Short answer: Use a published, validated instrument that already measures your construct rather than writing your own. Choose the scale most used in comparable nursing populations, obtain written permission from the developer, check whether it has been used before on a South African sample, and report its origin, subscales, scoring and reliability coefficient in your methodology chapter.
Why should you not write your own nursing questionnaire?
Because an instrument you invent has no psychometric history, and your examiner knows it. A self-made questionnaire obliges you to prove, inside a single honours research report or mini-dissertation, that your items measure what you claim they measure. That means expert content review, a pilot, item analysis, a reliability coefficient and some form of construct validity evidence. It is a full study on its own, and it is the single most common reason a nursing methodology chapter comes back from a supervisor with the comment that the instrument is not defensible.
A validated scale hands you that work already done. When you use the Practice Environment Scale of the Nursing Work Index, you inherit two decades of published psychometrics, a scoring key, established subscales and comparison figures from other countries and other units. Your discussion chapter then has something to compare against, which is exactly what an examiner is looking for when they ask whether your findings are situated in the literature.
There is one legitimate exception. If your construct is genuinely local — compliance with a specific South African protocol, experiences of a particular provincial programme — no international scale will exist and you will have to build one. In that case the instrument development becomes part of your contribution and you must show the full evidence needed to prove a questionnaire is valid, not just a Cronbach’s alpha.
Which validated scales appear most often in South African nursing research?
These are the instruments that recur in South African nursing dissertations and journal articles. The list is not exhaustive; it is a starting point for a literature search in your specific area.
| Construct | Instrument | Notes on access |
|---|---|---|
| Burnout | Maslach Burnout Inventory — Human Services Survey (MBI-HSS) | Licensed commercially through the publisher, charged per administration. Budget for it in rand before you commit. |
| Burnout (no-fee alternative) | Copenhagen Burnout Inventory (CBI) | Freely available for research use; three subscales covering personal, work-related and client-related burnout. |
| Work engagement | Utrecht Work Engagement Scale (UWES) | Free for non-commercial research on request. Has a substantial South African validation literature, which makes it easy to defend locally. |
| Nursing practice environment | Practice Environment Scale of the Nursing Work Index (PES-NWI) | Widely used in staffing and retention studies; five subscales. Request permission from the developer. |
| Occupational stress | Nursing Stress Scale (NSS) and its expanded revisions | Older instrument, well cited, item wording sometimes needs contextual adaptation. |
| Patient safety culture | Hospital Survey on Patient Safety Culture (HSOPSC) | Produced by a public agency and available at no charge with its own scoring guidance. |
| Moral distress | Moral Distress Scale — Revised (MDS-R) | Has separate versions for nurses in different settings; use the version matching your population. |
| Newly qualified nurse transition | Casey-Fink Graduate Nurse Experience Survey | Strong fit for community service nurse studies, which are a distinctly South African research opportunity. |
| Patient satisfaction with nursing care | Newcastle Satisfaction with Nursing Scales (NSNS) | Requires patient participants, which raises your ethics burden considerably. |
| Job satisfaction | Minnesota Satisfaction Questionnaire (MSQ) or the McCloskey/Mueller Satisfaction Scale | Check licensing; the short forms are usually sufficient for a research report. |
How do you choose between two scales measuring the same thing?
Rank them on five criteria, in this order.
- Population match. An instrument validated on hospital nurses may behave differently among primary health care nurses in a rural clinic. Prefer the scale used in studies whose participants look like yours.
- Prior South African use. If a scale has already been administered to South African nurses and the reliability held, cite that study and half your justification is written.
- Length. A 90-item battery administered to nurses on shift will produce a poor response rate and a lot of missing data. Shorter validated forms exist for most constructs and are almost always the better choice for a research report.
- Cost and permission. A commercially licensed instrument that you cannot afford is not an option, however good it is.
- Scoring transparency. You need the scoring key. Some instruments publish item wording but withhold scoring; you cannot analyse what you cannot score.
Do you need permission to use a validated scale?
Yes, and you must be able to show it. Even where a scale is described as freely available, developers almost always ask you to request use, and many supply the current version, the scoring syntax and a translation register in return. Send a short email stating your degree, institution, study title, sample and intended sample size, and keep the reply. Attach that permission email as an appendix. Ethics committees increasingly ask for it, and an examiner who finds a licensed instrument reproduced in an appendix without evidence of permission has found a copyright problem, not a methodological one.
Where a scale is commercially licensed, the licence normally forbids reproducing the full instrument in your appendix. The convention is to include a sample of two or three items with a note that the full instrument is available from the publisher under licence. Say so explicitly rather than leaving the appendix mysteriously thin.
What must you do before using an international scale on a South African sample?
Three things, and they belong in your methodology chapter as separate subsections.
Contextual adaptation. Items referring to structures that do not exist here — specific insurance arrangements, staffing grades with no local equivalent — need adjusting. Any change to item wording must be reported, because it means your instrument is no longer identical to the validated original.
Translation, where needed. If you are administering the scale in isiZulu, Sesotho, isiXhosa or Afrikaans, a casual translation is not acceptable. The standard is forward translation by two independent translators, reconciliation, back-translation by a translator blind to the original, and review by a committee that includes a subject expert. Report the procedure and name the languages.
Re-established reliability. Whatever the published alpha was, you report the alpha obtained in your sample. A translated or adapted instrument that returns a low coefficient in your data must be discussed honestly rather than quietly omitted.
How many participants does your chosen instrument need?
The scale itself influences the number. A study that only describes scores can work with a smaller sample than one that runs a factor analysis, and any subscale-level comparison across units multiplies the requirement. Work the number out from a power analysis tied to the specific test you intend to run, not from a rule of thumb — the logic is set out in the guide to deciding how many participants your study needs, and it is the section your supervisor will interrogate first.
