The Ethics Of Educational Research

Robert G. Burgess

• Publisher: Routledge
• Year: 2005
• eBook ISBN: 9781135389000
• Topic: Education
• Subtopic: Education General

This collection of papers examines ethical issues in different kinds of social research including surveys, ethnography and historical research. Contributors also deal with ethical problems involved in examining controversial issues in education. Hence the book is largely about the ethics of the conduct of social investigation, rather than an analysis of the technical procedures themselves.

Part One
Ethical Issues and Research Methods

1
Ethics and Tactics: Issues Arising from an Educational Survey

David Raffe, Ivor Bundell and John Bibby

Introduction

Discussions of research ethics tend to focus on qualitative or experimental methods. Survey researchers probably consciously confront ethical problems less frequently than most other social researchers; respected textbooks of survey methods make no reference to ethical issues (Moser and Kalton, 1971; Hoinville and Jowell, 1978).

True, survey respondents typically experience less inconvenience and intrusion than the subjects of other research studies; and a survey respondent, buried in a large sample, may feel less threatened by the publication of results. But surveys too have their ethical problems. The collection, storage and linkage of personal databases contain at least the potential for abuse. And the greater external validity claimed for survey research makes the use of its results even more critical. As the Radical Statistics Education Group (1982, p. 3) has noted, ‘the use of “statistics” and “computers” is often thought to lend an aura of infallibility to research results’ which may be ‘used to silence the legitimate concerns of those wishing to speak up for their own interests’.

The relative neglect of ethical issues in survey research may partly arise from the research process itself. Survey researchers, especially those using mail surveys, typically have little direct personal contact with their subjects; yet it is in the context of direct personal relationships that ethical issues are most often raised. In this chapter we discuss how ethical issues arise in the practice of survey research. This practice conditions whether or not an ethical principle is acknowledged in a particular case, how it is interpreted and how conflicts between principles are resolved. Our discussion is based on our own experience of conducting the Scottish Young People’s Survey, formerly the Scottish School Leavers’ Survey.

The Scottish Young People’s Survey (SYPS)

The Scottish School Leavers’ Survey was first carried out by the Centre for Educational Sociology (CES) in 1971 (Burnhill, McPherson, Raffe and Tomes, 1987). It was then a postal survey of a sample of qualified school leavers in Scotland from the previous school session. There was another similar survey in 1973. Since 1977 the survey has been conducted biennially, and has covered all types of school leavers, including the unqualified. Sample fractions have ranged from 10 per cent to nearly 40 per cent.

Since 1985 (with a smaller pilot study in 1984) the survey has developed in two further ways. First, in addition to surveying the previous year’s leavers, it now covers an overlapping sample of the previous year’s fourth year (equivalent to the English fifth year), including both leavers and stayers. Second, this sample is followed up in subsequent surveys, currently extending to age 19. These changes coincided with the change of name (to the Scottish Young People’s Survey) and with the introduction of the Youth Cohort Study in England and Wales, which drew on the Scottish experience (Clough and Gray, 1986). The SYPS continues to use postal questionnaires to collect both attitudinal and ‘objective’ data covering a range of topics, particularly concerning education and the labour market.

Initially funded by the Scottish Education Department (SED), the survey is now funded by the SED, the Manpower Services Commission (MSC), the Industry Department for Scotland and the Department of Employment. However the 1977 and 1981 surveys were also funded by the Social Science Research Council (SSRC, now the ESRC) as part of the Collaborative Research Programme (see below). The CES has received continuous SSRC/ESRC support since 1974, and may expect continuity for a further eight years from 1987 when it becomes a designated Research Centre of the ESRC. The balance between government and Research Council funding is an important ingredient of the CES’s status as an applied but independent research body.

