In control or at the mercy of others? Navigating power dynamics in online data collection with UK secondary school students

ABSTRACT

Research involving young people is a challenging process that requires managing relationships with diverse individuals and groups, including the young participants and their various gatekeepers. While it is normally assumed that the researcher is in overall control of their research, by using a Foucauldian conception of ‘power as effects’ that operate in the form of relations and through discourse as the articulation of norms, this paper discusses how, in practice, the researcher can lose control over their research and so be forced into making substantial compromises concerning the nature and extent of the data they can collect. I do this by reflecting on my experience of conducting research involving UK secondary school students using online data collection methods during the Covid-19 pandemic. I identify several factors that generated power effects which influenced the conduct of the research, including: an ethics review that relied on a simplistic discourse concerning young participants’ (in)competence; my own self-regulation of my conduct in respect of ‘ethical’ research; my ‘positionality’ in the field; and a researcher's general dependence on participants and gatekeepers to complete their research. I conclude by reflecting on how these factors may impact upon the conditions for viable social research involving young people.

KEYWORDS:

• Young people
• gatekeepers
• research ethics
• positionality
• power as effects

Introduction

Research involving young people, and especially those under the age of informed consent, can be challenging as it entails negotiating the complex power dynamics which can arise when dealing with diverse groups of ‘gatekeepers’ – including university research ethics committees, parents and school teachers – as well as the young participants themselves. Likewise, research involving young people is a site where strong assumptions and discourses mingle with and compete against each other to influence how such research is conducted and evaluated. In particular, it is frequently assumed that, as adult researchers exercise control over the research process (Farrell 2005, McGarry 2016, Lohmeyer 2020), their young participants are subject to their power which, in turn, can lead to various forms of regulation being placed on the conduct of the research at the ethics review stage. Yet this assumption is often held alongside an increasing commitment to the competence and agency of young people which would suggest that they are unlikely to be the passive tools of researchers.

In contrast to the first assumption above, this paper will discuss what can happen when an adult researcher is not always in full control of their research and how, in some situations, they lack any real control at all. I do this by drawing on my recent experience of conducting online and remote research involving UK secondary school students during the Covid-19 pandemic. Using a Foucauldian conception of ‘power as effects’, I argue that several factors can combine and interact to weaken a researcher's control over their research, leading to substantial compromises in the nature and extent of the data they are able to collect. These factors include: the assumptions and discourses that often underly the ethics reviews of such research and the researcher's ethical self-regulation of their own conduct; the ‘positionality’ of the researcher in relation to those whom they encounter in the field; and the wider issue of the extent to which any researcher is dependent upon their gatekeepers and research participants for their project to be completed successfully. In my study, the power effects arising from these factors were further exacerbated by the requirement to use online and remote data collection methods due to the pandemic and the chosen research approach.

Drawing on the work of Foucault, I begin by discussing how I conceptualized the forms of power that I experienced during my research. I then discuss the key relationships and factors that are often involved in generating power effects in research involving young people, how these appeared in my study and how they acted to constrain my ability to conduct it in the manner that I had intended. I conclude by discussing the wider implications of the Foucauldian approach and my own experience for social research with young people. Above all I suggest that, if it is to remain viable, the ethics regulation of research with young people needs to take proper account of the power dynamics and effects involved in such research.

Power as effects in social research

While power has been conceptualized in various ways (e.g. Lukes 2005), most conventional approaches see it as an individual's or a group's ability to intentionally exercise control over what others do (Spencer and Doull 2015). For example, Runciman (1999) suggests that power always involves the ability of one party to exercise forms of economic, ideological or coercive control over another. These conceptualisations of power are often associated with the idea that society is divided into groups by structural factors such as class, gender, race/ethnicity and disability that contribute to inequalities and domination. Power relations are also seen to arise in individual consensual interactions where one party possesses more resources or expertise than another (e.g. between a doctor and their patient). However, and despite these variations, the unifying characteristic of power so conceptualised is that power is inherently unequal and zero-sum: the more power one group has, the other will have less.

