Feeling Part of the School and Feeling Safe: Further Development of a Tool for Investigating School Belonging

ABSTRACT

A sense of school belonging has been associated with a number of positive outcomes for young people Two core aspects are feeling part of the school, and feeling safe. This paper examines qualitative data from a mixed methods questionnaire designed to inform schools on the barriers and supports to participation in school life and the relationship with students’ feelings of belonging. It contributes to the limited existing qualitative data and can be set alongside students’ quantitative measures to provide new insights. We explore responses from 595 students, to the open questions of a survey. Students frequently mentioned relationships with teachers and peers, often in the context of feeling safe to be themselves, suggesting that feeling part of the school and feeling safe are intertwined. The comments reveal the importance of being recognised and accepted and not having their identity purely defined in relation to attainment on a narrowly defined curriculum.

KEYWORDS:

• School belonging
• school connectedness
• safety
• feeling part of the school

Introduction

A sense of school belonging has been associated with a number of positive outcomes for young people and a mediating factor for offsetting anti-social and health risk behaviours. The role of schools in fostering student well-being is particularly relevant in the current context in which children and young people have experienced considerable periods of time outside school. Transitioning back and re-connecting with school takes on heightened significance with new groups of young people vulnerable to exclusion (Daniels et al 2019). Pre-pandemic, OECD data indicated that 20% of 15 year olds feel like an outsider at school (OECD 2019). It is particularly timely that schools (and governments) look at systemic aspects that contribute to a lack of connectedness between young people and schools, rather than pathologize and exclude those who don’t have a sense of belonging.

School connectedness and school belonging have an extensive literature largely based on quantitative studies and their meta-analysis. While these are valuable in indicating relationships between a variety of measures of attainment and well-being, they provide limited insight into the complexity and diversity of learners’ views. In contrast there are few studies which gather qualitative data and which surface learners’ views and analyse these in the context of students’ self-ratings of belonging. Our purpose here is two- fold: namely to start to address that gap through developing a mixed-method questionnaire that schools can use to learn about the barriers and supports to participation in school life and students sense of belonging; and secondly, to examine the relationship between the quantitative ratings of aspects of belonging, feeling part of the school, and feeling safe, and students views about the barriers and supports to positive feelings.

This forms part of a wider project in which we have been working collaboratively with schools to develop a tool that supports them in responding to the diverse needs of learners, and to shift the focus away from individual deficit-based responses to a more systemic approach, one that recognise the ways in which the policies, practices and environment of the school can be experienced as exclusionary. Systemic responses require the views of all learners and we know from previous nationwide surveys that the choice of instrument for many schools is the online questionnaire (Porter 2015).

Literature review

There are many different descriptors of belonging and a variety of terms that are used interchangeably, with school connectedness often used as an overarching indication of students’ relationship to school” (Libbey 2004, 274). We draw here on the work of Goodenow (1993) who defined students’ sense of belonging as “ … the extent to which students feel personally accepted, respected, included and supported in the school social environment” (p.80), a definition that is one of the most frequently used in research (Allen and Kern 2017). St-Armand, Girard, and Smith (2017)’s review identifies four defining attributes of belonging: feeling positive emotions (closeness to others, a sense of usefulness) positive relationships with peers and teachers; actively participating, in and outside the classroom; and harmonisation, aligning with situations or people (i.e. not feeling the odd one out). Allen and Kern (2017) identify consistencies despite the use of different terms in focusing on having a place within the school, attachment to others, and a sense of inclusion. These are typically operationalised into “school-based relationships and experiences”, “student-teacher relationships” and “student general feelings about school as a whole” (Allen and Kern 2017, 17).

Despite variation in terminology there is a strong degree of consensus about the importance of school belonging with a wide body of empirical literature exploring the relationship between students’ sense of belonging or connectedness to school and aspects of young people’s wellbeing. These include studies identifying the positive relationship between belonging and self-esteem (Watson 2017); academic outcomes (Niehaus, Moritz Rudasill, and Rakes 2012); mental health (Millings et al. 2012); and health risk behaviours (Resnick, Bearman, and Blum et al. 1997). Indeed, a sense of belonging is often described as a buffer or protective factor (Millings et al. 2012).

