bullying policies. On average, school policies had 61% of items. Chi-square comparisons found an increase in policy coverage from 2008 to 2022, notably for mentioning cyber bullying and many types of bias-based bullying; but comparisons are limited by different sampling procedures. Despite good coverage in some areas, fewer than 25% of policies mentioned responsibilities of other school staff, suggested how to help the pupil(s) doing the bullying to change their behaviour, gave advice to parents about bullying, or discussed specific powers to deal with cyberbullying and out-of school bullying. For 131 schools, correlations of self-report scores on bullying victimisation and perpetration with the overall policy score were negative but very small. Ways to improve school policy coverage, and the impact they may have, are discussed.

KEYWORDS:

• Bullying
• victim
• schools
• England
• policy
• content-analysis

Introduction

In England and Wales, the School Standards and Framework Act (SSFA) 1998; Section 61(4), required headteachers to “determine measures … to be taken with a view to … preventing all forms of bullying among pupils. The measures … shall be publicised … in a written document … and [be made] generally known within the school and to parents … at least … once in every school year. This was reiterated in the Education and Inspections Act (EIA) 2006 Section 89(1)(b), which recommended that the anti-bullying policy should form part of the overall school behaviour policy, although a school may decide to issue a separate and specific anti-bullying policy.

The current DFE guidance, Preventing and Tackling Bullying: Advice for Head Teachers, Staff and Governing Bodies (2017), states that “Successful schools have policies in place to deal with bullying and poor behaviour which are clear to parents, pupils and staff so that, when incidents do occur, they are dealt with quickly. However a school chooses to define bullying for the purposes of its own behaviour policy, it should be clearly communicated and understood by pupils, parents, and staff. Successful schools create an environment that prevents bullying from being a serious problem in the first place. School staff, headteachers and governors are best placed to decide how best to respond to the particular issues that affect their pupils. There is no single solution to bullying which will suit all schools” (DfE, 2017, p. 10).

A separate document, Cyberbullying: Advice for Head Teachers and School Staff (2014) states that such policies should cover cyberbullying: “Whole-school policies and practices designed to combat bullying, including cyberbullying, should be developed by and for the whole school community” (DfE, 2014, p. 3). The recent document Behaviour in schools: Advice for headteachers and school staff (DfE, 2022, p. 8, 9) states that a school behaviour policy should aim to “prevent all forms of bullying (including cyberbullying, prejudice-based and discriminatory bullying)” and that it should be “clear and easily understood by pupils, staff and parents”.

Anti-bullying policies set a framework for the actions of the school as regards bullying, and the policy should be readily available in a user-friendly form for parents, pupils, and all members of the school community. In England and Wales it is largely left to schools to devise their own policies, and they vary in scope and quality. There is some web-based guidance available, from a website for school governors, https://schoolgovernors.thekeysupport.com/school-improvement-and-strategy/policies/behaviour-policies/bullying-policies, and through the AntiBullying Alliance website, https://anti-bullyingalliance.org.uk/tools-information/all-about-bullying/whole-school-and-setting-approach/anti-bullying-policies and the Diana Award guide, https://www.antibullyingpro.com/resources/webinar-rewriting-your-schools-anti-bullying-policy-primary-focus and https://www.antibullyingpro.com/resources/webinar-rewriting-your-schools-anti-bullying-policy-secondary-focus.

Nevertheless, the nature and coverage of school anti-bullying policies varies considerably. Smith et al. (2008) devised a 31-item scoring scheme for the coverage provided by school anti-bullying policies, and applied it to 142 school policies, from 115 primary schools and 27 secondary schools, gathered in 2002 from one county in England. On average, schools had about 40% of the items in their policies. In 2008, six years later, Smith et al. (2012) reported a follow-up in the same county; they analysed 217 policies, from 169 primary schools and 48 secondary schools. A slightly expanded 34-item scoring scheme was used. On average, schools had about 49% of the items in their policies, a modest increase. Most included a definition of bullying: including reference to physical, verbal, material and relational forms, and clarifying the difference from other kinds of aggressive behaviour; statements about improving school climate; how sanctions will depend on type or severity of incident; and contact with parents when bullying incidents occurred. However, there was low coverage of cyberbullying; homophobic bullying; bullying based on disabilities; faith; teacher-pupil bullying; responsibilities beyond those of teaching staff; following up of incidents; and specific preventative measures such as playground work, peer support, inclusiveness issues, and bullying to and from school.

Purdy and Smith (2016) analysed 100 policies from schools in Northern Ireland, gathering the data in 2012. They used a 36-item scoring scheme. On average, schools had 52% of the items in their policies. Most schools included reference to physical, verbal, relational, material and cyberbullying but a minority mentioned racist, homophobic, sexual, adult/teacher-pupil bullying or bullying related to disability or religion. Overall the policy scores compared favourably with earlier studies carried out in England; however, a low percentage of Northern Ireland policies gave detailed information about how incidents of bullying would be recorded, who would coordinate this, and how the data would be used to evaluate the effectiveness of the policy.

Context of the current study

The first aim of the current study was to provide a more recent (2022) survey of the coverage of anti-bullying policies in schools in England, based on 200 schools across the country. A second aim was to compare this coverage with those policies obtained earlier in England, in 2008.