The above view of power is often held to characterize the relationships between adults and young people. In such a view, age differences are seen as a key factor in producing such power imbalances (Lee 2001) with young people typically considered to be both physically and psychologically immature compared to adults (Woodhead 2011). In social research, this age-related power inequality is often assumed to apply to the relationship between the adult researcher and their young participants, which can be further exacerbated by the prestige or symbolic status the adult ‘researcher’ may bring into the relationship. Given these assumptions, and especially where it involves vulnerable groups of young people or certain topics, it is considered necessary that any such research is subject to close ethics regulation to prevent the intended or inadvertent exploitation of the researched by the adult researcher.

While the prevalent assumption in the field of research involving young people is as described above – i.e. adult researchers are in a position of power over the young participants (McGarry 2016, Aaltonen 2017, Davidson 2017) – there is also an increasing emphasis on young people's competence and capacity for agency. A further related assumption that the exercise of agency by young people is both beneficial in general and, more specifically, in terms of the outcomes of research has also meant that participatory research methods have been promoted with the explicit aim of redressing any adult–child power imbalances and to ‘empower’ the latter (Alderson 2001, Burke 2005, Kellett 2010). Indeed, some researchers use whether such a method and/or approach was adopted as a significant criterion to evaluate individual studies. Yet participatory research methods and approaches have themselves been criticized for not necessarily better reflecting the perspectives of young people as opposed to those of adult researchers, and for not always producing more authentic and valid knowledge (Gallacher and Gallagher 2008, Hunleth 2011, Spyrou 2011, Tisdall and Punch 2012).

In contrast with the above view, Foucault (1975[1991], 1978[1990]) suggests that power is distributed across society and is diffused in and through knowledge and discourse. Crucially, under this approach, power is not treated as a disposition or possession (Foucault 2003) and does not always derive from the decision or intention of an individual subject (Foucault 1978[1990]). Instead, power operates in the form of relations and through discourse as the articulation of norms (Brown 2008). Thus, any analysis of power so conceived should focus on its field of application where it produces effects rather than on its source (Foucault 1980). Foucault also conceptualizes power relations as lying on a continuum between ‘strategic games between liberties’ and ‘the states of domination’. Furthermore, while it is unrealistic to abolish such power relations (as no relationship can be free of power), one should always seek to engage in them with as little domination as possible (Foucault 1997 cited in Simons 2013, p. 316).

In social research this notion of ‘power as effects’ is particularly pertinent in contexts where access to data is dependent on the relationships between the researcher and those whom they encounter in the process. In research, as in other social fields, diverse discourses, inextricably associated with the particular social and historical conditions of the context, can also compete with one another, with some being more dominant and influential than others. Individuals who subscribe to, or are under the influence of, a particular discourse (or discourses) may therefore respond to the actions of the researcher in ways that can facilitate or hinder the latter's ability to conduct their research. Most importantly, in contexts where power is dispersed across a range of research relations and throughout the research process, the researcher may not have as much control over their research or its participants as is assumed by the first view of power and may even be manipulated or exploited, especially when they are not positioned as part of a dominant social group (Thapar-Björkert and Henry 2004).

Yet, while researchers themselves also participate in discourses, their own reflexivity, or what Foucault (1988) refers to as a ‘technology of the self’, can be an instrument through which they become aware of a set of truths by which to judge their own behaviours and those of others (McCabe and Holmes 2009). Through such reflexivity, as McCabe and Holmes (2009) suggest, researchers can become aware of how their own social positions and views can influence their research and its outcomes. Indeed, it was through such reflexivity that I was able to identify the discourses involved in my own study, my subjectivity in respect of ‘ethical’ research, my unique positionality in relation to my young participants and their gatekeepers, and how all of these may have generated power effects that determined the nature and scope of the data I was (or not) able to collect.

As will become clear, my own recent experience of conducting research with young people has led me to view the ‘power as effects’ approach discussed above as the one which more adequately describes the power dynamics I experienced with the young participants and their gatekeepers, including the research ethics committee and their teachers. However, and before considering my own experiences, I first need to discuss in more detail the key relationships and factors that can generate such ‘power as effects’ in research involving young people.

Power as effects in research involving young people

Research ethics regulation and young people’s ‘competence’

As noted above, ethics regulation is essential for balancing the effects of power that can harm any individuals involved in research, including both the participants and the researcher. However, research ethics reviews are a relatively recent innovation and can take various forms depending on the institution involved. In particular, UK university research ethics regulation has varying procedures reflecting different institutional values and beliefs in the extent of the role of ethics committees that can also ‘vary from individual to individual within the ethics review process’ (Vadeboncoeur et al. 2016, p. 226). These differences can result in a tendency for ethical requirements to differ between institutions.