Single studies have produced mixed results in identifying which factors have the biggest impact on school belonging. Allen and Kern 2017) meta-analysis examines the association between school belonging and ten groups of variables across 51 studies. Teacher support and personal characteristics (e.g. conscientiousness, social intelligence, positive affect, hope, coping ability) were found to have the largest effect sizes. Surprisingly, given the emphasis on friendship elsewhere in the literature (St-Armand, Girard, and Smith 2017) parents and teachers play a more important role than peers. However, Allen and Kern 2017) ecosystemic approach to analysis draws attention to the complexity of singling out particular sources of impact, and the dangers of inferring a causational relationship. This may explain why, despite its importance, much less attention has been given to school belonging compared to academic attainment.

Research indicates that many children feel unsafe in schools (Cohen and Geier 2010) but conversely, for some young people school can be a safe haven (Allen and Kern 2017). Allen and Kern 20177) quantitative meta-analysis places school safety as having a moderate effect on feelings of school belonging, and linked with a number of positive outcomes. Feeling safe is related to academic achievement, mental and physical well-being (Lacoe 2020; Lenzi et al. 2017). Research indicates that some young people are more vulnerable to not feeling safe than others. Previous analysis of quantitative data in our study McDermott et al 2020) found that responses to items on feeling safe in lessons, and feeling safe during break and lunch times, exerted significantly greater weight in predicting how connected students with special needs and disability (SEND) feel than it did for students who did not disclose SEND. Studies of other minority groups also suggest lower levels of feeling safe and lower levels of school connectedness tied to absenteeism (Pampati et al. 2020). Pampati at al.’s (2020) study of transgender students finds that they did not experience positive teacher support, believing that staff did not care about them or treat them fairly. Notably the focus of many of the studies are American disadvantaged youth where levels of feeling safe are lower (Lacoe 2020).

One consequence of the reliance on quantitative measures or scales with little use of open questions, is that student voice is under-represented (Nind, Boorman, and Clarke 2012; Craggs and Kelly 2018a). Craggs and Kelly (2018b) provide a meta-synthesis of a relatively low number (8) of small-scale qualitative studies. Four main concepts are generated through their meta-synthesis: positive interactions or intersubjectivity; individual identity and feeling accepted; experiences of being part of a group; safety and security with an over-arching concept of “feeling safe to be yourself in and through relationships with others in the school setting” (p.9). Safety and security are largely constructed within the context of relationships, particularly those with their peers. This contrasts to quantitative studies where safety is typically placed within a context of school physical environment and linked with rules and discipline leading to a narrow conceptualisation of safety (Thapa et al. 2013; Lacoe 2020). An approach that includes the collection of qualitative data enables one to explore its meanings for young people.

The focus of an earlier analysis McDermott et al 2020) concerned students’ responses to closed questions, indicating that our proposed questionnaire had good levels of internal consistency across four subscales: “Emotional Security and Comfort (Belonging), being in the wider School Environment, the Academic Environment, and the Social Environment”. Having previously statistically analysed the questionnaire with respect to construct validity and internal reliability on the quantitative measures, here we privilege the qualitative data and investigate elements of validity and bias and examine what we learn through collecting learner views. The analysis is driven by the following research questions:

1. Is there bias in who responds to the open questions?

2. What can we learn from the open comments that provide insight into the barriers and supports to positive feelings of school belonging and what are the implications for school practices?

3. What is the relationship between students’ measures of connectedness and their comments?

4. How do students feel the questionnaire can be improved?

Development of the questionnaire

Full details of the development of the questionnaire are reported elsewhere, but in summary through an iterative process of piloting and refining in collaboration with volunteer schools, we collected data using a 5-point scale to explore how students felt at different times and places and during different types of lessons (reported in (McDermott et al 2020). The Emotional Security and Comfort (Belonging Scale) was comprised of 11 items, five of these were from the original Goodenow (1993) scale and a further six that reflected more recent literature on connectedness and addressed issues of safety, security, and being oneself, previously trialled (Daniels et al 2019). The items addressed personal acceptance (people being friendly and being able to be oneself), respect (from students and staff), sense of belonging (being part and proud), but also safety/security (lessons and non-lessons). Of central relevance here were the open questions inviting students to tell us a bit more: What things (if any) help you feel a part of the school; What things (if any) prevent you from feeling a part of the school; Can you tell us a bit more, following a question concerning feelings of safety in classroom and break and lunch times. These open questions were added following two pilots, and discussion with the school of the findings.

