Construct creation from research questions

Abstract

Purpose Construct Creation (CC) is a methodological problem occurring when a research process, instead of measuring an extant construct in the participant’s mind, creates the construct. The purpose of this paper is to argue that CC derives from problems around ecologically invalid research and attitudinal responses developed on the spot, both resulting from self-generated validity. Design/methodology/approach A between-subjects design was used to explore whether the personification prime (PP), a component of brand personality (BP) methodology, influenced the CC of BP for rocks. Analysis of qualitative data on how participants made their BP ratings in the absence of a PP was also completed. Findings Findings revealed that a methodology can enable CC in the participant’s mind, despite the construct being ecologically invalid prior to them participating in the study. Analysis also revealed that participants will use varied, and sometimes elaborate, strategies to enable CC and provide researchers with the answers to their questions. Research limitations/implications Previous research has drawn attention to CC as a problem but the implications of prior research have so far been “sidestepped”. Consequently, this paper demonstrates CC and why it is a problem, while rebutting some arguments made in prior research for sidestepping CC. Practical implications CC is a potentially serious methodological problem that can result in invalid findings informing or misdirecting theory used by practitioners. As such, this paper proposes methods to ameliorate CC and improve ecological validity of future research. Originality/value This study will contribute to methodological literature by refocusing attention to the currently neglected problem of CC and by proposing a model of CC by participants.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 12 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.