Yet, and despite this variation, there is also evidence that ethics reviews have become increasingly institutionalized and bureaucratized (Allen 2009, Sleeboom-Faulkner et al. 2017, Iphofen 2020) and often more geared towards the protection of universities, funding bodies and other organizations against the threat of litigation (Whitney 2023) than with resolving genuine ethical dilemmas. These developments may also have resulted in the more frequent adoption of an overly instrumental attitude towards research ethics, with some researchers regarding the ethics review stage as a ‘hurdle’ to surmount early in the research process (Reid et al. 2018) or as a procedural hindrance with some even seeing it as a ‘charade’ (Mcareavey and Muir 2011).

But despite these frustrations, the institutional discourse of ethical research that ethics reviews enact can also have ‘constitutive effects for researcher identities that are “ethical” and what in practice might represent “ethical” research’ (Allen 2009, p. 395). These effects, which resemble Foucault’s (2008) concept of ‘governmentality’, can lead to researchers’ self-monitoring and regulation of their own conduct in the field. This kind of self-government can thus affect researchers’ reflexive attention to ethical issues as autonomous beings to varying degrees on the continuum between strategic games of liberties and states of domination. Consequently and, in conjunction with their more overtly procedural requirements, ethics reviews can serve to both limit and shape the researcher's conduct of their research.

Given that young people's competence and agency are contested concepts (Kim 2023), ethics reviews for research involving young people can also be a complex site of competing discourses. In particular, there can be a major tension in the discourses that guide the ethics regulation of such research between, on the one hand, a commitment to protect the participants from harm and, on the other, to ensure that their ‘competence’ is respected and that they can exercise autonomy in relation to researchers. Yet, paradoxically, if young people are held to have the competence and capacity to exercise agency in such contexts, this would also suggest that they are unlikely to be merely the passive subjects of the power of the adult researcher and any other adults involved in the process (e.g. on-site gatekeepers).

However, and while it has been argued that ‘children who are mature enough to reflect on their experiences should be participants in consideration’ of ‘ethics policies and guidelines pertaining to social research with children’ (Grover 2004, p. 91), van den Hoonaard (2002) observes that research ethics committees tend to overstate the potential risks to participants, while discourses concerning ‘safeguarding’ and risk management remain dominant (James 2021). These discourses, in turn, can prevent researchers from using child- or youth-centred methodologies and, paradoxically, make them feel that they have acted ‘unethically’ by complying with the requirements of research ethics committees (Allen 2009, p. 395). Furthermore such ‘risk-averse’ ethics reviews can risk infantilising researchers by limiting their ability to manage some of the risks of a study themselves (Lohmeyer 2020, p. 52).

A related topic is young people's competence to give informed consent. Helseth and Slettebo (2004, p. 299) suggest it is generally agreed that children aged 7 and older ‘are at a developmental stage that allows researchers to include them in the process of informed consent’, although the research details must be provided in a way which they are able to understand. In the UK, the concept of ‘Gillick competence’ respects the autonomy of a person aged under 16 if they are considered to have ‘sufficient maturity and intelligence’ and suggests their consent cannot be overruled by a parent/guardian (Griffith 2016). Gillick competence is also ‘task specific’ where ‘the degree of maturity and intelligence needed’ of a child ‘depends on the gravity of the decision’ (Griffith 2016, p. 245). While Gillick competence originated from a legal judgement concerning the provision of contraception to sexually active underage girls without parental consent, decisions about participation in most school-based social research are usually of far less consequence. Furthermore, in school-based research, teachers acting as on-site gatekeepers will assess the value of the research and any associated risks before allowing the researcher access to their students, so providing further reassurance concerning any risks associated with seeking consent only from students.

However, in the UK, for research involving persons aged under 16, research ethics committees regularly stipulate that parental consent must be obtained and strict requirements are often laid down about how this is to be done. For instance, ‘opt-in’ parental consent (where the parent/guardian's explicit, usually written, consent must be obtained before their child can participate) is frequently insisted on, as opposed to ‘opt-out’ parental consent. Given the practical burdens this can entail, some institutions explicitly discourage research with young people ‘unless circumstances [are] so “exceptional” [as] to justify their inclusion’ (Sherwood and Parsons 2021, p. 450) and, where it does occur, such requirements can restrict the research design, participant recruitment processes and, ultimately, a study's findings.