The questionnaire also asked participants, whether they had a disability or difficulty (with a number of examples) which had gone on for a year or more; and who they turned to for sources of support referred to later as SEND status.

Methodology

Following University ethical approval, we made a presentation to a University-Local Authority liaison group explaining the project and invited secondary schools to take part in this collaborative project, in which we would share the data with the schools and meet later to discuss the findings. We suggested that they focus on pupils aged 13/14 (often seen as particularly turbulent times with the highest rates of fixed and permanent exclusions). Schools chose whether they administered the online questionnaire to the entire year group through their own dedicated url. Three volunteer schools are represented in this data set and the questionnaire was completed by 595 students, who were assured both of their anonymity and that they could omit any questions they didn’t want to answer, and hence their right to silence (Lewis 2010).

To provide some further context, the three schools had very similar mean levels of feeling part of the school and feelings of safety (see Table 1) both of which averaged above the mid-point of a 5-point scale; 5 = completely true, they feel part of the school/safe, and 1 = not at all true. Student responses with respect to attendance in all schools spanned the range 6 = never take time off school to 1 = yes, all the time.

Table 1. Responses to quantitative questions by school.

Closed Question	Dale	Kings	Xenon
Q19 I feel like a real part of the school	N = 151 Mean 3.2 SD 0.94 Range = 1–5	N = 195 Mean = 3.37 SD 1.1 Range = 1–5	N = 230 Mean = 3.22 SD 1.04 Range = 1–5
Q27 I feel safe in this school during lessons	N = 151 Mean = 3.94 SD 0.94 Range = 1–5	N = 192 Mean = 4.14 SD 0.97 Range = 1–5	N = 230 Mean = 3.92 SD 0.97 Range = 1–5
Q28 I feel safe in this school at break and lunch times	N = 150 Mean = 3.91 SD 0.96 Range = 1–5	N = 200 Mean = 4.14 SD = 0.94 Range = 1–5	N = 233 Mean = 3.91 SD 0.91 Range = 1–5

Table 2 sets out the response rates to the five open questions that are the focus of this paper, revealing that a lower percentage of students from Kings school answered each of the open questions.

Table 2. Responses rates to open questions.

Open Question		Dale	Kings	Xenon
N		153	203	239
Q20 What helps you feel part of the school	Number of respondents	93	103	163
	Proportion	61%	51%	68%
Q21 What stops you feeling part of the school	Number of respondents	61	71	132
	Proportion	40%	35%	55%
Q29 Can you tell us a bit more about safety	Number of respondents	57	72	153
	Proportion	37%	35%	64%
Q42 Do you have any ideas for making the survey better? Perhaps questions we should ask?	Number of respondents	99	127	205
	Proportion	65%	63%	86%

Data analysis

Analysis of the qualitative data was informed by Craggs and Kelly’s (2018b) framework and that of St-Armand, Girard, and Smith (2017) adopting an iterative approach to the analysis of content (Schreier 2013) resulting in a hybrid strategy, with both deductive and inductive elements, whereby the framework was applied and the category of “other” revisited to identify emerging new themes and the data recoded. In essence this led to some differences between the coding of responses to “part of the school” and “safety” open questions. The question concerning “missing school” and “Ideas for survey” were coded through an analysis of emergent codes. Consistency was examined through calculating the degree of agreement between two independent coders of between 20% and 27% of responses with agreement levels of 82–95% and an average agreement level of 88%.

Statistical analysis was carried out using SPSS 25. Chi-square (χ²) tests were conducted in order to identify group differences in response rates for qualitative question, in particular examining the potential influence of salient demographic variables (such as gender or SEND status) on student’s observed rate of response (McHugh 2013).

The point-biserial correlation coefficient (r_pb), a special case of the Pearson (product-moment) correlation involving one continuous variable (X) and a dichotomous variable Y (r_XY = r_pb), was employed in estimating the magnitude of association between coded qualitative responses and associated Likert scale responses (Chen and Popovich 2002; Bruning and Kintz 1997).

Results

1. Differences in Response Rates to Open Questions.

Table 3 sets out our analysis of whether students exhibiting different demographic traits demonstrated significantly differing likelihood of responding to the open questions. Notably students’ likelihood of responding to any of the three open questions did not differ to a significant extent based on gender or self-reported SEND status. Conversely, the likelihood of student response was found to be a significantly associated with school, with fewer students than expected from Kings school responding to the open questions and more than expected responding from Xenon, a pattern that held for all three questions.

