Some questions resist the first answer because the first answer is too eager. I returned to one question three times, each time with a different kind of material: a statistic, a conversation, and a long silence in the record.
First return: the visible measure
The number was useful because it set a boundary. It did not explain the experience, but it prevented the discussion from becoming entirely atmospheric. A measure can be a beginning even when it is not an explanation.
What the number leaves out
Measurements flatten sequence. They tell us how much, not always when, for whom, or after what. The missing relations were the reason to return.
Second return: the situated account
A conversation restored sequence. The same event appeared differently when described from inside a day rather than from above a dataset. The account also introduced uncertainty, which was more honest than a neat conclusion.
Third return: the unresolved remainder
The silence was not empty. It marked the point where the available material stopped. Keeping that remainder in the archive is part of the research: a later entry can return without pretending the question was already settled.
More research entries collect the returns rather than only the answers.