Relations with young people in research

Likewise, the underlying assumption that an adult researcher is in overall control of research with young participants can be misleading given the general characteristics of researcher-participant relationship. One such characteristic is that, as the successful completion of a study is the researcher's responsibility alone, they are usually highly dependent upon the cooperation of participants who are unlikely to be highly committed to the research or, indeed, to be interested at all. This is especially so where the research does not reflect the participants’ own interests or concerns, is initiated from a theoretical agenda or where they have been encouraged to cooperate by other adults in positions of power (e.g. teachers or parents). The resulting precarity in their commitment to the study can be further exacerbated by a requirement frequently laid down by research ethics committees that researchers continually remind their participants that they can end their participation at any time and can decline to answer any questions without giving a reason for either (Miller and Bell 2012). While not suggesting either of these entitlements should be abolished, it is important to understand that such requirements can have consequences for the quality of the research produced, especially in a situation where participants’ level of commitment is already low.

Some research suggests young people can exercise power (seen in the first sense above as being an individual's ability or possession) in their interactions with the researcher, although this is sometimes interpreted as a manifestation of their overall powerlessness. For example, Tinson (2009, p. 29) suggests withholding information, changing the subject or even fabricating responses are strategies that adolescents use to deal with their lack of power. But it is also important to recognize, regardless of their motivation, that such strategies can also have a disempowering effect on the researcher. Davidson (2017) discussed how she successfully managed the participants’ exercise of power through her prolonged presence in the field in her study with young people from a disadvantaged housing estate in the UK. However, in many research sites, and particularly in school-based research, the researcher's presence in the field can be strictly time-limited, making such approaches more difficult.

Such problems are magnified when data is collected online and remotely, as was required for research conducted during the Covid-19 pandemic. While some research suggests that, particularly for focus groups, online data collection can elicit rich data (Keen, Lomeli-Rodriguez and Joffe 2022), building the necessary rapport often depends on a lengthy relationship between the researcher and individual participants (Weller 2017), which can be more difficult remotely compared to in-person interviews (Seitz 2016). And while exchanging emails with participants prior to an online interview can help develop rapport (Deakin and Wakefield 2013), doing so with young people can come into conflict with their school's safeguarding policies. Young people themselves may also feel uncomfortable about exchanging emails with a researcher whom they have not yet met even through a video call.

On-site gatekeepers and researcher positionality

Finally, the researcher's relations with on-site gatekeepers, such as teachers, may also involve power effects that can influence the research process. Here it is very important to understand that schools can receive numerous requests to carry out research in their institutions and that this can impose significant burdens on the staff and add to the already considerable pressures associated with administration, testing and inspection. Perhaps as a result, Crowhurst and Kennedy-Macfoy (2013, p. 457) suggest that unlike the way they are often described in social research textbooks as ‘monolithic, neutral and static figures’, on-site gatekeepers can impact on research in various ways: they can grant (or deny) access to the site, provide (or refuse) assistance in participant recruitment, and help (or hinder) the fieldwork. Likewise, they can also actively influence the design of the data collection tools and procedures and the sample composition (see Agbebiyi 2013 for an example of such an expanded influence).

The researcher's positionality relative to the on-site gatekeepers can also be crucial. ‘Positionality’ relates to how the identity of the researcher is constructed by those they meet in the field, based on characteristics that are the subject of their ‘interest, speculation and contestation’ (Thapar-Björkert and Henry 2004, p. 367). In practice, this means the identity of a researcher is constructed by gatekeepers from a range of characteristics, including the researcher's age, gender, race/ethnicity, class and, importantly, their institutional status: a professor or tenured academic may be able to command more authority than a research associate or research student. This can also mean that on-site gatekeepers’ ‘unplanned’ influence is more likely to occur when the academic authority of the researcher is perceived as relatively weak. Hence, in Agbebiyi’s (2013) study, the on-site gatekeepers’ involvement and influence, from the design of the data collection tools to the final sample composition, might not have happened to such an extent had the author not been a student researcher.

