Consistent scores and stable categories are separate questions

Reading the synthesis at the right level

Capraro and Capraro reviewed score consistency across studies using the Myers-Briggs Type Indicator. Their synthesis concerns the scales and research settings examined. It should not be treated as a single answer to every question about category boundaries, repeated type assignments, or the performance of an original preference questionnaire.

A boundary example without a type verdict

Consider two fictional respondents whose planning-preference values are forty-nine and fifty-one on a continuous response index. A system that draws a category line at fifty would give them different labels, even though their answers may be almost indistinguishable. If one person changes a single response next month, the label could switch while the underlying description remains much the same. The visible drama would come from the reporting rule rather than a demonstrated transformation of personality.

Machiavelli Preference Patterns therefore keeps its proposed dimensions continuous and does not assign a four-letter type. Even then, a difference of two points should not be narrated as a meaningful personal distinction without direct evidence. A participant may plan more extensively for shared work than for a solo task, or develop ideas through discussion only when the audience is trusted. Those variations are worth exploring rather than compressing into a permanent category.

What technical repeatability does establish

A deterministic scoring function returns the same result for the same answer set. That is an engineering property. It does not show that responses are consistent over time, that statements are understood similarly, or that an interpretation is useful. Keeping these layers separate matters when a polished dashboard makes the numbers appear authoritative. A report should explain the calculation while leaving empirical questions visibly open.

A more informative repeated study

An original study could ask participants to complete the questionnaire again after a specified interval and record whether their work setting changed. Analysis could examine each dimension, individual statements, and the effect of oppositely keyed wording. Participant interviews could investigate why a changed answer occurred rather than automatically treating every change as error. The report's language would also need testing: do readers understand a preference as revisable and contextual, or do they still infer a fixed type? A useful instrument must examine the consequences of its presentation as well as the repeatability of its numbers.

Continue the reading

Reliable scales do not automatically establish reliable types →