The same field can take different types under different analytical tasks

What it is

Task-dependent field typing means that one stored field may serve different analytical roles in different questions. Year can identify cohorts, order events, or support interval calculation under appropriate calendar assumptions. A satisfaction score can be a response category, an ordered level, or an approximately quantitative summary when a measurement model permits it. The chosen type must be recorded with its task and assumptions rather than permanently attached to a column name.

Why it happens

Data type describes relations used in the current inference, not the appearance of characters. A postal code is written with digits but primarily identifies an area; age is quantitative but can be grouped for an eligibility decision; a timestamp can identify an event, order it, or bound a duration. Aggregation, binning, ordering, and derivation create new field roles, but do not automatically make all arithmetic on the new representation meaningful.

Studying it

Choose one field and define classification, ordering, difference, trend, and threshold tasks. Compare answer quality under alternative type declarations and encodings. Ask participants to explain permissible comparisons to reveal whether field semantics or graphic suggestion drives the answer. Record derivation, unit, binning, and aggregation, and test whether legends, caches, or filters from one task leak into another.

Where it stops holding

“Can take different types” does not license arbitrary redefinition. Transformations must preserve provenance, units, and domain constraints. An identifier does not acquire interval meaning merely because it can be sorted, and averaging ordered item responses does not by itself establish a ratio scale. Sampling, measurement error, and privacy purpose further constrain cross-task reuse.

Applying it

• Record type at the field-by-task-by-view level, not only at the data-dictionary column level; preserve both source and derived types.
• Give each binning, aggregation, or unit transform a named derived field with thresholds, formula, and version.
• Recheck axes, color scales, summaries, and accessible descriptions when the task changes; do not inherit order or magnitude implications from the previous task.
• Let users inspect and correct the current role while blocking computations and comparisons unsupported by the measurement assumptions.

Related

• Same group: U1.08.1 Nominal, ordinal, and quantitative data need different encoding channels · U1.08.2 Encoding nominal data with a quantitative channel implies a nonexistent order · U1.08.3 Encoding quantitative data with a nominal channel discards magnitude differences · U1.08.4 Determine data type before choosing a chart type
• Adjacent: U2.01.3 The same data supports several tasks; draw a chart for each · U2.01.2 Define the question before choosing the chart
• Search terms: task-dependent data type · field role · measurement assumptions