The power dynamics with on-site gatekeepers can be a sensitive issue to write about since, although they remain anonymous in publications, gatekeepers may still be recognizable to themselves and others; and, if they dislike the way they are described, this may affect their relations with both current and future researchers. For example, Campbell et al. (2006) were concerned about how their critical paper on the role of on-site gatekeepers might impact on their on-going relationships with their gatekeepers. While an open and reflective discussion with gatekeepers concerning their impact in the research process may help mitigate such risks, this is not always either feasible or effective.

My own study, which was conducted during the recent pandemic, clearly illustrates the various ways in which the factors discussed above – those relating to the role of research ethics committees, the autonomy of young participants and teachers as on-site gatekeepers – can generate ‘power as effects’ in the research process. While I do not claim that any of these factors are illegitimate, by both highlighting them and their consequences, my intention is to suggest that such power effects need to be given more attention, especially at the ethics review stage.

Navigating power dynamics in research with UK secondary school students during the pandemic

My study aimed to investigate whether and to what extent UK secondary school students held critical views about socioeconomic inequality and whether and how these views might relate to their sense of agency concerning their future occupational choices. The research was intended to help understand whether and how young people's views about inequality might contribute to the UK's low levels of social mobility (e.g. Elliot Major and Machin 2018) which has emerged as an issue of significant political and popular concern (Littler 2018).

As the study's focus was on the impact of young people's socioeconomic status on their views, and to help increase the possibility of recruiting enough voluntary participants, I planned to recruit at least two schools from areas with very different socioeconomic characteristics. I had also designed the research as a mixed-methods study: an initial questionnaire to collect data from a relatively large sample of students to analyze the patterns in the variables associated with the topic; and then, based on the findings from this data, to conduct a smaller number of semi-structured in-depth interviews (with those participants who were willing to participate in this second stage) to examine the underlying processes and mechanisms in more detail.

I conducted this research as an unpaid visiting researcher at the academic institution where I obtained ethics approval of the study. Yet while a ‘visiting’ researcher and originally from East Asia, I was not literally ‘visiting’ the university from another one in the UK or from one located abroad. Instead, after holding various fixed-term contracts at several British universities for about a decade since completing my PhD (also in the UK), and while having previously conducted school-based research in the country, I had been left with nowhere to continue my research career. At this point this university kindly gave me a position of an associate researcher to use its facilities (e.g. library, desk space) and this then enabled me to secure a small grant from a major funder to cover the direct costs of the research.

Power effects of the ethics review

As noted above, UK university research ethics systems have varying values and procedures, meaning that, depending on their institution's approach to ethics reviews, researchers can have different or inconsistent experiences when conducting their research (Lees, Walters and Godbold 2021). In my case, an issue arose at the ethics review stage concerning the required mode of parental consent. In a mixed-methods study, and especially at the questionnaire stage, it is important to obtain a suitable sample size. To achieve this, I had planned to use ‘opt-out’ parental consent to help maximize the number of participants. However, the ethics committee of the university where I was based viewed this proposal as rendering the research ‘high risk’. In response to my enquiry about what risks opt-out parental consent might create, the research ethics office responded that, where this had previously been used, the university had received parental complaints and, to further justify the requirement, they referred me to the concept of ‘Gillick competence’. However, and as was noted earlier, such a competence would require only having to obtain the consent of the participants (assuming they have sufficient maturity and intelligence) who were already secondary school students, not that of their parents.

But, after a prolonged discussion with the research ethics office, and to proceed with the research, I submitted an application which stipulated the use of opt-in parental consent for both the questionnaire and interview stages. Yet, when subsequently negotiating access to the schools, the teachers suggested using opt-out consent, especially for the questionnaire stage, as they did not think the research risked any harm to their students, and because obtaining opt-out parental consents would require less time and effort. In particular, the teachers suggested that parents do not necessarily object to their child's participation in research but, and especially for those from certain socioeconomic backgrounds, can have more pressing priorities than reading a research information sheet and signing and returning a consent form.

However, even when informed of this, the research ethics office would not modify their position. And, as was expected by the teachers, the need to obtain opt-in parental consent meant I was only able to achieve a relatively small overall sample size. This was mainly due to mismatches between parents’ and their children's wishes: in some cases, the parents consented to their children's participation but the children themselves did not; while, in others, the pattern was reversed. Indeed, the number of such mismatches far exceeded that of matches. A further consequence was that I was also unable to recruit participants from the full range of year groups in any of the participating schools. This, in turn, made it difficult to use the questionnaire data to examine the wider patterns in participants’ views and, instead, almost reduced its role to providing background information about the participants whom I interviewed at the second stage. In brief, and due to the restrictions placed upon it at the ethics review stage, the research became a highly compromised mixed-methods study.

