Instruments and Scales for a Social Work Dissertation in 2026: Choosing, Adapting and Reporting Measures

Social work research occupies an uncomfortable methodological position. The constructs that matter most to the profession — wellbeing, family functioning, resilience, the quality of a helping relationship, the corrosive effects of the work on the practitioner — are precisely the constructs that resist tidy measurement. Meanwhile, the populations who could tell you most about them are, almost by definition, the populations that ethics committees protect most carefully.

The result is that measurement decisions in a social work dissertation carry an ethical weight they do not carry in most disciplines. Choosing an instrument is not only a psychometric question about reliability coefficients; it is a question about what it is defensible to ask a person who is receiving a service, in a research encounter they cannot easily decline, about experiences that may be actively distressing.

This guide sets out the instruments that are realistically available to a student researcher, the psychometric evidence you need to establish before using them, and the ethical architecture that has to surround their use.

First: does your study need a standardised instrument at all?

A substantial proportion of social work dissertations are qualitative, and rightly so. If your research question concerns how service users experience a process, how practitioners construct professional judgement, or what meaning participants attach to an intervention, a semi-structured interview schedule is your instrument and no scale is required. In that case the methodological work lies in topic guide design, sampling and analytic rigour — our walkthrough of coding interviews in NVivo step by step covers the analytic side.

Standardised instruments earn their place when you need to:

• measure change over time, such as before and after an intervention;
• compare a sample against published norms or a comparison group;
• describe the level of a construct in a population precisely;
• supply the quantitative strand of a mixed-methods design.

If none of those applies, adding a scale to a qualitative study is decoration, and examiners recognise it as such.

Instruments organised by what social work dissertations actually measure

Mental health and wellbeing

• WEMWBS / SWEMWBS — positively worded mental wellbeing measures, free after registration and widely used in UK service evaluation. The short seven-item version is well suited to service settings where burden matters.
• PHQ-9 and GAD-7 — public domain screeners for depressive and anxiety symptoms, with established thresholds. Extremely well documented, but see the ethical caveats below before administering them.
• Kessler K6 / K10 — brief non-specific psychological distress measures, free, and designed for population and service surveys rather than diagnosis.
• WHO-5 Wellbeing Index — five items, free, exceptionally low burden, useful when your questionnaire is already long.
• CORE-10 and CORE-OM — psychological distress measures developed for and widely embedded in UK counselling and psychological services, free for clinical and research use.

Children, young people and families

• Strengths and Difficulties Questionnaire (SDQ) — the standard brief behavioural screening measure for children and young people, with parent, teacher and self-report versions, free for research use in paper form and available in many languages.
• Family Assessment Device (McMaster FAD) — family functioning across seven dimensions, widely used in family research.
• Parenting Scale and various parenting-efficacy measures — check licensing individually, as some parallel instruments in this area, including the Parenting Stress Index, are commercially published.
• Adverse Childhood Experiences questionnaire — freely available, but ethically the most demanding instrument on this list, since it asks directly about abuse, neglect and household dysfunction. It requires a full distress protocol and, in most student projects, a genuinely strong justification.
• Child and Adolescent Needs and Strengths (CANS) and the Outcomes Star family of tools — widely used in practice, but the Outcomes Star tools are licensed and typically accessed through the agency that already uses them rather than by an individual student.

Social support, isolation and connectedness

• MSPSS — twelve items, three sources of support (family, friends, significant other), free and cleanly structured.
• UCLA Loneliness Scale (Version 3) and the De Jong Gierveld six-item scale — the standard loneliness measures, with the latter distinguishing emotional from social loneliness.
• Social Provisions Scale and the Duke Social Support Index — useful where the theoretical framing is about functions of support rather than sources.

Practitioner wellbeing and the workforce

Dissertations on social workers themselves — burnout, retention, secondary trauma, supervision quality — are among the most feasible student projects, because the sample is professionally accessible and the ethics route is generally lighter than research with service users.

• ProQOL (Professional Quality of Life Scale) — measures compassion satisfaction, burnout and secondary traumatic stress; free for research use with attribution, and almost purpose-built for social work workforce studies.
• Copenhagen Burnout Inventory and the Oldenburg Burnout Inventory — free alternatives to the commercially licensed Maslach Burnout Inventory, which most students cannot afford.
• Secondary Traumatic Stress Scale — targeted measure of indirect trauma exposure.
• Perceived Stress Scale (PSS-10) — free, general appraisal of stress over the past month, with four reverse-scored items.
• Utrecht Work Engagement Scale — free for non-commercial research, useful as the positive counterpart to burnout.

Process, relationship and satisfaction

• Working Alliance Inventory, short revised form (WAI-SR) — measures the bond, task and goal components of the helping relationship, with client and practitioner versions.
• Client Satisfaction Questionnaire (CSQ-8) — brief service satisfaction; check licensing terms, which have varied over time.
• Goal Attainment Scaling — not a fixed scale but a method for constructing individualised outcome measurement, which suits social work’s person-centred logic better than many standardised instruments and is free to use.