There has been only modest evidence that having a good policy translates into lower rates of school bullying or violence. In Welsh schools, a significant association was reported between lower levels of bullying, and pupils reporting that the school had clear rules on bullying (Lambert et al., 2008). Woods and Wolke (2003) found few associations of policy scores with measures of bullying in 34 English primary schools, but the criteria used for scoring policy quality in this study are debatable (see Smith et al., 2008). Smith et al. (2012) related policy scores to pupil self-report survey data on perceptions of and experiences of bullying, available for 78 schools. Most were not significant, although schools with high scores for the section on strategies for preventing bullying did have significantly fewer pupils reporting bullying others (and also fewer being bullied, though non-significantly).

Nikolaou (2017) examined the impact of state anti-bullying laws in states of the USA, implemented from 2000 to 2015. These were effectively policies at state level rather than directly at school level. The analysis showed that schools in states with such laws had up to 8.4% less reported school bullying incidents compared to those in states without anti-bullying laws; these effects were much stronger in states where there was a specific clause in the law defining the term bullying.

Llorent et al. (2021) examined school climate policy documents (including bullying) in 22 schools in Andalusia, Spain, in 2015. The quality of school climate policy documents was not significantly related to social and emotional competencies among students, or to bullying victimisation, cyber perpetration and cyber victimisation. However, high bullying perpetration was related to low quality of the school climate policy document.

In a review, Purdy (2021, p. 264) concludes that while schools are commonly required to create and publish anti-bullying policies, guidance is often inconsistent, meaning that there is considerable variation in the scope and content of the policies, which are generally written to adhere to very diverse district, state, or national guidance and/or statutory requirements; while it might be argued that the anti-bullying policy can provide a useful summary for the school community of the school’s approach to addressing bullying, there are rather mixed results in terms of the association between the measured quality of the policy and the incidence of bullying victimization and perpetration. A third aim of the current study was to examine whether school policy scores related to self-reported bullying victimisation and perpetration scores.

This current study reports first on an analysis of policies from 200 schools, participating in the first year of the United Against Bullying (UAB) programme, organised by the Anti-Bullying Alliance in England. This programme is for schools across England and has a whole-school approach. The overall aim is to reduce bullying and improve pupil wellbeing. It has a particular focus on those children who are most vulnerable and more likely to be victims of bullying behaviour, including disabled pupils and those with special educational needs (SEN/D), and those with other protected characteristics. The programme started in February 2022 with schools completing a school audit, and baseline pupil questionnaire (on wellbeing at school and involvement in bullying). Through the next few months, schools developed and implemented an action plan to tackle bullying and undertook training. These were supported by an UAB online Hub, 12 online training courses, and an interactive Parent Information Tool. By late June 2022, schools reported on progress and a total of 228 schools completed the course. School policies were obtained from 200 of these schools.

In addition, as part of the evaluation of the UAB programme, pupils completed a self-report questionnaire on well-being. This included questions on experiences of bullying victimisation and perpetration; these were available at baseline for 131 of the 200 schools, and allowed an assessment of whether these correlated with overall coverage of policy scores.

Aims of the current study

The first aim was to analyse anti-bullying policies from a wide sample of schools in England, to see how satisfactory their coverage was. The researchers used a slightly revised and expanded version of the scoring scheme from Smith et al. (2012) and Purdy and Smith (2016). The study also compared coverage in primary and secondary schools. A second aim was to make some comparison with findings from the 2012 study (data gathered in 2008), bearing in mind that different schools were sampled. A third aim was to examine whether experiences of bullying victimisation and perpetration correlated with the total policy score.

Method

Participants

Anti-bullying policies were obtained from 200 schools which had completed the United Against Bullying programme by June 2022. Policies were obtained through direct contact with the schools and request via email, or the schools’ websites where certain policies are uploaded as per the legal standard requirements. These were from 136 (68%) primary schools and 64 (32%) secondary schools. The self-reported experiences of bullying victimisation and perpetration were available from 131 of these schools (94 primary; 37 secondary).

Procedure

The scoring scheme for the anti-bullying policies started by examining the coding scheme used by Smith et al. (2012), with 34 items, expanded to 36 items in Purdy and Smith (2016). Given developments in the last decade, some revision and expansion seemed justified. This was discussed further with colleagues and practitioners (see Acknowledgements), and then a revised version trialled on 10 anti-bullying policies. A few modifications were made, and a further revised scheme used on another 10 policies from the main sample. Further minor clarifications were again made. The final list of items can be seen in Table 1. The 42 categories were divided, as previously, into four sections: (A) 15 categories concerning the definition of bullying behaviour; (B) 11 categories concerning reporting and responding to bullying incidents; (C) 9 categories concerning recording bullying and communicating and evaluation the policy; (D) 7 categories concerning strategies for preventing bullying.

Table 1. Number of schools scoring yes on each item (percentage in brackets), for each section (A, B, C, D) and for total anti-bullying content (in bold); for all schools, and separately for primary and secondary schools. Significant differences between primary and secondary schools shown in bold.

