13 Research methods, statistics, critical review and evidence-based practice13.7 Qualitative research
13.7.5 From qualitative material to measurement
2 min read
Written by Resident Doctor Committee · Published
New in the 2027 syllabus
The College added this topic for exams from January 2027.
Turning qualitative material into numbers
Quantitising converts qualitative material into defined numerical categories or counts. Keep a record of how the original material was coded.
Instrument development can use interviews to generate candidate items. You then test how people understand and answer them before evaluating the instrument quantitatively.
A count depends on the unit you choose, such as participants, mentions or documents. If one person raises the same point several times, that is still one participant.
Common scale formats
Format | How responses become a score | Main limitation |
|---|---|---|
Thurstone scale | Judges give each statement a value for how favourable it is. Respondents then pick the statements they agree with | Scores depend on how well the statements and judges represent the attitude |
Likert scale | Responses to several ordered agreement items are scored and added together | Neighbouring response options are not automatically equally spaced |
Visual analogue scale | A mark on a continuous line is measured from one labelled end | Meaning depends on the end labels and whether the person understands the task |
Guttman scale | Items are in a cumulative order, so agreeing with a harder item implies agreeing with the easier ones | Real answers may break the cumulative pattern |
A Likert-type item is ordinal: the options have an order, but the gaps between them are not known to be equal. Treating a total across several items as roughly continuous needs justification from its properties and the planned analysis.
Reverse scoring brings items worded in the opposite direction into line before they are combined. Always follow the defined scoring rule when reversing an item.
Worked example of coding
Suppose each interview is coded once for whether the person described a transport barrier.
Interview | Barrier described |
|---|---|
A | Yes |
B | No |
C | Yes |
D | Yes |
Code Yes as 1 and No as 0. This gives 1, 0, 1, 1.
Add up the codes. 3 of 4 interviewees described a barrier, which is 75% of this sample.
Do not read 75% as the prevalence in the wider population. The sample was purposively selected (chosen deliberately, not at random).
Check what was lost
Content validity asks whether the final items still cover the construct you meant to measure, including less frequent experiences that matter clinically.
Reliability is about whether the measure gives consistent results. Validity is about whether the way you intend to interpret the scores is justified.
Numerical coding can strip out context and ambiguity. Report the coding rules and their limits alongside the quantitative findings.