Evaluating an instrument before you commit to it

Search the literature properly rather than accepting the first scale you encounter. Google Scholar, PsycINFO and Social Care Online will each surface validation studies; look specifically for a validation in a population resembling yours, because a measure validated on university students may behave quite differently among care leavers, kinship carers or people using mental health services.

Establish four things for every instrument, and record the source for each:

1. Internal consistency — Cronbach’s alpha above .70, ideally with McDonald’s omega reported alongside. You must compute this in your own sample as well as citing the developer’s figure; reliability is a property of scores, not of instruments.
2. Factor structure — confirm any subscales you intend to analyse separately have been replicated in comparable populations.
3. Sensitivity to change — this is specific to intervention research and frequently overlooked. A scale can be highly reliable and still be too stable to detect the change your intervention produces over eight weeks.
4. Reading level and administration burden — many participants in social work research have interrupted educational histories, English as an additional language, or cognitive or communication needs. A scale written at university reading level is a validity problem, not merely an inconvenience.

If you plan to examine the structure in your own data, our walkthrough of exploratory factor analysis in SPSS covers the sequence and the decisions at each step.

Permission, licensing and adaptation

Verify the licence for every instrument before your ethics submission and keep the evidence. Some are public domain, some free with registration, some free only for non-commercial research, and some — including the Maslach Burnout Inventory, the Parenting Stress Index and several child behaviour batteries — require a paid licence and sometimes a qualified purchaser. Discovering this in week nine is a genuine crisis; discovering it in week two is a five-minute substitution. Our guide to getting permission to use a validated questionnaire covers what to email, what counts as evidence, and what to do when there is no reply.

Do not modify items to fit your setting. Changing wording, dropping items or altering the response format severs the link to published norms and reliability evidence, and many licences prohibit it outright. Where adaptation is genuinely necessary — a different language, or a version for participants with learning disabilities — follow a documented procedure and report it as an adaptation rather than an adoption; our guide to back-translation and cross-cultural adaptation sets out the standard steps.

The ethical layer that social work adds

This is what distinguishes measurement in social work from measurement in most other fields, and it is where dissertations are most often sent back by ethics committees.

Screening instruments generate obligations. If you administer the PHQ-9 and a participant scores in the severe range, or endorses the item about self-harm, you cannot simply record the number and move on. Your protocol must state in advance what you will do, who you will tell, and what support information you will provide. Committees expect a written distress protocol, and its absence is a routine reason for deferral.

Recruitment through a service is never fully voluntary. Where participants are receiving a service and are approached through it, there is an inherent risk that declining feels consequential. Separate the researcher role from the practitioner role explicitly, make the independence of participation unambiguous in the information sheet, and avoid recruiting people from your own caseload. Where you are researching your own workplace, the dual-role problem needs to be addressed head-on in your methods — our discussion of insider research, gatekeepers and the dual-role problem covers how to write that section credibly.

Instruments about harm require justification proportionate to their intrusion. Asking about adverse childhood experiences, domestic abuse or substance use may be entirely appropriate — or may be a question your dissertation does not actually need answered. Committees will ask why a less intrusive measure would not suffice. Have the answer ready.

Safeguarding overrides confidentiality, and participants must know that before consenting. Your information sheet must state plainly that disclosures indicating risk to a child or adult at risk will be passed on in accordance with local procedures. Promising unconditional confidentiality and then breaching it is both an ethical failure and a research-integrity one.

Capacity and proxy respondents. Where participants may lack capacity, the relevant legal framework governs the process, and most student projects are advised away from this territory. Proxy report — a carer completing a measure about the person they support — is a legitimate alternative, but proxy and self-report scores diverge systematically and that divergence must be discussed rather than ignored.

Our broader treatment of social work dissertation methodology and research ethics covers the surrounding design and governance questions, and the wider chapter architecture is set out in our complete guide to writing a social work dissertation.

Analysis implications you should anticipate

Small, clustered, non-normal data is the norm in social work research. Three consequences:

• Sample sizes are small. Twenty-five practitioners or eighteen service users is a realistic dissertation sample, which pushes you toward non-parametric methods and toward reporting effect sizes with confidence intervals rather than relying on p values. Our overview of non-parametric tests covers Mann-Whitney, Wilcoxon and Kruskal-Wallis and what each actually tests.
• Non-significant results are likely and are not failures. With that sample size, absence of evidence is the expected outcome, and our guide on what to do when results are not significant shows how to write it up as a legitimate finding.
• Participants may be clustered within teams, services or local authorities, violating independence. Where that clustering is substantial, our introduction to multilevel models for nested data explains the alternative; where the sample is too small to model it, name the limitation explicitly.

What to report about your measures

For each instrument, the measures subsection should give the full name and abbreviation with the development citation, what it measures, the number of items and subscales, the response format and anchors, a sample item, the scoring rule including reverse-keyed items and score range, previously reported reliability with the population it came from, and reliability in your own sample. Where you have used any cut-off, cite the source of the threshold and state whether it was validated in a comparable group. Where an instrument is used to describe rather than diagnose, say so — “scores above the clinical threshold” is accurate, “participants with depression” is not.

Frequently asked questions