	All Schools (N=200)	Primary Schools (N=136)	Secondary Schools (N=64)	Primary vs Secondary
A Definition of bullying behaviour (15 points)
1 Does the policy have a definition of bullying?	186 (93)	125 (92)	61 (95)	χ2(1)=.509 p=.354
2 Does the definition make it clear that bullying is different from other kinds of aggressive behaviour? (e.g. mention power imbalance)	116 (58)	79 (58)	37 (58)	χ2(1)=.018 p=.506
3 Mention physical bullying (hits, kicks)?	197 (99)	133 (98)	64 (100)	χ2(1)=.951 p=.461
4 Mention direct verbal or emotional bullying (threats, insults, nasty teasing)?	198 (99)	134 (99)	64 (100)	χ2(1)=.473 p=.680
5 Mention relational bullying (rumours, social exclusion)?	189 (95)	129 (95)	60 (94)	χ2(1)=.310 p=.403
6 Mention material bullying (hiding or damaging belongings, extortion of money)?	171 (86)	116 (85)	55 (86)	χ2(1)=.000 p=.572
7 Mention cyberbullying (email, text messages, internet, social media, gaming)?	192 (96)	128 (94)	64 (100)	χ2(1)=3.414 p=.064
8 Mention homophobic or biphobic bullying (due to sexual orientation)?	178 (89)	117 (86)	61 (95)	χ2(1)=3.384 p=.050
9 Mention transphobic bullying?	151 (76)	98 (72)	53 (83)	χ2(1)=2.395 p=.083
10 Mention racist bullying (or harassment)?	175 (88)	117 (86)	58 (91)	χ2(1)=.614 p=.297
11 Mention sexual and sexist bullying (or harassment)?	175 (88)	116 (85)	59 (92)	χ2(1)=1.563 p=.154
12 Mention bullying due to special education needs and disabilities?	154 (77)	105 (77)	49 (77)	χ2(1)=.047 p=.481
13 Mention bullying because of faith, religion or beliefs?	148 (74)	99 (73)	49 (77)	χ2(1)=.211 p=.392
14 As well as pupil-pupil bullying, mention adult/teacher-pupil bullying or vice versa?	66 (33)	45 (33)	21 (33)	χ2(1)=.001 p=.551
15 Mention the issue of bullying on the way to school or happening outside school? (e.g. shops, town centre; not cyber)	112 (56)	68 (50)	44 (69)	χ2(1)=5.747 p=.012
B Reporting and responding to bullying incidents (11 points)
*1 State what victims/targets of bullying should do (e.g. tell a teacher; should clearly apply to victims/pupils who experience bullying)?	111 (56)	68 (50)	43 (67)	χ2(1)=4.781 p=.020
2 Say how teaching staff should respond to a report of bullying (should specifically mention bullying, and be more specific than just ‘deal promptly’)?	127 (64)	90 (66)	37 (58)	χ2(1)=1.314 p=.161
3 Clearly mention the responsibilities of other school staff (teaching assistants, lunchtime supervisors, etc) if they know of bullying? (more specific than simply referring to ‘all staff’)	19 (10)	14 (10)	5 (8)	χ2(1)=.312 p=.392
4 Clearly mention the responsibilities of parents if they know of bullying (this can include knowing if their child has a behaviour problem if bullying is included elsewhere)?	96 (48)	63 (46)	33 (52)	χ2(1)=.353 p=.329
5 Clearly mention the responsibilities of pupil bystanders if they know of bullying?	89 (45)	55 (40)	34 (53)	χ2(1)=2.836 p=.063
6 State whether the approach used for bullying will depend on type or severity of incident? (it should be clear that the approach or sanctions apply to bullying behaviour)	84 (42)	51 (38)	33 (52)	χ2(1)=3.163 p=.052
7 Mention follow-up to see whether the approach or sanctions were effective?	126 (63)	90 (66)	36 (56)	χ2(1)=2.134 p=.097
8 Discuss what action will be taken if the bullying persists?	79 (40)	49 (36)	30 (47)	χ2(1)=1.854 p=.114
9 Suggest how to support for the victim/target? (more than just ‘we will support victims’)	119 (60)	80 (59)	39 (61)	χ2(1)=.081 p=.450
10 Suggest how to help the pupil(s) doing the bullying to change their behaviour (apart from sanctions)? (more than just ‘we will support …’)	42 (21)	24 (18)	18 (28)	χ2(1)=2.880 p=.067
11 discuss if, when or how parents will be informed? (‘parents will be informed’ is sufficient if it clearly refers to bullying)	181 (91)	125 (92)	56 (88)	χ2(1)=1.408 p=.177
C Recording bullying, communicating and evaluating the policy (9 points)
1 Say reports of bullying will be recorded?	185 (93)	127 (93)	58 (91)	χ2(1)=.815 p=.266 p=.266
2 Say who is responsible for co-ordinating the recording system?	79 (40)	55 (40)	24 (38)	χ2(1)=.158 p=.406
3 Show how records or survey data will be used to know whether the policy is working or not?	107 (54)	77 (57)	30 (47)	χ2(1)=1.923 p=.109
4 Clearly mention periodic review and updating of the policy?	170 (85)	112 (82)	58 (91)	χ2(1)=1.994 p=.114
5 Clearly mention consultation with parents about the policy?	69 (35)	53 (39)	16 (25)	χ2(1)=4,137 p=.029
6 Clearly mention consultation with pupils about the policy?	72 (36)	52 (38)	20 (31)	χ2(1)=.922, p=.212
7 Mention the role of the school governors (or governing body)?	167 (84)	118 (87)	49 (77)	χ2(1)=3.874 p=.041
8 Reference the Equality Act 2010?	108 (54)	66 (49)	42 (66)	χ2(1)=4.697 p=.021
9 Mention alignment with other relevant policies (e.g. safeguarding, behaviour)	164 (82)	110 (81)	54 (84)	χ2(1)=.229 p=.396
D Strategies for preventing bullying (7 points)
1 Mention any of encouraging co-operative behaviour, rewarding good behaviour, improving school climate, or creating a safe environment?	197 (99)	133 (98)	64 (100)	χ2(1)=.951, p=.461
2 Discuss general issues of peer support (ways of pupils helping victims/targets, beyond B5)?	76 (38)	48 (35)	28 (44)	χ2(1)=1.321 p=.160
3 Discuss specific powers and/or strategies to deal with cyber bullying	31 (16)	18 (13)	13 (20)	χ2(1)=1.664 p=.140
4 Discuss specific powers and/or strategies to deal with out of school bullying (not cyber)	45 (23)	23 (17)	22 (34)	χ2(1)=7.611 p=.006
5 Give advice for parents about bullying (clearly beyond B4)?	39 (20)	24 (18)	15 (23)	χ2(1)=.930, p=.218
6 Mention the preventative role of playground activities or lunchtime supervisors?	73 (37)	57 (42)	16 (25)	χ2(1)=5.370 p=.014
7 Discuss issues of inclusiveness (e.g. support for non English speakers; pupils with learning difficulties; vulnerable pupils)?	66 (33)	47 (35)	19 (30)	χ2(1)=.467 p=.303
Total (42 items)	25.65	25.28	26.42	t(198)=0.37 p=.714

