Qualitative research methods for evaluating training effectiveness
Public health training is most useful when new knowledge survives beyond the workshop and reaches frontline decisions. A sound evaluation should therefore move beyond attendance and satisfaction to trace changes in confidence, routine practice, service delivery and, where relevant, population outcomes.
Qualitative research methods for evaluating public health training effectiveness make those pathways visible. Interviews, focus groups, observation and document analysis help explain what participants learned, what they could apply, and what organisational conditions shaped transfer to practice.
Frame the inquiry around learning and practice
Begin with a concise logic model that links inputs, activities, outputs and intended outcomes. Define training effectiveness operationally: immediate learning gains, workplace application after 30–90 days, and longer-term service or population effects. Attendance and pre/post scores provide useful context, but qualitative evidence is needed to interpret learning transfer and implementation fidelity.
In Australia, that model must accommodate state and territory governance, local health networks and non-government partners. A course delivered to environmental health officers in regional Victoria may travel differently from one for nurses in a Sydney metropolitan service, so organisational role, resources and policy authority belong in the analysis from the outset.
Select participants for informative comparisons
Choose participants who can illuminate the evaluation question rather than simply those easiest to contact. Purposive sampling may include high and low performers, experienced and newer staff, participants who applied the learning and those who did not, plus supervisors and partner representatives. Set sample size through information power or a defensible saturation rationale, and document non-participation.
Australian public health teams are often multidisciplinary and geographically dispersed, so recruitment should not rely on a single profession or channel. In Melbourne, for example, staff may move between council, hospital and community settings; in smaller regional services, the trainer may also be the only person authorised to implement a new procedure.
Build inclusive discussion guides
Use discussion guides that move from concrete experience to institutional conditions. Semi-structured interviews can ask what changed after the session and what made that change possible. Focus groups can compare peer strategies, while probes about time, workload, leadership support and available tools reveal facilitators and barriers. Reserve sensitive topics for interviews where group members might feel unable to disagree.
For evaluations involving Aboriginal and Torres Strait Islander communities, follow locally agreed protocols and partner with community-controlled health organisations where appropriate. Community participation should be paid, and interpretation should recognise Indigenous data sovereignty. Facilitators can also review focus group guidance when adapting prompts, without replacing local cultural advice.
Collect consistent field data
Collect data through more than one mode when the question warrants it. Record interviews and focus groups with consent, produce expanded field notes after observations, and retain relevant artefacts such as checklists, referral pathways or revised policy drafts. A brief reflexive journal records the researcher’s influence on access, interaction and emerging interpretations.
Standardise core prompts across sites while allowing local examples, but train multiple note-takers if fieldwork is extensive. In the Australian training market, short online modules are often combined with workplace coaching, so participants may be interviewed while still developing competence. This supports analysis of implementation as a process rather than a single event.
Analyse transcripts and field notes
Analyse transcripts and notes systematically. Begin with familiarisation, then develop an initial codebook that combines deductive concepts from the logic model with inductive codes emerging from participants’ language. Use thematic analysis to examine patterns within and across cases, and retain illustrative extracts that show variation rather than presenting a single polished narrative.
Quality checks should include a second coder for a meaningful subset, an audit trail of coding decisions and reflexive memos that challenge early assumptions. Software can organise large datasets, but it cannot make analysis rigorous by itself. Report negative cases, disagreements and context-specific explanations so credible findings remain distinct from promotional claims.
Establish credibility and ethical care
Protect participants through informed consent, clear limits on confidentiality and secure, access-controlled storage. Follow the relevant ethics approval and applicable Australian privacy obligations, and disclose how de-identification will work in quotations. Small groups require particular care because colleagues may recognise one another even when names are removed.
Strengthen credibility through triangulation, member checking where participants wish to review summaries, and transparent reporting of sampling limits. Transferability depends on a thick description of setting, workforce and policy context, not on claiming findings apply everywhere. Keep participant quotations precise, yet never allow a vivid quotation to carry evidence beyond the data that support it.
Turn findings into training action
Turn analysis into prioritised actions by linking each theme to a decision owner, resource and timeframe. A recurring theme about limited supervisory support may justify coaching or protected implementation time, while low confidence in data interpretation may justify practice-based assessment. Separate findings that require a curriculum change from those needing system-level action.
Teams can compare course and field-placement approaches through the regional course catalogue, then adapt—not copy—their evaluation design to an Australian setting. Begin the next evaluation by mapping the logic model, recruiting a small purposive participant group and booking interviews for the first month after delivery. Store the resulting evidence in a decision register that assigns an owner and deadline to every proposed training change.