Jmis Theory Development

Theory Development (jmis-theory-development)

When to trigger

• Your claims are descriptive ("IT spend correlates with performance") with no economic or behavioral mechanism
• You built an artifact or ran a regression but cannot say what IS knowledge it advances
• A reviewer says "the theoretical contribution is unclear" or "this is a technical/empirical exercise"
• You need hypotheses, propositions, or signed predictions derived before you look at results

Theory takes a different shape in each JMIS style

JMIS spans empirical, economic-analytical, and design-oriented IS work, so "develop theory" means different things. Pick the row that matches your study; JMIS's distinctive pull is toward the economics and management of IS, so even behavioral papers benefit from naming the value or incentive at stake.

IT business value: make the mechanism, not the correlation, the theory

A JMIS IT-value paper is not "more IT, more performance." State which IT capability, acting through what organizational process, under which complementary conditions (process redesign, human capital, governance), produces which outcome. Theorize the complementarities and contingencies — JMIS reviewers reward a contingent value mechanism over a universal "IT pays" claim, because the field already knows the productivity-paradox debate.

Platform and economics-of-IS: derive signed predictions a priori

State the economic force in words before any math: who has what incentive, what information, and how does the IT-mediated market structure change behavior? Then write comparative-static predictions (e.g., "stronger same-side network effects raise the platform's optimal openness"). For an analytical model, the contribution is the insight the model yields, not the model's existence. Avoid HARKing — predictions precede estimation.

Behavioral IS: keep the artifact load-bearing

If your hypotheses would survive deleting the technology, they are not IS theory. Anchor the mechanism in a feature of the system (recommendation transparency, gamification, anonymity, automation) and theorize why that feature shifts cognition or behavior, and where the effect reverses. Generic TAM/UTAUT re-tests are a desk-reject risk; extend or challenge a mechanism instead.

Make boundary conditions and level of analysis explicit

IS effects are contingent on the artifact, the user, the firm, the market, and the period. Name the level (individual, firm, platform, market) and the scope where the theory holds and breaks — an honest scope condition beats an over-claimed law.

Worked vignette: from "AI improves productivity" to a JMIS mechanism (illustrative)

A draft claims "generative-AI tools raise analyst productivity." As stated it is a correlation with no IS theory. The JMIS mechanism names which capability of the tool (it lowers the cost of drafting but not of verification), acting through what organizational process (task reallocation toward judgment-heavy work), under which complementarity (it pays only where the firm redesigns review workflows and has the human capital to vet output), producing which contingent outcome (productivity rises for routine tasks, is flat or negative for novel ones where verification cost dominates). That contingent, complementarity-laden statement is theory; "AI helps" is not. The hypotheses then follow a priori, including where the effect reverses.

Referee pushback mapped to a theory fix

• "This is descriptive — where is the mechanism?" → State the IT-enabled force, the channel, the population, and the reversal condition before any estimation.
• "This is just TAM/UTAUT again." → Anchor the mechanism in a specific system feature and theorize why that feature shifts behavior; extend or contest a mechanism rather than re-confirm acceptance.
• "Your model has no insight beyond its setup." → Present the comparative static that surprises — the result the model yields that intuition would not predict.