Given institutional variations and my own experiences, when I began a full-time paid post in another institution (where I am based at the time of writing this paper) towards the end of the project, I made enquiries to explore whether I might have had a different outcome with the study's ethics approval had I done it at the new institution. However, after several email exchanges with its research ethics office and other colleagues, the impression I developed was that it would have been similar. While variations do exist, the difficulty concerning obtaining parental consent in research involving young people seems to be especially widespread. Therefore, and as was discussed earlier, due to the burden of obtaining such consents, some institutions actively discourage researchers from involving young people in their research (Sherwood and Parsons 2021).

Power effects in relations with on-site gatekeepers

Given the focus of the research on the role of young people's socioeconomic status, I was fortunate to be able to recruit three schools in locations with different socioeconomic characteristics: an affluent town in the London commuter belt; a city in the east of England undergoing a process of economic transition; and a town to the east of London with a history of traditional manufacturing employment. Gaining access to these schools and afterwards agreeing the timing to collect data was, however, a protracted process lasting over eighteen months. This was initially because research participation was (understandably) a low priority for most school leaders and then because of the restrictions due to the Covid-19 pandemic.

Campbell et al.’s (2006) ‘revolving door’ model of continuously negotiating access with gatekeepers aptly represents both the long period of time I had to wait for the data collection to begin and the data collection itself. As I had no physical access to the schools due to the pandemic, I could not help but rely on the teachers for almost all aspects of the online/remote data collection process (revised from the originally planned face-to-face methods before the pandemic), including sending out participation invitations to students, informing parents about the research to obtain their consent, and placing the participating students who I was scheduled to interview in front of the computer at the agreed dates and times. Everything that occurred was due to the teachers’ voluntary work for me, for which I was immensely grateful, but I was also very careful to continue to retain their good will. This sometimes felt like walking on thin ice because there was frequently no response to my emails and no certainty as to whether the interviews would take place as planned if other priorities intervened.

Given that the teachers were finding time out of their busy schedules to help me, I often accepted what they offered to organize rather than insisting upon what I needed to properly fulfil my study. For example, in one school, instead of asking how many interviews I needed to undertake and how long each one should last, the teacher gave me a list of the dates and times when they could coordinate the online interviews. I accepted this even though the total time suggested would not allow me to interview all the participants I had identified on the basis of their questionnaire responses. Consequently, the 40–45 min I had initially planned for each interview was reduced to a maximum of 30 min in order for me to speak to enough students to explore how the patterns in their responses might relate to the factors I was investigating. However, and in practice, some interviews were much shorter than 30 min due to delays in the students arriving while, in other cases, it was only discovered that the students were absent from school at the time they were supposed to be interviewed.

I also relied on the gatekeepers for any additional information, other than what was publicly available on websites, to help me understand the schools as research contexts, and thereby what my participants had to say. Some information was refused on data protection grounds: for instance, while I had hoped to corroborate the socioeconomic status of my participants, as suggested from information in their questionnaire responses, against their free school meals eligibility, the provision of this information was declined. But some of my other requests for information were met with very short answers while some information was not provided at all. Yet I was often hesitant to follow up on my enquiries because I feared that persistent requests (even if these were made due to previous ones having been unanswered) might risk exhausting the hospitality being offered, so closing off the research relationship.

Here my own positionality in relation to the gatekeepers may also have affected the amount of time I was allotted for interviews and the information I was able to obtain. In particular, my institutional position as a ‘visiting’ researcher discussed above may have highlighted my temporary and ambiguous academic status while, and despite my prior experience of researching in UK schools, my ‘foreignness’ might have suggested a limited understanding of the British context. These factors, combined with my other characteristics (e.g. my reserved demeanour), may have acted to reduce my academic authority in the eyes of my gatekeepers.

None of the above should be seen as an attempt by the gatekeepers to deliberately control or influence the research: the hierarchical power structures and multiple levels of gatekeeping in schools (Fecke et al. 2022), the teachers’ normal responsibilities and the additional pressures imposed on them due to the pandemic (Beattie, Wilson and Hendry 2022, Kim, Oxley and Asbury 2022) were all important factors. Instead, the above is intended to show how ‘power as effects’ were operating – deriving especially from my dependence on the gatekeepers for the successful completion of the research – which determined where, when, and how much data I could collect and thereby the ultimate findings of the study.