The changes made to the earlier scoring protocol (Smith et al., 2012) were as follows:

Section A:

One new item was added. This was A9, “Mention transphobic bullying”. This is a topic gaining recognition in recent years (Apostolidou, 2021). Also a category A15, “Mention the issue of bullying on the way to school or happening outside school? (for example, shops, town centre; not cyber)” was moved up from D6 previously and the expansion added. This issue has received more focus recently (for example, Betts et al., 2019; Vodden & Noret, 2019). DfE guidance (n.d.) states that “Head teachers have the legal power to make sure pupils behave outside of school premises (state schools only).This includes bullying that happens anywhere off the school premises, for example on public transport or in a town centre”. Minor changes were made for A2 “Does the definition make it clear that bullying is different from other kinds of aggressive behaviour?”; the clarification “ (for example, mention power imbalance)” was added. For A7 “Mention cyberbullying” the clarification was expanded from “(email, text messages)” to “(email, text messages, internet, social media, gaming)”. For A8, “Mention homophobic bullying” was expanded to “Mention homophobic or biphobic bullying (due to sexual orientation)”. A11 (formerly A10) was expanded from “Mention sexual bullying (or harassment)” to “Mention sexual and sexist bullying (or harassment)”. A12 was expanded from “Mention bullying due to disabilities” to “Mention bullying due to special education needs and disabilities”. A13 was changed slightly, from “Mention bullying because of faith or religious beliefs” to “Mention bullying because of faith, religion or beliefs”.

Section B:

No new items were added, but the wording of several items was slightly altered. For B1, from “State what victims of bullying should do (for example, tell a teacher; should clearly apply to victims/pupils who experience bullying)?” to “State what victims/targets of bullying should do (for example, tell a teacher; should clearly apply to victims/pupils who experience bullying)”. For B6, from “State whether sanctions applied for bullying will depend on type or severity of incident? (it should be clear that the approach or sanctions apply to bullying behaviour)” to “State whether the approach used for bullying will depend on type or severity of incident? (it should be clear that the approach or sanctions apply to bullying behaviour)”. For B7, from “Mention follow-up to see whether the sanctions were effective” to “Mention follow-up to see whether the approach or sanctions were effective”. For B9, from “Suggest how to support for the victim? (more than just ‘we will support victims’)” to “Suggest how to support for the victim/target? (more than just ‘we will support victims’)”.

Section C:

Included the two additional items added by Purdy and Smith (2016), namely C5 “Clearly mention consultation with parents” and C6 “Clearly mention consultation with pupils”, but adding “about the policy” in each case. In addition, three new items were added, C7 “Mention the role of the school governors (or governing body)”, C8 “Reference the Equality Act 2010” and C9 “Mention alignment with other relevant policies (for example, safeguarding, behaviour)”.

Section D:

One item was added, ‘D3 “Discuss specific powers and/or strategies to deal with cyber bullying” (which schools have, see DfE, 2014). Minor changes were made for D2, expanding “Discuss general issues of peer support (beyond B5)” to “Discuss general issues of peer support (ways of pupils helping victims/targets, beyond B5)” and D7 (formerly D6) from “Discuss issues of inclusiveness (for example, support for non-English speakers; pupils with learning difficulties)” to “Discuss issues of inclusiveness (for example, support for non-English speakers; pupils with learning difficulties; vulnerable pupils)”.

Policy scoring

The total number of pages of the policy document was counted; this included any cover pages, but any appendices were not counted. Then, each of the 42 categories was rated one or zero, based on whether the item was or was not present in the policy; thus, the total policy score could range from 0 to 42. Independent scoring of 20 policies resulted in an overall inter-rater agreement of 87%. Statistical analyses were carried out using SPSS version 27.

