Begin with a decision question
Research should help answer a real question: what is known, where is the information weak, and which priorities need further examination? A large folder of data is not a diagnosis.
The scope may combine official public records, existing client information, structured field observation and voluntary conversations. Each method has limits that belong in the findings.
What a useful research output contains
- A source register with dates, geography and definitions.
- An issue map that distinguishes observation from inference.
- A review of existing public work and organisational information.
- A list of unanswered questions and evidence gaps.
- A decision brief with recommendations and stated limitations.
Evidence before precision
Historic election data must be described with the correct year, constituency and units. Small or convenience samples should not be presented as representative opinion polls. The report should not manufacture vote estimates, percentages or causal claims to look authoritative.
A finding may challenge the original hypothesis. That is a useful result when it prevents resources being committed on the basis of a mistaken assumption.
Responsibilities and safeguards
The client supplies information it has the right to share and identifies access constraints. Research participants should understand the purpose and take part voluntarily. Collection should be proportionate to the assignment.
We do not build persuasion plans around caste, religion or other sensitive personal characteristics. Voter lists and private identity records should not be uploaded through the initial application. Scope, methods, timelines and access arrangements are agreed before research begins.
Questions about the work
Is this a guaranteed election forecast?
No. A constituency diagnosis identifies evidence, gaps and priorities; it does not create a reliable win probability.
Can research be commissioned separately?
Yes, subject to fit, conflict review, a defined question and agreed deliverables.