Power effects in relations with young participants

I now discuss the power effects that arose in my relations with the young participants themselves. Here it is possible to argue that some of the power effects I experienced may have been influenced by the particular approach the study took. As was noted earlier, research that is derived from young people's own interests and concerns or which involves them from the planning stage may attract greater commitment from them, although such research will entail its own forms of power dynamics to negotiate (this would also apply to the on-site gatekeepers given their own interests and concerns). In comparison, my study was initiated from a theoretical interest in the mechanisms underlying the reproduction of socioeconomic inequality. While it focused on the roles of young people's own views and agency, they were the factors to examine rather than those to actively engage with as would be in the former type of research. Young people therefore may relate less to this kind of research – or even show little or no interest at all if it cannot be explained why it is relevant to them – which, in turn, can affect the power dynamics between them and the researcher.

For example, while my participants were entitled to choose whether to take part in the research, a decision not to do so on their part, and especially during the data collection, could have had serious consequences for the study, given the small sample size. While Tyldum (2012, p. 199) suggests ‘various forms and degrees of institutional, economic and emotional pressures are widely used to recruit’ participants although it is taboo to acknowledge this, I was not in a position to apply such pressures even if I had wanted to. Moreover, my own ethical self-government may even have further contributed to the small final sample size. In particular, my subjectification as an ‘ethical’ researcher, alongside my knowledge of the teachers’ potential influence over their students, meant that I repeatedly stressed in my communications that the study required voluntary participation. Indeed, the only influence I sought to exercise during the participant recruitment processes was to ask the gatekeepers (sometimes to their annoyance) to send their students email reminders about the study.

As noted above, participants are not required to maintain their commitment to the research after giving their initial consent. While researchers should always respect their young participants’ own reasons for taking part (Lohmeyer 2020), any lack of cooperation can affect the successful completion of a study. In this research, although my participants had all volunteered after reading a research information sheet and signing a consent form, this did not always indicate a genuine interest in contributing to the study, perhaps for the reasons noted above. For example, some respondents only took about five minutes to finish the online questionnaire, whereas, when piloting it with other secondary school students, it had taken around 15–20 min to complete. In a similar vein, one teacher said to me that some of the students would be happy to skip a class with the excuse of being a research participant.

My limited influence over the young people – compared to that of the teachers and the schools – was also felt when the teachers suggested that the interviews would be better to be held during school time. Although I had wanted to avoid interrupting their learning by interviewing them outside of normal school hours, I soon realized that, had I done this, I would have struggled to complete the study. When, after having gained the gatekeeper's permission, I tried to arrange interviews with students who I was unable to meet at the times arranged by one of the schools, only two out of seven students I contacted responded even after reminder emails, while the other five did not reply even to withdraw from the study.

The interviews with the young participants were themselves processes of continuously gauging my relations with them and thus what I would or would not be able to ask or hear from them. Here, too, the dynamics of power as effects were clearly in evidence. In particular, and while it is typically older and more mature participants who are less likely to be influenced by the researcher (McLeod 2007), relations with younger ones are not always easier to manage. Indeed, in my study it was the youngest students in Years 7 and 8 with whom I struggled the most during the interviews. For example, there were some participants (albeit a minority) who, throughout the time, showed no intention of giving anything but short and superficial responses. In a couple of cases, due to their apparent attitudes and my concern over their possible reactions, I was not always able to ask probing questions.

In particular, I felt very uncomfortable throughout the interview with one Year 7 male student whose attitude and tone of voice I found aggressive – although this may not have been what he had intended. For instance, when I asked whether he would consider himself to be rich or poor, he responded that ‘I’d stay where I am now’. After cautiously probing what he meant by this, to which he responded just ‘working class’, I was able to probe further by indicating my own lack of clarity about the meaning of the term. However, his demeanour meant I felt unable to ask ‘why’ he wanted to stay that way. In other interviews, I might have been able to use my implied ignorance about the UK class system as a ‘foreign’ researcher to probe further. Yet, in this case, I was too worried if he would interpret such probing as devaluing his identity and react in ways that I would have found difficult to address remotely.