Bullying victimisation and perpetration

Pupils at the schools (participating in the UAB programme) filled in a self-report questionnaire which included 10 items measuring being bullied, and bullying others. The five items of being bullied experiences were: I am hit, pushed or kicked by other pupils; I am called mean names by other pupils; Other pupils stop me from joining in with them; Other pupils say bad things about me when I’m not there; Other pupils are mean or rude to me online. The five items about bullying others were correspondingly: I hit, push or kick other pupils; I call other pupils mean names; I stop other pupils joining in with me; I say bad things about other pupils when they aren’t there; I am mean or rude to other pupils online. All items were responded to on a four-point Likert scale, never (0), a little (1), a lot (2) or always (3). Being bullied and bullying others were analysed separately, and categorised as never, ever, or frequently (bullied, or bullying others). “Never” meant that the pupil ticked “Never” for all 5 items. “Ever” meant that the pupil ticked “a little” or more for at least one item. “Frequently” meant that the pupil scored 2 or more for at least one item. Average school scores for ever and frequent bullied, or bullying others, were entered into correlations with total policy score. These scores were available for 131 of the 200 schools (94 primary, 37 secondary).

Results

Length of policies

The mean number of pages for the policies was 9.5, with a range from 1 to 35. Secondary schools had a mean of 10.9 pages on average, longer than for primary schools with a mean of 8.7 pages. The correlation between total anti-bullying policy content scores and the number of pages of each of the school policies was significant, r(200) = .421, p < .001, meaning that longer policies covered more criteria.

Overall item scores

Table 1 shows the number of schools (percentage in brackets) whose policies were scored for presence of each of the 42 categories. Looking first at the findings for all schools (n = 200), an overall average of 61% of categories was present.

Most categories in Section A were present in the majority of school policies; the different types of school bullying were generally well covered. An exception was A14 “As well as pupil-pupil bullying, mention adult/teacher-pupil bullying or vice versa?” which only one-third of policies mentioned.

Section B categories were mostly scored in around half of policies, but with two noticeably low scores: B3 “Clearly mention the responsibilities of other school staff (teaching assistants, lunchtime supervisors, etc) if they know of bullying? (more specific than simply referring to ‘all staff’)” was only mentioned by 9.5% of schools, and B10 “Suggest how to help the pupil(s) doing the bullying to change their behaviour (apart from sanctions)? (more than just ‘we will support …’)” by only 21%.

Section C categories varied in scoring, with most schools saying that reports of bullying would be recorded (C1). The lowest scores, just over one-third of schools, were for C5 and C6, whether the policies mentioned consultation with parents, or pupils.

Section D categories showed the lowest scores, apart from D1 – almost all school policies mentioned “encouraging co-operative behaviour, rewarding good behaviour, improving school climate, or creating a safe environment”. Other items were only present in a minority of policies, with particularly low scores for discussing specific powers to deal with cyberbullying (D3) and out- of school bullying (D4), and giving advice to parents about bullying (D5).

Comparing primary and secondary schools

Type of school (primary, secondary) differences were analysed by chi-square for individual criteria, and t test for total scores, with the results shown in Table 1. No differences were found between primary and secondary schools for overall content score. Individual criteria were compared and a total of eight differences were found (shown in bold in Table 1). Primary school policies were more likely to include criteria C5 (consultation with parents), C7 (role of school governors) and D6 (role of playground supervisors), all arguably more salient in smaller schools and with younger pupils. Secondary school policies were more likely to include criteria A8 (homophobic bullying), A15 (bullying outside school), B1 (what victims of bullying should do), C8 (reference the Equality Act), and D4 (powers to deal with out-of-school bullying). Apart from B1, these might be seen as more relevant in the secondary sector.

Comparing the present data with scores from Smith et al. (2012)

There were 34 items in common between the survey carried out in 2008, and the current survey carried out in 2022. A few items changed slightly in wording (see above), and Table 2 shows the earlier wording. Comparisons are made for primary schools and secondary schools separately; these were analysed by chi-square for individual criteria, and t tests for section and total scores, with the results shown in Table 2. Given limitations to the comparison, and the number of comparisons made, the survey only highlights appreciable differences found in both primary and secondary schools, significant at p < .01level and bolded in Table 2.

Table 2. Comparing the policy scores from 2008 to 2022, for each of the 34 items of the content analysis that were in common. Mean scores for each section (A, B, C, D) (standard deviations in brackets), and the number of schools (percentage in brackets) scoring yes on each item. Differences significant at p < .01or beyond, and found in both primary and secondary schools, are in bold.

