design-a-mixed-methods-study

name	design-a-mixed-methods-study
category	school
description	Design a mixed-methods study with a justified integration strategy, aligned samples, ethical collection, and explicit inference limits. Use when qualitative and quantitative evidence must answer connected parts of one question.

design-a-mixed-methods-study

Mix methods at the level of reasoning, not just data collection.

When to use

• Use when neither qualitative nor quantitative evidence alone answers the research question.
• Do not add a second method merely for appearance or use one stream to erase contradictions in the other.

Preconditions

• Define the question, population, decisions, ethics, access, expertise, resources, and likely inference.

Procedure

1. State what each evidence stream contributes and why integration is necessary.
2. Choose convergent, explanatory sequential, exploratory sequential, embedded, or multiphase design with a rationale.
3. Align constructs, units, sampling frames, timing, recruitment, consent, and analysis populations.
4. Define instruments, interview or observation protocols, data quality, reflexivity, and stopping rules.
5. Predefine where integration occurs through sampling, instrument design, analysis, joint display, or interpretation.
6. Plan handling for disagreement, missingness, unequal credibility, vulnerable groups, privacy, and withdrawal.
7. Pilot both streams and the integration mechanism.
8. Write an analysis plan that separates stream-specific results, integrated inference, uncertainty, and limits.

Failure plan

• If one stream fails, report the changed design and restrict the inference.
• If streams conflict, investigate construct, timing, sampling, and measurement before forcing consensus.

Worked example

A school combines attendance records with student interviews to explain why a mentoring program reaches some groups but not others, using a joint display rather than averaging unlike evidence.

Done

• A mixed-methods study plan documents questions, rationale, design, samples, instruments, ethics, integration, analysis, and inference limits
• Pilot, alignment, missingness, disagreement, privacy, withdrawal, and joint-display checks verify executable integration