Table 3. Likelihood of responding to open questions by group.

Question	Group	χ^2	df
What things (if any) help you feel part of the school?	Gender	1.975	2
	SEND	.156	1
	Attendance	5.842*	1
	School	13.966***	2
What things (if any) prevent you from feeling part of the school?	Gender	2.412	2
	SEND	1.513	1
	Attendance	6.659**	1
	School	19.857***	2
Can you tell us a bit more [following questions about feeling safe	Gender	1.235	2
	SEND	.058	1
	Attendance	.357	1
	School	36.254***	2

*** p < .001

** p < .01

*p < .05

It is with this recognition of particular sampling disparities that we turn to examine what students said and the relationship to other parts of the questionnaire.

2. Responses to: what things (if any) help you feel part of the school?

For the open question regarding the elements and aspects which helped students feel part of the school, responses (see Table 4) were most commonly identified as referencing extra-curricular activities “Events like concerts, competitions etc where I can take part” (n = 128, 21.5% of students); social, friendship or activity groups “when people invite me into their friendships when I look like I am lonely” (n = 92, 15.5% of students); and feeling supported “probably the way everyone is like, kind, helpful and caring and so on”, (n = 79, 13.3% of students). Underpinning these responses were indications of their feelings: “when we get congratulated or recognized for something”, or “helping out” or “getting an actual part in a play, or “getting points for your team”. These are indicative of feelings of worth, of making a contribution, and being recognised for it.

Table 4. Frequency of coded responses: “What things (if any) help you feel part of the school?”and their relationship to measures of Feeling Part of the School.

Codes	Frequency of Coded Responses	Percentage of Students (%)	Relationship to ratings Q19 I feel like a real part of the school
Extracurricular	128	21.5	.124**
Social, Friendship, Activity	92	15.5	.115**
Feeling Supported	79	13.3	.124**
Sharing Symbolic Resources	49	8.2	.064
Curriculum	46	7.7	−.011
Other	31	5.2	−.080
Nothing	21	3.5	−.323***
Individual Identity	20	3.4	−.047
	359	60.3

*** p < .001

** p < .01

Several response codes concerning characteristics or elements which helped students feel part of the school were identified as positively correlated with the closed Likert Scale question indicating the extent to which students feel like a part of the school (see Table 4), including feeling supported (r= .124, p = .003), extracurricular activities (r= .124, p = .003) and social, friendship, and activity groups (r= .115, p = .006). Whilst weak, these correlations indicate that students whose responses to this open question were coded in these categories had typically indicated a greater feeling of being a part of the school than those whose responses were not coded in these categories (including those who did not respond). Notably, students whose responses were coded as indicating that they did not identify any aspects or elements of school as helping them feel part of the school (i.e. responded “nothing”) tended to demonstrate less feelings of being part of the school (i.e. scored lower in the closed question; r = −.323, p < .001) than those students whose responses were not coded in this category (including those who did not respond).

3. Responses to: what things (if any) prevent you from feeling part of the school?

Regarding the elements and aspects which prevented students from feeling part of the school (see Table 5), responses were most commonly identified as referring to (not) feeling supported (n = 85, 14.3% of students), (not taking part in) social, friendship or activity groups (n = 45, 7.6% of students).

Table 5. Frequency of Coded Responses: “What things (if any) prevent you feeling part of the school? And their relationship to measures of Feeling Part of the School.

	Frequency of Coded responses	Valid Percent of Students (%)	Relationship to measures in Q19 I feel like a real part of the school
Feeling Supported	85	14.3	−.069
Social, Friendship, Activity	45	7.6	.027
Other	45	7.6	−.140***
Curriculum	35	5.9	−.132***
Nothing	28	4.7	.060
Extracurricular	27	4.5	−.039
Individual Identity	27	4.5	−.125**
Sharing Symbolic Resources	5	.8	−.023
Total	262	44.0

*** p < .001

** p < .01

Feeling like teachers don’t like you or feeling targeted and alone (feeling supported and respected)

Not taking part in things that help the school

Not being involved

Being excluded from group’s. having arguments with friends.