		Primary schools 2008 (N = 169)	Primary schools 2022 (N = 136)	Primary schools 2008 vs 2022	Secondary schools 2008 (N = 48)	Secondary schools 2022 (N = 64)	Secondary schools 2008 vs 2022
A	Definition of bullying behaviour (13 items)	6.4 (3.0)	9.5 (7.0)	t(1) = 5.20***	8.1 (2.7)	9.9 (15.5)	t(1) = 4.81**
1	Have a definition of bullying?	148 (87.6)	125 (92.0)	χ2(1) = 3.98*	44 (91.7)	61 (95.3)	χ2(1) = 3.76*
2	Does the definition make it clear that bullying is different from other kinds of aggressive behaviour?	120 (71.0)	79 (58.1)	χ2(1) = 3.54*[decrease]	32 (66.7)	37 (57.8)	χ2(1) = 3.64* [decrease]
3	Mention physical bullying (hits, kicks)?	140 (82.8)	133 (97.8)	χ2(1) = 3.1*	46 (95.8)	64 (100)	χ2(1) = 3.84*
4	Mention direct verbal bullying (threats, insults, nasty teasing)?	136 (8.5)	134 (98.5)	χ2(1) = 3.59*	45 (93.8)	64 (100)	χ2(1) = 3.86*
5	Mention relational bullying (rumours, social exclusion)?	128 (75.7)	129 (94.9)	χ2(1) = 6.54**	42 (87.5)	60 (93.8)	χ2(1) = 3.84*
6	Mention material bullying (damage to belongings, extortion of money)?	117 (69.2)	116 (85.3)	χ2(1) = 6.64**	38 (79.2)	55 (85.9)	χ2(1) = 8.79*
7	Mention cyberbullying (email, text messages)?	45 (26.6)	128 (94.1)	χ2(1) = 1.31***	25 (52.1)	64 (100)	χ2(1) = 10.89***
8	Mention homophobic bullying?	32 (18.9)	117 (86.0)	χ2(1) = 1.96***	22 (45.8)	61 (95.3)	χ2(1) = 10.90***
9	Mention racial bullying (or harassment)?	101 (59.8)	117 (86.0)	χ2(1) = 5.96**	37 (77.1)	58 (90.6)	χ2(1) = 3.71,*
10	Mention sexual bullying (or harassment)?	72 (42.6)	116 (85.3)	χ2(1) = 1.96***	32 (66.7)	59 (92.2)	χ2(1) = 5.64**
11	As well as pupil-pupil bullying, discuss the issue of adult/teacher-pupil bullying or vice versa?	12 (7.1)	45 (33.1)	χ2(1) = 1.93***	6 (12.5)	21 (32.8)	χ2(1) = 10.91***
12	Mention bullying due to disabilities?	20 (11.8)	105 (77.2)	χ2(1) = 1.81***	12 (25.0)	49 (76.6)	χ2(1) = 10.91***
13	Mention bullying because of faith or religious beliefs?	10 (5.9)	99 (72.8)	χ2(1) = 1.96***	6 (12.5)	49 (76.6)	χ2(1) = 10.85***
B	Reporting and responding to bullying incidents (11 items)	5.6 (2.0)	5.2 (3.8)	t(1) = 1.18*	5.9 (2.2)	5.7 (8.9)	t(1) = 0.15
1	State what victims of bullying should do (e.g. tell a teacher; should clearly apply to victims/pupils who experience bullying)?	106 (62.7)	68 (5.0)	χ2(1) = 3.80*[decrease]	32 (66.7)	43 (67.2)	χ2(1) = 1.72*
2	Say how teaching staff should respond to a report of bullying (should specifically mention bullying, and be more specific than just “deal promptly”)?	113 (66.9)	90 (66.2)	χ2(1) = 4.21	28 (58.3)	37 (57.8)	χ2(1) = 3.19* [decrease]
3	Clearly mention the responsibilities of other school staff (teaching assistants, lunchtime supervisors etc) if they know of bullying? (this should be more specific than simply referring to “all staff”)	28 (16.6)	14 (1.3)	χ2(1) = 3.84*[decrease]	2 (4.2)	5 (7.8)	χ2(1) = 3.82*
4	Clearly mention the responsibilities of parents if they know of bullying (this can include knowing if their child has a behaviour problem if bullying is included elsewhere)?	98 (58.0)	63 (46.3)	χ2(1) = 3.79*[decrease]	24 (50.0)	33 (51.6)	χ2(1) = 3.84*
5	Clearly mention the responsibilities of pupil bystanders if they know of bullying?	94 (55.6)	55 (4.4)	χ2(1) = 3.66*[decrease]	36 (75.0)	34 (53.1)	χ2(1) = 6.49** [decrease]
6	State whether sanctions applied for bullying will depend on type or severity of incident? (it should be clear that the sanctions apply to bullying behaviour)	110 (65.1)	51 (37.5)	χ2(1) = 6.64**[decrease]	37 (77.1)	33 (51.6)	χ2(1) = 6.46** [decrease]
7	Mention follow-up to see whether the sanctions were effective?	54 (32.0)	90 (66.2)	χ2(1) = 6.61**	19 (39.6)	36 (56.3)	χ2(1)-6.64**
8	Discuss what action will be taken if the bullying persists?	37 (21.9)	49 (36.0)	χ2(1) = 3.84*	13 (27.1)	30 (46.9)	χ2(1) = 6.41**
9	Suggest how to support for the victim? (more than just “we will support victims”)	83 (49.1)	80 (58.8)	χ2(1) = 3.80*	24 (50.0)	39 (60.9)	χ2(1) = 3.34*
10	Suggest how to help for the pupil(s) doing the bullying to change their behaviour (apart from sanctions)? (more than just “we will support …”)	71 (42.0)	24 (17.6)	χ2(1) = 6.60**[decrease]	23 (47.9)	18 (28.1)	χ2(1) = 6.64** [decrease]
11	Discuss if, when or how parents will be informed? (“parents will be informed” is sufficient if it clearly refers to bullying)	154 (91.1)	125 (91.9)	χ2(1) = 3.1	44 (91.7)	56 (87.5)	χ2(1) = 3.31* [decrease]
C	Recording bullying, communicating and evaluating the policy (4 items)	2.3 (1.4)	2.7 (2.0)	t(1) = 1.30*	1.8 (1.4)	2.7 (4.2)	t(1) = 2.23*
1	Say reports of bullying will be recorded?	130 (76.9)	127 (93.4)	χ2(1) = 6.65**	34 (70.8)	58 (90.6)	χ2(1) = 6.55**
2	Say who is responsible for co-ordinating the recording system?	73 (43.2)	55 (4.4)	χ2(1) = 3.21*[decrease]	18 (37.5)	24 (37.5)	No Change
3	Show how records or survey data will be used to know whether the policy is working or not?	62 (36.7)	77 (56.6)	χ2(1) = 5.35**	13 (27.1)	30 (46.9)	χ2(1) = 6.30**
4	Mention periodic review and updating of the policy?	119 (7.4)	112 (82.4)	χ2(1) = 3.80*	22 (45.8)	58 (90.6)	χ2(1) = 9.81**
D	Strategies for preventing bullying (6 items)	2.1 (1.0)	2.3 (1.7)	t(1)= .18	2.3 (1.2)	2.2 (3.4)	t(1) = 0.22
1	Mention any of: encouraging co-operative behaviour, rewarding good behaviour, improving school climate, or creating a safe environment?	147 (87.0)	133 (97.8)	χ2(1) = 3.89*	40 (83.3)	64 (100)	χ2(1) = 6.56**
2	Discuss general issues of peer support (beyond B5)?	43 (25.4)	48 (35.3)	χ2(1) = 3.86*	22 (45.8)	28 (43.8)	χ2(1) = 3.62* [decrease]
3	Discuss advice for parents about bullying (beyond B4)?	72 (42.6)	24 (17.6)	χ2(1) = 6.65**[decrease]	25 (52.1)	15 (23.4)	χ2(1) = 6.62** [decrease]
4	Mention the preventative role of playground activities or lunchtime supervisors?	50 (29.6)	57 (41.9)	χ2(1) = 6.64**	10 (20.8)	16 (25.0)	χ2(1) = 3.83*
5	Discuss issues of inclusiveness (e.g. non English speakers; pupils with learning difficulties)?	5 (3.0)	47 (34.6)	χ2(1) = 1.79***	3 (6.3)	19 (29.7)	χ2(1) = 10.86***
6	Mention the issue of bullying on the way to school or happening outside school?	29 (17.2)	68 (50)	χ2(1) = 1.78***	11 (22.9)	44 (68.8)	χ2(1) = 6.60**
	Total anti-bullying policy content (34 items)	16.3 (4.7)	23.6 (8.1)	t(1) = 9.84***	18.1 (5.2)	20.5 (6.1)	t(1) = 4.80***