Throughout its existence the survey has obtained sample names and addresses from the SED. However, at a critical point in its development it needed to supplement these with names and addresses obtained from the local authorities. This arose from the desire to extend the coverage of the 1977 survey, and increase the sample size, in connection with the SSRC-funded Collaborative Research Programme which ran from 1975 to 1982 (Gray, McPherson and Raffe, 1983). This aimed to break down some of the boundaries between researchers, practitioners and others by helping non-professional researchers to influence the survey design and to analyze survey data. In the course of this programme the CES adopted a Code of Practice (McPherson and Raffe, 1979). Among other things this protected the confidentiality of data concerning individual sample members and schools; subject to this it provided for open access to the data which are stored in the Scottish Education Data Archive. Other clauses of the code, for example concerning wide consultation over the questionnaire, have fallen into abeyance with the termination of funding for collaborative research, but the principles of anonymity and public access to the data are preserved, for example in current contracts with government departments which refer to the code.

The survey is also subject to other regulatory codes. The Data Protection Act (1984) requires users of personal data held on computers to register with the Data Protection Registrar’s Office established by the Act. Most medium or large surveys store personal data on computer and are thus covered by the Act; academic survey researchers usually register through their parent institution—in the case of CES, the University of Edinburgh. The broad aims of the Data Protection Act are embodied in eight principles. However data held for ‘historical and research’ purposes are exempted from the principle which gives individuals the right of access to personal data about themselves, provided the data are not made available in a form which identifies individuals. Research data also have partial exemption from two further principles, with the effect that such data may be held indefinitely and the use of the data for research purposes need not be disclosed at the time of data collection.

Of the principles which do concern research data, the two most important are probably the principle which states that personal data ‘shall not be used or disclosed in any matter incompatible with’ the purpose for which they are held, in this case research; and the principle which requires ‘appropriate security measures’ to be ‘taken against unauthorized access to, or alteration, disclosure or destruction of the data. A further principle, which many researchers might find troublesome, states that personal data should be ‘accurate’, although this is framed more to protect individuals against damage resulting from inaccuracy, than to enhance the validity of measurement in social research. The main effect of the Data Protection Act on the work of the CES has been to ensure a thorough review of current security practices regarding personal data, both on-line and off. The task of registering has also involved a considerable amount of extra work.

Since the CES uses government data, particularly in the construction of the sample, it is affected by the Code of Practice of the Government Statistical Service (GSS, 1984). The main purpose of this code is to protect the confidentiality of data collected by government departments from ‘statistical units’ such as individuals, households or (in some cases) schools. There are only three circumstances under which these data may be subsequently transferred to other departments or organizations in a form which identifies individual respondents: if the transfer is specifically provided for by law; if the respondents have given their consent; or if the information is transferred to bona fide researchers and for statistical purposes only, and the transfer has prior written ministerial authorization. Before 1983 the SED only provided the CES with details of sample members if they had consented to take part—or, more precisely, if they had not withheld their consent. This resulted in a level of opting-out which, although not large in absolute terms, was skewed towards the less qualified. We discuss some of the implications of this below. Since 1983, ministerial authorization has been given for the transfer and sample members have not had a prior opportunity to opt out.

The thrust of the GSS code (1984) is to protect individual confidentiality, and not to restrict access to information. The transfer of anonymous data is specifically allowed for, and the preamble to the code notes that ‘if proper safeguards on confidentiality are applied, it is to the general advantage if data collected for statistical purposes inside government are also available to outside analysts and researchers’ (p.).

Some Ethical Principles

The above discussion of regulatory codes has raised ethical issues relating to accuracy, confidentiality, breadth of consultation, rights of access, and continuity of purpose. However, ethical principles range far wider than regulatory codes, including the CES’s own Code of Practice. We do not attempt here either to review the literature or to propose a comprehensive ethical code for survey research. However we suspect that the following set of principles, or something similar, would attract fairly wide assent.

Ethical principles in survey research, we suggest, concern the proper conduct of relationships between researchers and three groups of people: resource-providers, data subjects, and a broad amorphous group that we shall call ‘the public’. There is of course an overlap here: data subjects are key resource-providers in the research production process.

