ARCH9090 Chap.9 Triangulating and Interpreting Findings
Triangulating and Interpreting Findings
Triangulation compares evidence across methods, participant groups and contextual sources to strengthen or revise interpretation. This chapter shows how to preserve differences between questionnaire patterns, interview explanations and documentary context, then use agreement and contradiction productively.
Findings are written with calibrated confidence, visible sample limits and a clear account of how one source changes the meaning of another.
What this chapter covers
- 01
Triangulation across methods, participant groups and contextual sources
- 02
Questionnaire patterns, interview themes and documentary evidence as distinct products
- 03
Agreement that increases confidence without proving universal truth
- 04
Contradiction as evidence of group, timing, wording or power differences
- 05
Quotations as thematic illustration rather than prevalence evidence
- 06
Calibrated finding language that states support, limits and remaining uncertainty
Worked example · free
Integrate conflicting plaza-safety evidence
- 1Identify the questionnaire's daytime recruitment and broad perception measure. The marks assigned here are independent AskSia practice weighting and are not a published University mark allocation.
- 1Treat interviews as time-specific explanations rather than population prevalence.
- 1Explain why low reporting may not refute experience and why observation does not prove cause.
- 1State that perceived safety varies by time and group, with a credible concern near one exit.
- 1Recommend a bounded test and monitoring plan rather than a universal conclusion.
Key terms
- Convergence
- A pattern in which independent evidence streams support a compatible interpretation.
- Contradiction
- A meaningful difference between evidence streams that requires explanation rather than removal.
- Theme
- A recurring interpreted pattern in qualitative material supported by transparent examples and analysis.
- Inference boundary
- The limit on what a method, sample and context can defensibly support.
Triangulating and Interpreting Findings FAQ
Does triangulation mean three methods are required?
No. It means comparing evidence from different angles, which may include methods, participant groups or sources. The value lies in how the comparison tests interpretation. Collecting several data sets without integration does not create triangulation.
What should happen when sources disagree?
Check whether they reached different groups, asked different questions, refer to different times or were shaped by unequal power. Preserve a defensible difference when it matters. Forcing agreement can erase the very condition the engagement needs to explain.
How should confidence be expressed?
State which sources support the finding, which groups or contexts it covers, what complicates it and what remains uncertain. Use calibrated verbs such as suggests, indicates and was reported by. Precision makes a client recommendation more usable, not less persuasive.
Why keep negative cases during analysis?
A response that resists the dominant theme may reveal a boundary condition, a missing subgroup or a category that does not fit participant experience. Examine it rather than treating it as noise. The report need not give one case the same weight as a repeated pattern, but explaining why it differs can improve the finding and identify where a recommendation should be targeted or tested before wider use.
Label conclusions as convergent, complementary or unresolved so the client can see whether confidence comes from agreement, added context or explicit uncertainty.
Assessment move
Create a triangulation matrix organised by the project questions or themes, not by method. Give questionnaire patterns, interview explanations, contextual sources, group differences and uncertainty separate columns. For each row, write whether evidence converges, complements or contradicts, then explain why.
Practise converting overclaims into bounded findings: residents believe becomes respondents reported; the project caused becomes participants linked; everyone wants becomes a preference among the reached group. Pair every quotation with the theme it illustrates and every chart with recruitment and measure context. Before writing a recommendation, ask whether the finding supports an immediate action, a pilot or further inquiry.
That distinction protects the client from false certainty and shows reflective control of the evidence. Add a row for negative or deviant cases so evidence that resists the dominant theme remains visible. For every contradiction, test four explanations: different recruitment, different question meaning, different timing and different freedom to speak.
Do not select one explanation without evidence; state which is plausible and what further check would distinguish them. Write the same finding at three confidence levels and choose the version that matches the design. Then append the client consequence, showing why this interpretation matters to an option or action.
Review charts and quotations as a pair: the chart may show a pattern while the quotation explains one experience, but neither should be asked to perform the other's role. End the matrix with an inference boundary covering the reached public, place, period and decision. This single sentence keeps synthesis useful without turning it into false consensus.