| Area | What I work with | Boundary to preserve |
|---|---|---|
| Psychometrics and cross-cultural adaptation | Translation and adaptation, structure, reliability and validity, measurement invariance and brief-form candidates. | Evidence remains specific to the population, language, construct and scoring method. |
| Cross-sectional multivariable research | Associations, path models, sensitivity analyses and latent profiles. | Patterns can be described, but causation and temporal order cannot be established. |
| Intervention development and implementation evaluation | Delphi appraisal, feasibility, acceptability, process and participation reporting. | Feasibility is not efficacy; satisfaction is not clinical improvement. |
| Theory and critical review | Conceptual distinctions, mechanism mapping, ethics and reporting frameworks. | A proposed framework is not a validated framework. |
Principles across studies
- Keep constructs separate from indicators, for example digital well-being is not screen duration.
- Separate measured change, personally valued change and causal attribution.
- Report programme participation, activity completion, research response and analysis inclusion separately.
Understanding these methods
Measurement invariance
This asks whether a measure has comparable meaning across groups, rather than whether average scores happen to be similar. For example, different interpretations of an item may limit direct score comparisons. My Chinese-language measurement work considers this alongside structural, reliability and validity evidence.
Latent profile analysis
This explores patterns across several indicators, such as different configurations of sleep, digital experience and support. My resource research examines associations with flourishing and sensitivity to model choices, without treating statistical profiles as fixed types of people.
Delphi expert appraisal
Several rounds of expert feedback are used to organise views and refine content. My online programme development uses a two-round modified Delphi approach. Expert endorsement of content does not establish efficacy.
Implementation feasibility and efficacy
Feasibility concerns whether a programme can be delivered and how participants evaluate it; efficacy concerns whether an intervention causes meaningful improvement. My single-cohort pilot supports implementation learning, while exploratory pre–post changes do not replace controlled evaluation.