In many cases responses revealed interrelated themes

constantly feel like teacher are delibratly picking on me and putting me down to make me feel worthless and putting themselves out . to make me feel dumd stupid and inscure about myself and my learning only a very small amount of teachers make me feel good about myself and succeed to help me identify what I am good at while others just tell me I don’t care about anything I do its just them delibeatly picking on me and I hate this school and the fact I am here (individual identity-feeling supported and respected)

knowing that there are always going to be people judging you if you do something or act differently, that thought is always at the back of my mind”. (identity, feeling supported and respected)

A number of responses fell into the “other” category (n = 45, 7.6% of students) but a number of these (24/45) indicated that students “Did not know” or were unsure or responded with question marks “?????”.

Several response categories regarding characteristics or elements which prevent students from feeling part of the school, were identified as negatively correlated with the closed question indicating the extent to which students feel like a part of the school (see Table 5) including individual identity (r = −.125, p = .003), and curriculum (r= .132, p = .001): when i get down because i can’t do the work. It makes me feel useless and that i am not part of the school. Although weak, these correlations indicate that students whose responses to the open question were coded in these categories had typically signified a lower sense of feeling part of the school (i.e. in ratings on question 19) than those whose responses were not coded in these categories (including those who did not respond). These results may be interpreted as demonstrating that negative experiences related to individual identity, curriculum, and feeling supported are associated with a diminished sense of feeling a part of, and connected to, the school.

4. Responses to: Can you tell us a bit more [following questions concerning safety in classroom and at break and lunchtimes.]

Responses concerning students’ sense of safety during lessons, and at break and lunchtime, (see Table 6) were most commonly identified as referring to relationships with students (n = 71, 9.7% of students):

Table 6. Frequency of Codes Can you tell us more about safety and their relationship to measures of Feeling Safe.

Codes	Frequency of Codes	Valid Percent of Students (%)	Relationship to Ratings in Q27 I feel safe in this school during lessons	Relationship to Ratings in Q28 I feel safe in this school at break and lunchtimes
Relationships – Students	71	9.7	−.085*	−.080
Other	52	8.7	−.024	.020
Relationships – Teachers	46	6.3	−.025	.007
No	46	7.7	−.090*	−.053
Space Systems	39	5.3	−.027	−.124*
Individual Identity	31	5.2	−.048	−.059
Curriculum	26	3.6	−.076	−.030
Total Answered	239	40.2

* p < .05

I feel its difficult for me to be my self in such big groups of people in lessons, prefer to sit next to a friend so I feel more combatable and can get on with my work without worrying …

people in my class make fun of me and some of my friends becuz the way we look, size and make fun of our personality and it stops some people from being them selves in class and being too nervous to take part.

I don’t really feel like I can be open and be myself.

The issue of space was raised in some of the responses, particularly with respect to lunch:

there is a lack or places to go at lunch

the canteen is wayyyyy to small to feel comfortable

people running around barging into everyone and making people drop there lunch

Reasons “other” than those categorised by identified codes (n = 52, 8.7% of students), included 46 (7.7%) students who explicitly declined to elaborate on their sense of safety (i.e. stated no, or IDK).

Several response codes (see Table 6) were identified as negatively correlated with student responses to the related rating questions (question 27 and 28). Students who discussed student relationships tended to indicate a lower sense of security during lessons (r = −.085; p = .041) than those whose response was not coded in this category (including those who did not respond). Interestingly, responses explicitly refusing to – or expressing no desire to elaborate on feelings of safety (i.e. responding “no”) were also negatively associated with student feelings of safety during lessons, such that students declining to elaborate tended to indicate a lower sense of security during lessons (r = −.090, p = .030). Students whose responses referred to space systems typically demonstrated a lower sense of security during break and lunchtime (r = −.124, p = −.027) than those students whose response was not coded in this category (including those who did not respond).

5. Responses to: Do you have any ideas for making this survey better? Perhaps questions we should ask?

Over a third of students responded to being asked for suggestions for the survey, although the largest category (see Table 7) contained variants of the response “no” (I don’t know, not sure, no idea etc). However almost 10% of students made suggestions for questions that we should ask. Interestingly the largest group of questions concerned how people felt (N = 19) including reference to their mental-welling being: e.g. this survey could say; “Do you feel like you are in the right place” or “Do you feel like you can express yourself at school”. Another predominant theme was to suggest we ask questions about their teachers (N = 15) “do u have good relationship with the teachers?” or more directly “who is the worst teacher”; and similarly questions about lessons (N = 14) “You could ask what lessons are better than others and why so you can improve the ones that aren’t the most popular”. Seven students suggest we ask directly about bullying and a further five about relationships with peers and their ability to make friends.