*** p < 0.001; ** p < 0.01; * p < 0.05.

On average, the total policy score increased on these 34 items, very appreciably for primary schools (from 16.3 to 23.6, or 48% to 69%), less so for secondary (from 18.1 to 20.5, or 53% to 60%). Overall, 66% of items were covered. There were marked increases in Section A (definition), notably in mentioning cyber bullying (A7) and many types of bias-based bullying (homophobic, sexual, disabilities, faith; A8,10,12,13). There was also an increase in mentioning adult/teacher bullying or vice versa (A11), but from a low baseline.

For Section B (reporting), changes were relatively small. There was an increase in mentioning follow-up to see whether sanctions were effective (B7); but a decrease in mentioning grading of sanctions (B6), and helping pupils doing the bullying to change their behaviour (B10).

For Section C (recording) there was improvement in coverage for stating that reports would be recorded (C1), and how the data will be used (C3).

For Section D (strategies for prevention) there was an increase in coverage for issues of inclusiveness (D5), and for mentioning bullying on the way to or outside school (D6). However, there was a decrease in covering advice to parents about bullying (D3).

Correlations between policy scores, and self-reported bullying victimisation and perpetration

The correlations amongst these variables are shown in Table 3, for the 131 schools where all the data were available. In line with much previous literature, there are quite high correlations between the variables of ever and frequent involvement and of victimisation and perpetration (Espino et al., 2022). The main interest here is on the correlations of these four variables, with the total policy score for each school. These are all negative, in line with the expectation that schools with greater policy coverage may have lower pupil reported bullying victimisation and perpetration. However, the correlations are very small (not reaching statistical significance), indicating at best a very small impact of the policy on these behavioural measures.

Table 3. Correlations of total policy score with 4 measures of bullying victimisation and perpetration, across 131 schools.

	Policy score	Ever victim	Frequent victim	Ever bully
Ever victim	−.018
Frequent victim	−.076	.564**
Ever bully	−.049	.328**	.615**
Frequent bully	−.101	.177*	.566**	.726**

Discussion

The analysis of school anti-bullying policies presented here gives an idea of their content, and strengths and weaknesses in coverage, for a sample of 200 schools across England. These schools are not a random sample, as they chose to be part of the ABA “United Against Bullying” programme; as such, they might be expected to be more concerned about the issue than schools chosen randomly. However it is a national sample, and there is the advantage that most of the schools provided self-report data on bullying victimisation and perpetration rates, to compare with policy scores.

The school policies varied greatly in length, from 1 to 35 pages. There was a positive correlation of length with coverage, and one page is certainly too short to cover the range of items in the four sections of the coding scheme. The mean length was around 10 pages. However, and even as part of a general behaviour policy, those policies much longer than this sometimes had unnecessary material and risked compromising the requirement that “it should be clearly communicated and understood by pupils, parents, and staff” (DfE, 2017).

