Reliable scales do not automatically establish reliable types

The substantive question

Capraro and Capraro’s reliability synthesis reported generally strong estimates for MBTI scales alongside variation across studies. A later review of Form M examined a broader psychometric literature. These findings should be read at the level of the specific forms, samples, and estimates investigated. They cannot be transferred to MPP, and scale reliability alone does not settle whether type categories are valid.

For MPP, reliability work should examine each original dimension and the effect of context. A person’s preferred way to develop an idea may vary with familiarity, audience, and time pressure. Advance structuring may vary with whether a task is shared. Those influences can be substantively meaningful rather than random noise. The study must specify whether the score is intended to summarize a broad preference or a particular setting. It should also inspect whether positively and oppositely keyed items behave as intended. A smooth interface and deterministic score calculation provide technical consistency, not measurement evidence.

A worked interpretation

A reader might report more outward processing after joining a collaborative project. The change could reflect a new preference, more opportunities to talk, or a different reference situation. A retest number alone cannot distinguish these explanations.

If a future version offers uncertainty ranges, the ranges should reflect data from that version and a relevant population. Until then, the report should avoid declaring small changes meaningful or interpreting a shift across the midpoint as a new identity. The reader can still examine what changed in the context and whether their preferred approach is serving their chosen purpose.

A question worth testing

A retest protocol could ask respondents to use the same chosen context, then compare that result with a broader instruction in a separate study. The distinction would clarify whether MPP describes context-specific preferences or more general tendencies. Neither design is automatically superior. The important requirement is that the report’s claim matches the administration frame and that users are not encouraged to generalize beyond what was studied.

How we would evaluate MPP

Reliability work for MPP would retain continuous scores throughout analysis. Researchers should examine consistency of each preference summary and how it changes across administrations, rather than first dividing participants into categories and treating a changed label as the primary outcome. A person near any arbitrary dividing point can cross it after a very small response change. MPP avoids that specific reporting problem by assigning no type, but continuous scores still require evidence about uncertainty and stability.

A retest study should ask whether the participant imagined the same kind of decision. Advance structuring may look different for travel, collaborative work, and an unstructured weekend. Criteria-led deliberation may differ when comparing familiar options and making an unfamiliar choice with incomplete information. Researchers would record those contexts and test whether a general preference summary is defensible. They should also examine whether reading the first report changes how participants understand the later questions. That would be relevant to the response process even if the numerical score becomes more stable. Until these studies exist, MPP cannot identify a fixed cognitive style, claim a reliable change threshold, or promise that a profile will remain constant across roles and circumstances.