Table 7. Frequency of responses to question asking for survey suggestions.

	Frequency	Valid Percent (%)
Improved Security	20	3.4
Formatting	23	3.9
Clarity	16	2.7
Suggested Questions	58	9.8
No/IDK	93	15.6
Total Answered	216	36.3%

While many of these respondents suggested questions that were quite personal, a reoccurring theme from those whose comments concerned security (n = 20 3.4% of students) was a desire to ensure anonymity and not to give any personal information: “I would prefer to put no details, more confidential”. A theme that ran through comments about clarity and formatting was the challenge of open questions and being quite clear about what was being asked. There was also a desire for a shorter survey or at least “shorter questions” and “more time to answer questions”, and “Making the questions more specific and less of the ‘Can You Say More’ please”, suggesting the need for fewer open questions.

Discussion

With reference to our first research question, no group appeared to be under-represented in responding to the open questions. This is particularly important given that open questions typically have much lower response rates (Denscombe 2008). There was no significant gender (including transgender) differences in the likelihood of responding, nor with respect to whether a student disclosed a special need or disability. School appeared to make a difference in the likelihood of responding, suggesting that there could be differences between schools in students’ experience of being listened to, but this needs to be set alongside the higher mean scores at Kings for feeling part of the school and safe. Their students may simply have had less to say.

Our checks of validity (Research question 3) through examining the open question responses against the closed question measures revealed a large degree of consistency. For example, when we looked at the most frequent codes for what makes you feel part of the school, these were positively correlated with students’ ratings of the extent to which they felt part of the school, and conversely those who responded “nothing” helped them feel part of the school gave lower ratings for their feelings. There was similar consistency with respect to what prevents you feeling part of the school as in this instance the top responses were negatively associated with their ratings. Similarly, with the issue of safety, students who responded with respect to issues of space had a lower measure of safety at break and lunch times. These findings give some confidence in the veracity of student responses. We now turn to examine what we learn from the open comments that provides insight into the barriers and supports to positive feelings of school belonging (RQ3).

Feeling part of the school

Our analysis of students responses to being asked “What helps you feel part of the school”, indicates that positive experiences of feeling supported, extra-curricular activities and students’ social, friendship and activity groups are elements of the school most closely associated with students’ sense of feeling a part of the school community. This is largely consistent with other quantitative studies (Allen and Kern 2017) including studies of minority groups who emphasise importance of teacher support (Pampati et al. 2020). The qualitative data here provided additional insights on the importance of both making a contribution, whether it be fund raising, scoring points for the team, or the success of a communal activity AND being recognised for this. Equally striking, however, was the significant association in our data between answering “nothing” would help and students’ lower scores for feeling part of the school, suggesting that these students felt no sense of optimism or agency in bringing about change.

Responses when asked what prevented students from feeling part of the school were highly consistent with these aspects and referenced an absence of feeling supported, and social, friendship or activity groups. Our analysis of the relationship between how students responded and the extent to which they felt part of the school indicated that the negative experiences relating to individual identity, curriculum and feeling supported are associated with a diminished sense of feeling part of a school. Students, for example, commented on not being able to do the work, often with reference to particular subjects, and the impact this had on their self-esteem.

Taken together the responses to these two open questions indicate the importance for young people of not being defined by their performance in class. Shochet and Smith (2014) argue for the importance of the classroom environment as a mediating effect on mental health and we would add to this the negative impact of school hierarchies of ability, of being made to feel “dumb” and on students sense of value as learners (Brown 2017) and their wider identity. Extra-curricular activities provide opportunities to develop common interests with others and a sense of agency in making a contribution.