Two practical South African considerations. Nurse response rates in facility-based studies are usually lower than in the international literature because of shift patterns and staffing pressure, so inflate your target. And if you are sampling across more than one facility, each additional site adds a separate permission process, not just extra participants.
What do the ethics committees need to see about your instrument?
Health research in South Africa clears more gates than most other disciplines, and the instrument is part of every one of them.
- Your university’s health research ethics committee. Submit the full instrument, the permission email and your consent documents. Generic timelines and common causes of delay are covered in the overview of what ethics clearance involves at a South African university.
- Provincial department of health approval. Research conducted in public facilities requires provincial approval, normally registered on the national health research database, and the province will review your instrument for participant burden.
- Facility-level permission. The hospital chief executive or clinic operational manager gives the final go-ahead, and often sets conditions about when and where you may approach staff.
Build the sequence into your timeline. These approvals run consecutively, not in parallel, and the instrument you submit at stage one is the instrument you are held to at stage three. Changing a scale after clearance means an amendment, and an amendment means waiting again.
How do you write the instrument up in the methodology chapter?
Give the examiner, in this order: the instrument’s full name and abbreviation, the developers and year, the number of items, the subscales and what each measures, the response format, the scoring direction including any reverse-scored items, the reliability reported in the original study, reliability reported in any prior South African use, any adaptation or translation you performed, and the alpha you obtained. That is one dense paragraph or a short table, and it answers almost every instrument question an examiner can raise. If you are working through the rest of that chapter, the step-by-step build of a methodology chapter shows where the instrument subsection sits relative to design, sampling and analysis.
Then carry it forward. Once data collection closes, your reverse-scoring and subscale computation must happen before any analysis — the mechanics of doing that cleanly, with a syntax trail an examiner can audit, are covered in the walkthrough of analysing questionnaire data in SPSS.
What if no validated scale exists for your construct?
Then say so in the chapter, and say it with evidence: name the databases you searched, the terms you used and the instruments you rejected with reasons. An examiner will accept a self-developed instrument that arrives with that justification and a proper development procedure. What they will not accept is a questionnaire that appears without explanation, as though scales were a matter of personal preference.
The middle option is often best: take a validated scale for the part of your construct it covers, and add a small self-developed section for the local element, reporting the two separately. Be explicit that the composite is not itself validated as a whole.
Getting the instrument section written
The instrument subsection is short, technical and highly structured — which makes it the easiest part of the methodology chapter to draft properly and the easiest to leave until it is too late. Tesify helps you build that section from your own decisions: the scale you selected, the permission you obtained, the adaptations you made and the reliability you calculated, arranged in the order an examiner reads them, with your references formatted to your faculty’s guide. Start your research report with Tesify and get the methodology chapter into a shape you can send to your supervisor this week.
Frequently asked questions
Can I use a validated scale without paying if I am a student?
Sometimes. Many developers waive or reduce fees for unfunded student research, but you must ask in writing and you must not assume. Commercially published instruments generally charge per administration regardless of student status. If the fee is unaffordable, choose a no-fee alternative measuring the same construct rather than using the licensed scale without permission.
Does the South African Nursing Council approve research instruments?
No. The regulator governs nursing practice, education and professional conduct, not the approval of individual research instruments. Instrument approval comes from your university’s health research ethics committee, and access to public facilities comes from the relevant provincial department of health and the facility itself.
What Cronbach’s alpha is acceptable for a nursing scale?
A coefficient of 0,70 is the conventional minimum for research use, and values above 0,90 on a short subscale can indicate redundant items rather than excellence. Report alpha for the total scale and for each subscale separately. A subscale falling below 0,70 should be reported and discussed as a limitation, not deleted silently.
Can I shorten a validated scale to improve my response rate?
Not on your own initiative. Deleting items breaks the validated structure and invalidates comparison with published norms. If length is a problem, use an officially published short form of the instrument, which will have its own psychometric evidence, and cite that version specifically.
Do I need to translate my questionnaire into an African language?
It depends on your participants. Professional nurses are educated in English and English administration is usually defensible, which you should state and justify. Studies involving patients, community members or auxiliary staff generally do require translation, and the ethics committee will expect a described translation procedure rather than an informal one.
Should I pilot a scale that is already validated?
Yes, with a small group drawn from the target population but excluded from the main sample. The pilot is not retesting the psychometrics; it checks comprehension, completion time, instructions and data capture. Report what the pilot changed, even if the answer is that nothing changed.
Can I combine two validated scales in one questionnaire?
Yes, and it is common practice — for example a burnout measure alongside a practice environment measure to test a relationship between them. Obtain permission for each instrument separately, keep the sections clearly labelled, score them independently and report reliability for each. Watch total length, because two full instruments plus demographics can exceed what a nurse on shift will complete.
Where do I find whether a scale has been used in South Africa before?
Search the instrument name together with South Africa in your university’s databases, then search institutional repositories, which hold theses that used the instrument and often reproduce it. South African nursing and health sciences journals are the highest-value sources because they report local reliability figures you can cite directly.
What if my reliability comes out lower than the published figure?
Report it and explain it. Common causes are a small sample, a population differing from the validation sample, translation effects, or a specific item behaving oddly. Inspect the item-total correlations, describe what you found, and treat the affected subscale cautiously in your interpretation. A discussed low alpha is a limitation; a hidden one is a finding against you.
Is a validated scale enough on its own for a master’s dissertation?
Using an existing instrument is entirely acceptable at master’s level; the contribution comes from the population, the setting and the question, not from building a new tool. What must be original is the argument: why this construct, in this South African setting, matters and what your findings add to what is already known.