Using a slightly revised and expanded coding scheme, overall item coverage was 61%.

Notably (see Table 1), most policies mentioned the various types of bullying (Section A), said that reports of bullying would be recorded (C1), mentioned the role of school governors (C7) and encouraged cooperative behaviour and a safe environment (D1). However, there were some clear weaknesses. For example, only one-third of school policies mentioned adult/teacher-pupil bullying or vice versa (A14), only one policy in 10 clearly mentioned the responsibilities of other school staff (teaching assistants, lunchtime supervisors, etc.) if they know of bullying (B3), and only one in five suggested how to help the pupil(s) doing the bullying to change their behaviour (apart from sanctions) (B10). Although almost all policies mentioned cyberbullying (A7), only one in seven discussed specific powers to deal with cyberbullying (D4), although they have these powers (DfE, 2014). Also, only one in four discuss specific powers to deal with out-of school bullying (D4), although they have these (DfE, 2017). Only one-fifth of policies included advice to parents about bullying (D5). All of these are clearly areas where many school policies could give better coverage.

Primary and secondary schools did not differ much in overall coverage; there were some differences on individual items, for example more mention of the role of lunchtime supervisors (D6) in primary school policies, more mention of homophobic bullying (A8) in secondary policies; these differences were largely understandable in terms of school size and age range of pupils.

Despite some shortcomings in many policies, there is evidence that their coverage is improving. At a very global level, item coverage was 40% in 2002, 49% in 2008, 52% in Northern Ireland in 2012, and now 61% in 2022. The item scale used increased over this time period, but Table 2 shows the 2008 to 2022 comparison for very similar items, with an increase from 49% to 66%. Comparisons over time are also limited by the changing sample of schools; in particular the earlier (2002, 2008) samples were from one county in England, whereas the 2022 data is a national sample. Given these limitations, this study only highlighted appreciable differences, significant at the p < .01 level, and consistent across primary and secondary sectors. Even so, the changes shown in Table 2 should be treated with caution.

One obvious and expected increase is in policies mentioning cyberbullying, from 27% to 94% in primary schools and 52% to 100% in secondary schools. This form of bullying was still relatively new in 2008, but is now fully recognised as a major type of bullying, with its own set of advice from DfE (2014); it is reassuring that almost all policies now mention this. There are also very large increases in mention of bias- or prejudice-based bullying, notably for homophobic bullying (A8), sexual bullying/harassment (A10), disabilities (A12) and faith/religious beliefs (A13). The Public Sector Equality Duty (Ministry of Justice, 2012) means that schools that receive public funding are also required to ensure their policies are effective at tackling prejudice-based bullying, so these changes are to be expected. There are also increases in coverage of adult/teacher-pupil bullying or vice versa (A11), but as noted above from a very low baseline (from about 1 in 10 to 1 in 3 policies).

There has been an increase in items concerning use of the policy, notably mentioning follow-up to see if sanctions were effective (B7), recording reports of bullying (C1) and saying how these will be used (C3). Mentioning the issue of bullying outside school increased (D6, recoded as A15 in Table 1), even though only one in four discuss specific powers to deal with it (Table 1, D4). Discussing issues of inclusiveness increased from a low baseline (D5). However, there were some worrying decreases on a few items: these were appreciable and consistent for stating whether sanctions will depend on type or severity of incident (B6), suggesting how to help pupils doing the bullying to change their behaviour (B10), and discussing advice for parents about bullying (D3).

The third aim of the review was to examine whether policy coverage was associated with rates of bullying victimisation or perpetration in the schools. The limited previous research has generally found quite small associations, even if usually in the expected direction. This is exactly what was found in the current review. The four relevant correlations in Table 3 are all negative (lower bullying rates in schools with better policies), but very small, and not statistically significant. It would appear from this, and previous findings (Purdy, 2021), that rates of bullying victimisation or perpetration depend on many factors. It is known that these include general school and class climate, proactive curricula activities, consistent and effective action by teachers, and support from parents. A good school anti-bullying policy may provide an important framework, and a support for pupils and parents; but even more important may be how the policy is actually communicated and put into practice, throughout the school community.

Implications for practice/conclusion

It may be useful for schools to review their policies, in line with the revised 42-item scoring scheme, and in light of the findings reported here. Educational psychologists might encourage schools by pointing out the suggested policy improvements over time, while also focussing on common areas of development needed, notably mentioning the responsibilities of other school staff, suggested how to help the pupil(s) doing the bullying to change their behaviour, giving advice to parents about bullying, and being explicit about the specific powers that they have to deal with cyberbullying and out-of school bullying. Engaging students and parents as well could be useful, also in evaluating the length and intelligibility of policy documents.

Acknowledgments

Thanks to Martha Evans and Liffy McDonnell at the AntiBullyingAlliance for giving access to the list of United Against Bullying schools; Susanne Robinson and Catherine Culbert for providing the bullying victimization and perpetration scores for each school; and to all of them and also to Noel Purdy for comments on the revision of the scoring scheme. Special thanks to Goldsmiths Research Internship Programme (GRIP) for organising and funding the internship for Ifraah Kidwai.

Disclosure statement

No potential conflict of interest was reported by the authors.