10. Design Thinking for Research

Iteration as a research virtue

Most research training rewards locking in the protocol early and running it through. Funders expect a fixed design. IRBs expect protocol stability. The result: many studies that are internally consistent but contextually fragile, because nothing was iterated.

Iteration is a research virtue — when it has stop-rules. The discipline of design thinking turns iteration from drift into rigor by pre-specifying what would license a change and how the change would be made.

What to iterate on

The temptation is to iterate only on the instrument: the interview guide, the survey items, the analysis plan. These are the easy targets and the visible ones.

But the elements most often needing iteration — and least often getting it — are procedural:

• Recruitment channels. Where are participants actually reached? Through which trusted intermediaries? In what language and tone?
• Consent process. Is the consent document readable? Where does it confuse? Is it being read in the right setting?
• Compensation structure. Is the amount adequate? Does the form (gift card, cash, voucher) match the context? Does it trigger tax or benefits complications?
• Scheduling logistics. Can participants actually navigate the booking system? Are evenings/weekends available?
• Setting. Does the data collection happen somewhere participants are comfortable? Is the setting culturally appropriate?

A protocol that iterates the interview guide but not the recruitment is iterating where it's easy, not where it matters.

Stop-rules turn iteration into rigor

Iteration without stop-rules is procrastination. Stop-rules turn it into a disciplined design move. A stop-rule has three components:

1. What triggers stop. "When two consecutive cohorts of three participants raise no new feedback on the consent."
2. Who decides. "The PI, in consultation with the community advisory board lead."
3. What gets locked. "The consent document is locked at v3; subsequent feedback is logged but not implemented mid-study."

Without these, iteration is iteration forever. With them, you have a documented design move that ends.

A worked example: recruitment iteration

A team studying caregiver burnout in a multi-ethnic community designed v1 of recruitment: flyers in three languages at clinics, with a phone number to call. After two weeks, zero calls.

Rather than locking the design and pushing through, they paused for iteration. v2: same flyers, plus a partnership with two community organizations who could refer warm leads. After two weeks, four calls — but all from one ethnic group.

v3: in addition to the partnerships, individual outreach by two community health workers who were members of the under-represented groups, paid for two hours per week. After two weeks, balanced recruitment.

Stop-rule applied: "Once recruitment from each target group reaches one per week for two consecutive weeks, v3 is locked and the rest of the recruitment runs as designed." The iteration was documented; the lock was clear; the study had defensible recruitment.

A standard non-iterating design would have pushed through with v1 and produced a "convenience sample" with the predictable demographic skew. The iteration cost time. It bought representativeness.

Iteration on consent

The consent form is the document that participants encounter first and that shapes their understanding of the study. It is also the document most likely to be written in legal-academic language and inflicted unchanged.

A practical iteration cycle:

1. Draft consent in plain language.
2. Read it aloud to three potential participants in their setting.
3. Note where they pause, ask questions, or reread.
4. Revise specifically those sections.
5. Re-read with three new participants; lock when no new confusion appears.

The investment is typically 4–6 hours of researcher time. The protection against misunderstanding-driven dropout is substantial.

Iteration on instruments

The instrument iteration cycle most teams know: cognitive interviewing for surveys, pilot interviews for guides. The transdisciplinary additions:

• Cognitive-interview the instrument with members of each sub-population, not just one
• Pilot interviews with two iterations: an early one to refine probes, a second to lock the guide
• Include practitioner or community partners in instrument review, not only researchers
• Document which items changed and why, for the audit trail

A locked instrument from cycle 1 is rare and worth flagging — it usually means the cycle missed something.

When not to iterate

Some studies have constraints that limit iteration:

• Funder timelines that don't allow for pre-data-collection iteration
• IRB processes that treat each instrument change as a protocol amendment
• Sites with limited tolerance for design changes
• Cross-site studies where consistency is paramount

In these cases, the iteration burden moves elsewhere: more extensive pre-protocol empathy work, more thorough literature scanning, more careful pilot studies before the main study locks. The point isn't to iterate forever; it's to choose deliberately when and where to iterate.

Iteration in team workflow

Iteration isn't only a design move; it's a team workflow. Practical patterns:

• Weekly 30-minute "iteration review" meetings during early data collection
• A logged backlog of design questions surfaced by team members
• Standing decision criteria for when an iteration becomes a protocol amendment
• A retrospective at the end of each phase: what worked, what didn't, what would we change?

These structures make iteration cheap. Without them, even researchers committed to iteration default to "we'll get to it later."

A retrospective discipline

Even when iteration didn't happen during a study, retrospective reflection has value: what should we have iterated on? Every researcher carries a list of "things I'd do differently next time" that don't make it into papers. Making the list explicit — and treating it as a personal iteration backlog — converts experience into design discipline.

Closing

Iterate on procedure as much as on instrument. Stop-rules turn iteration from drift into rigor. Consent, recruitment, and scheduling usually need more iteration than they get. When time constraints prevent iteration, move the burden to pre-protocol empathy and literature scanning. Treat team workflow as part of the design.

Next: Module 3 begins with reframing quantitative inquiry — numbers as one way of knowing, not the gold standard.