My positionality may also have contributed to the wider power dynamics between the participants and me. While it is unknown how my identity was constructed by each of the young participants, those of my characteristics they were able to notice, alongside any information they might have had about me and the research, may have influenced their attitudes towards the interviews. In some cases, as indicated above, my positionality may have enabled me to ask certain questions more easily although, at other times, it may not have or may have created other limitations. Perhaps, more tellingly, a teacher's comment when discussing their school's safeguarding policies, that ‘they are there to protect you as much as students’ made me wonder whether this would have been said to a researcher who was not a diminutive Asian woman with a foreign accent and unless they were conscious of the potential dynamics that might arise between their mainly white British students and me.

Conclusion

Social research involving young people is a site where strong and complex discourses mingle with and compete against each other concerning young people's competence, agency and vulnerability to influence how such research is conducted and evaluated. However, and despite an increasing emphasis on young people's competence and agency, the prevalent assumption is still that the adult researcher is in overall control of their research and is therefore in a dominant position via-a-vis the young participants. Yet, as I have shown in this paper, various factors can combine to make it difficult for the researcher to exert sufficient control over their research to secure its successful completion. By using a Foucauldian conception of power as effects, I have tried to make sense of the power dynamics involved in my recent study with UK secondary school students during the Covid-19 pandemic. It is important to underline that, according to this understanding, power effects are not always intended by the subjects involved in the relationships, nor are they necessarily illegitimate.

As noted above, ethics regulation is essential to help prevent harm to either the young research participants or the researcher themselves and can also have beneficial impacts on the researcher's own conduct in the field. In addition, ethics regulation can take different forms depending on the institution involved. However, and in my study, an ethics review which operated under the assumption that the young participants were not sufficiently competent to give their informed consent independently (despite them all being of secondary school age and despite the views of their teachers), which was reinforced by a concern with risk management at the organization level, created significant constraints on the research design and the sample size achieved – and consequently the scope and quality of the study itself.

The effects of ethics regulation were not limited to procedural requirements alone because, in compliance with the idea of being an ‘ethical’ researcher, I self-regulated my own conduct in my relations with the on-site gatekeepers and the young participants and this, as explained above, may have further constrained the scope of the study. Likewise, in my relations with the on-site gatekeepers and the young participants, my unique positionality as a foreign Asian woman with only a temporary and marginal institutional position may also have contributed to producing the particular power effects I experienced during the fieldwork. Finally, all these power effects may have been further accentuated by my reliance on online data collection methods and the additional pressures the pandemic had created for my on-site gatekeepers.

Given variations in university ethics reviews, the impact of the pandemic, and because the associated power dynamics may have played out differently in research using different designs and methods, my experience may not be representative. Nonetheless, I believe it highlights factors that frequently operate to produce power effects that can influence the quality of research involving young people. Conceptualizing power relations in the manner of Foucault, as lying on a continuum between ‘strategic games between liberties’ and ‘the states of domination’, those that I experienced during my research in terms of their effects on the final outcome appeared closer to the ‘domination’ end. This resulted from my dependence on the research ethics committee, the on-site gatekeepers and the young participants for the completion of the research. While this may be as it should be, the consequences for the practice of research must also be recognized, especially by research ethics committees.

If so, and in order for the power relations between research ethics committees and researchers to become more ‘positively’ productive (i.e. contribute to genuinely useful debates on ethical issues and help to generate more useful research findings), I would suggest that ethics reviews – and especially in institutions where such regulations are applied rather rigidly – be based on more realistic and nuanced views concerning young people's competence in relation to their characteristics (e.g. age, maturity, aspects of vulnerability, etc.), the particularity of the field in which the research is to be conducted, and the nature and topic of the study. Such reviews should also pay more attention to the multiple forms of power dynamics that are often involved in such research other than those arising between the researcher and the young participants, such as those between the on-site gatekeepers and the young people for whom they are responsible and the on-site gatekeepers and the researcher.

Acknowledgements

I am very grateful to Emeritus Professor Martyn Hammersley of the Open University, UK who read drafts of this paper and provided valuable guidance. I would also like to thank Professor Steve Jones and Dr Jenna Mittelmeier of the University of Manchester, UK for their comments on the first draft of this paper.

Disclosure statement

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

Additional information

Funding

This work was supported by the British Academy/Leverhulme Small Research Grants scheme under SRG19\190380.