With respect to providers of resources, such as funders or employers, researchers should clarify their obligations in advance, and honour them honestly. As a general rule, resource-providers should not have rights of confidentiality; nor should any intellectual property rights allow them to inhibit publication of results. This latter is especially important for methodological aspects of the research, which should be explicated for dissemination and public appraisal. The prior existence of suitable codes of conduct can strengthen the researcher’s hand in ensuring that these conditions are incorporated into any written contracts.

A particular group of resource-providers whose importance may be overlooked are the ‘gatekeepers’ who control access to sampling frames and/or sample members. Relationships between gatekeepers and researchers are often determined more by bargaining power and by the dictates of sound business practice than by any lofty ethical ideals. However, where the gatekeeper has a personal relationship with the data subject (for example, as proxy, parent, or teacher), researchers should be sensitive to the nature of that relationship, and to the private spaces which it maintains. This does not necessarily mean that one should accept the gatekeeper’s word as final, especially where the gatekeeper is in a position of considerable power over the data subjects.

With respect to data subjects, researchers should be conscious of their intrusive potential, and should seek to minimize any intrusion; the confidentiality of data must be respected and protected by positive measures; and data subjects should be told the purposes of the research and should have adequate opportunity to withhold their cooperation. (This last is the principle of ‘informed consent’.) ‘Snooping’ and ‘voyeurism’, and intrusive questioning which may merely reflect a hobby or passing interest on the part of the researcher, should be avoided (Huizer, 1973, p. 170).

With respect to the public, researchers should pursue openness, sensitivity, accuracy, honesty and objectivity in their choice of topic, methods, analysis and dissemination. This includes respecting the interests of different groups in society; avoiding research designs which preclude particular outcomes of the enquiry; disseminating findings fully, as well as widely; and facilitating the re-use of data.

In outlining these principles we have drawn upon a range of existing codes (British Sociological Association (BSA) 1973; International Statistical Institute (ISI) 1986; Market Research Society (MRS) 1986; Association des Administrateurs de l’INSEE, 1985) as well as upon recent and forthcoming statistical texts (Moore, 1985; Bibby and Moore, forthcoming). There are, of course, more general issues underlying most of these principles: many of them reflect the perception that ‘knowledge is power’ and enjoin a continuous resensitizing to power relationships, not only between researchers and data-subjects but also those involving resource-providers and the public. A corollary of this last point is that it is not only the researcher who should act ethically. The principles outlined above are, nevertheless, not comprehensive. Among other things they make no reference to relationships within a research team, which probably deserve close ethical scrutiny particularly given the currently limited opportunities and careers in research.

The application of these principles in any given situation will require judgment upon which two honest researchers could honestly disagree. In addition, several tensions in these principles are readily discernible:

– between principles and the feasible actions in any given situation;

– between different ethical principles; and

– between groups, as mentioned above, and other individuals or groups; for example, in emphasizing the ‘right’ of (randomly selected) data subjects to withhold consent, do we infringe the ‘right’ of groups to which they belong to be adequately represented in the data?

These tensions will be apparent in the following three sections, which reflect on ethical decision-making in the context of the CES surveys, and discuss the ways in which it is influenced by the practice or tactics of research. They describe how the practice of survey research may influence, respectively, whether an ethical issue is recognized in any particular context, how general principles are applied to particular situations, and how conflicts between principles are resolved.

Non-decision-making and Ethical Issues

We have discussed ethical principles in terms of the researcher’s relations with three categories of people. For certain purposes these may be regrouped into two main groups. The first consists of all those on whom the researcher is dependent for resources or data; it includes the providers of resources and those data-subjects from whom data are collected directly. The second group comprises those on whom the researcher is ...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• Series Editor’s Preface
• Preface
• Ethics and Educational Research: An Introduction
• Part One: Ethical Issues and Research Methods
• Part Two: Ethical Issues In Empirical Research
• Notes On Contributors