Feeling safe

Turning to feelings of safety, relationships with teachers and students featured most commonly in the responses. Students who discussed relationships with their peers tended to have a lower sense of safety in lessons and so too did students who declined to elaborate. It may be that conditions which cause students to feel somewhat unsafe during lessons prompt reticence and reluctance in specifying these conditions (e.g. should students feel unsafe due to issues involving teachers or other students they may be wary of disclosing such details). This would explain why this proved to be a harder question to respond to as denoted by fewer responses, and by the response “no” (they couldn’t say more). However, students wrote, often at length, about their difficulty with being themselves, feeling that they would be ridiculed for the way they looked or for being unable to do things. It was not always clear who made them feel this way e.g. I usually always feel on edge in case i accidentally say anything slightly wrong i will get shouted at and i just don’t know what to do; but in many responses it was clear that friends provided a buffer. A few students also mentioned feeling threatened, bullying and arguments between students. Consistent with this were comments about teachers, whether they were fair and their ability to control the class.

Students who referred to space systems typically demonstrated a lower sense of security during break and lunchtimes suggesting that for these students it was unsupervised times of the day that made them feel less safe. This is consistent with the review of Thapa et al. (2013) but here the responses of students suggested that it was due to being crowded. Small school spaces provide a number of pinch points that impact negatively on student interaction. Other studies suggest that spaces for lunch are particularly important in relation to feelings of connectedness (Davison, Ghali, and Hawe 2011), including simply having somewhere to go, either with friends or for sanctuary (Brown 2017) Responses to this open question revealed the importance of not simply seeing safety as an issue of discipline and behaviour management, or having a strong perimeter fence. For many students their responses were about being able to manage their work, their concern about being picked on, of not knowing the answer. The issue then shifts from how the teacher manages behaviour, to how they manage the learning environment when students are struggling.

Student feedback on the survey

We now turn to our last research question. Here were important reminders that students complete surveys under different conditions. Some students felt they did not have sufficient time suggesting that schools had not allocated enough time. Some said the survey was too long, and others that they would rather have closed questions. Student responses also suggested that the phrasing of open questions were too vague and ill defined: Making the questions more specific and less of the “Can You Say More” please. We had chosen not to be directive in our wording, which can imply a desired answer, but this may be uninspiring, and the lack of feedback that is available in face-to-face contexts present additional difficulties. The response rate to being asked about safety was particularly low but the answers revealed how safety was understood and experienced in many different ways, indicating that the phrasing needs to allow for all viewpoints. In future iterations of the questionnaire, the wording and frequency of open questions will be revisited. Students suggestions for additional questions often made reference to “feelings” which could provide a useful framing for open questions.

Although student responses suggested they were happy to talk about themselves, they had concerns about being identified. For ethical reasons, we gave students the opportunity to talk with someone of their choice if the questionnaire had raised issues for them, but this required them to give their name. A better option would have been for schools to make appointment times available with a range of adults for students to self-select. One of the advantages of online surveys is that students will give candid responses but if their sense of anonymity is compromised so too is the relative benefit of adopting this approach.

Significance and implications-

Our overall aim is to develop a robust tool that schools can use to make systemic responses that foster students’ feelings of connectedness to schools. Our previous analysis of quantitative data indicated that the questionnaire was a reliable tool, sensitive to differences in feelings of belonging (Author et al 2020). To this we can add that not only did the consistency of the open questions with the quantitative measures add further checks of validity, but they also contributed particular insights for schools that give further meaning to the measures. Through focusing on student views attention is directed towards particular systemic aspects of provision that serve as barriers to school belonging. Providing schools with a mixed methods tool enables them to understand what would improve school belonging but also to measure how effective their changes have been.

A sense of belonging is not a simple set of cause and effect associations. Rather these qualitative data have highlighted the inter-relatedness between how learners feel about themselves, access to activities they can choose to engage in (extra-curricular), and their relationship with others and, particularly importantly the support that friends provide. This indicates the need for a multi-strategy approach that may well require a culture shift within the school, rather than a single targeted intervention.

Student comments also revealed the complexity of feeling safe. While government and school policy may place the emphasis on behaviour management systems, a student perspective lies with positive and valued interactions with others. This argues for a shift away from improving behaviour, to policies and practices that concern improving relationships, of positive interactions, of being accepted and valued. As Brown argues (2017, 412) “to belong socially is to belong spatially”, so safety also relates to the allocation of space, within classes, including seating plans, having a space within school to gather with friends or simply be on ones own, and to movement around those spaces. The deterioration of the school estate (NAO 2017) coupled with the introduction of a “generic” one size fits all school design policy has resulted in crowded schools that provide a negative context for learning.

Positive interactions that lead to students feeling supported require staff to be sensitive to the emotional state of the young people (Greenwood and Kelly 2019). This includes being aware of students external lives including their experience of bereavement and loss of interaction with key people in their lives. Not all staff feel naturally confident in responding to strong emotions, suggesting that now more than ever students will benefit from having access to a key adult in the school who is capable of supporting those experiencing strong emotions. Rowe and Stewart (2011) found in their case study schools that valued staff combined the characteristics of being “understanding”, having “respect”, “listening to students”, “caring”, “supporting”, “interacting in a positive way”, showing “genuineness”, and “commitment” (p.407). Staff similarly need also to experience these characteristics in their interactions with others so that they are features of the whole school culture and reflected in school procedures and routines.

It is unclear whether school staff see their role to involve promoting positive peer relationships (Craggs and Kelly 2018a), and yet they are integral to students feeling safe in the classroom. Having trusted friends provides a buffer both in relation to feeling excluded (Hamm and Faircloth 2005) and to physical threats (Einberg, Lidell, and Clausson 2015). Staff may under-estimate the importance of this (Aldridge and McChesney 2018). From our data we can add the issue of safety in lessons. Anxiety concerning being asked a difficult question or being ridiculed for making an incorrect response can contribute to a sense of insecurity within school. This places particular responsibility on teachers to foster through their own responses to students, a positive style of pedagogic interaction.

Limitations

There are some limitations to the study that must be acknowledged. While the number of returns are adequate for testing out the robustness of the instrument, we must be cautious about generalising the findings to other age groups and the wider community of schools. Given that the aim is that this is a tool for schools to use, the coding of open questions is not without difficulty. This leads to some inference of meaning which we sought to offset through examining consistency between coders. For example, when people respond “no”, to the request for feedback on the survey they could mean that they won’t or that they can’t or that “No, I think it was a good survey”. There was less ambiguity where students responded “nothing” in response to what helps you feel part of the school a reminder that such should not be dismissed as a flippant or apathetic response. The explicit action of responding “nothing” has, in the context of the open-ended questions included here, been demonstrated as representing a clear and intentional response on the part of students, and consequently considered with care.

We are also mindful of the danger of assuming that frequency of responses is indicative of relative importance for participants, a limitation that could be overcome by schools using additional methods such as nominal group technique (Porter 2013). In many instances the comments themselves were indicative of the strength of feeling. However, as we have tried to portray here, a single word “nothing” can be equally meaningful.

Conclusion

The inclusion of open questions on surveys of connectedness are rare and yet they have the potential to provide some insights from a whole cohort or class of students, giving an anonymous space to students who have something to say, whilst also upholding the right to silence. The collection of qualitative data alongside quantitative provides an avenue for accessing the views of students who, on self-rating measures feel quite differently. Our analysis indicates that with a slight reframing of the open questions, this tool can provide schools with both a robust measure of school belonging and valuable insights to explain these.

The qualitative responses revealed the central role of relationships, both their propensity to enable students to feel part of the school but also their threat to individual identity and feeling accepted. In this sense there was a strong inter-relationship between categories of response. Extra-curricular activities made students feel part of the school, and provided a social activity in which they felt valued. Friends in turn were a buffer when there was a danger of being ridiculed or otherwise threatened. Feelings of safety proved to be multi-faceted, and highlighted the threat to self-esteem in lessons. There are particular implications for staff and for the development of systemic responses to foster positive interactions, both as a pedagogic style and in non-academic contexts. Feeling part of the school is firmly associated with a range of outcomes that are linked to student well-being. It is particularly timely, given the impact of the pandemic on access to education, for schools to think about building on their sense of community.

Disclosure of potential conflicts of interest

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the John Fell Fund, University of Oxford.

Notes on contributors

J. Porter

Professor Jill Porter is Honorary Research Fellow, Oxford University Department of Education.

T. McDermott

Tiarnach McDermott is DPhil Student at Oxford University Department of Education.

H. Daniels

Professor Harry Daniels is a member of the Department of Education, University of Oxford, a Fellow of Green Templeton College and Co-PI of the ESRC funded large grant “Excluded Lives: the Political Economies of School Exclusion.

J. Ingram

Dr Jenni Ingram is Associate Professor of Mathematics Education, Fellow of Linacre College, Oxford University Department of Education.