Bowling and colleagues (Reference Bowling, Sessa, Shaffer and Banks2026) argue that construct proliferation threatens cumulative knowledge in I-O psychology. I agree and believe this problem shows up across I-O, occupational health psychology (OHP), and much of the social sciences, where construct drift and near synonyms blur theory tests, weaken meta-analytic accumulation, and muddy conclusions about intervention effects. I use work–family research as an illustrative case because it is my substantive area and because it is a strong example (i.e., the literature is mature, interdisciplinary, and methodologically diverse, which makes both construct crowding and practical remedies easy to see). However, rather than a moratorium, I argue for construct stewardship and offer a three-prong framework that generalizes beyond work–family: (a) primary study standards that require domain maps and nearest-neighbor tests; (b) editorial routines that treat refinement and measure linking as the default path to contribution; and (c) training plus shared resources that lower the cost of integration. To make the third prong concrete, I highlight the utility of the Kanter Zotero Library, which is a curated Zotero collection of Rosabeth Moss Kanter Award nominated work–family articles assembled by the Center for Families at Purdue University (Center for Families at Purdue University, 2025).

Like many I-O areas (e.g., leadership, motivation, workplace aggression), work–family is especially vulnerable to construct crowding because it sits at the intersection of multiple disciplines, levels of analysis, and methods. The literature spans foundational constructs such as work–family conflict (Greenhaus & Beutell, Reference Greenhaus and Beutell1985) and has expanded to include positive interface concepts, boundary management, and dynamic approaches that capture within-person fluctuation and cross-domain processes (Allen & French, Reference Allen and French2023; Perry-Jenkins & Gerstel, Reference Perry-Jenkins and Gerstel2020). The field also has a long history of trying to clarify how work–family constructs relate to one another, including careful conceptual work that distinguishes constructs from mechanisms and outcomes (Edwards & Rothbard, Reference Edwards and Rothbard2000). More recent calls to put the family back into work–family research push the topic further into family processes and contexts (Kramer & Kramer, Reference Kramer and Kramer2021). This growth is welcome, but it also makes it easy to introduce new constructs faster than the field can evaluate their distinctiveness and usefulness.

Recent work–family research illustrates both progress and crowding. Episodic approaches show that work–family conflict and strain vary within people over time, and that conflict episodes can have near-term links with well-being (French & Allen, Reference French and Allen2020). Meta-analytic work on dimension-based work–family conflict helps adjudicate whether added distinctions among facets are theoretically and empirically useful (Hetrick et al., Reference Hetrick, Haynes, Clark and Sanders2024). At the same time, boundary management research provides a cautionary measurement example: Cobb and colleagues (Reference Cobb, Murphy, Thomas, Katz and Rudolph2022) identified dozens of boundary management scales, documented conceptual-operational misalignment, and noted limited evidence for convergent and discriminant validity across measures. Balance research offers a parallel pattern, where multiple meanings were clarified conceptually (Wayne et al., Reference Wayne, Butts, Casper and Allen2017) and then operationalized in new measures (Wayne et al., Reference Wayne, Vaziri and Casper2021) and longitudinal models (Wayne et al., Reference Wayne, Michel and Matthews2022). These are substantive advances, but they also show why stewardship needs to be explicit when constructs proliferate faster than direct comparisons.

Work–family provides an instructive case for how I-O and OHP can respond to construct proliferation. I propose a three-prong approach to construct stewardship and use work–family examples for each prong. Prong 1 focuses on primary research practices: Domain maps and nearest-neighbor tests should be routine for any proposed construct or measure. Prong 2 focuses on editorial practice: Journals can shift incentives by making refinement and linking the default path for contribution, not naming. Prong 3 focuses on training and shared tools: Curated bibliographies and construct-audit exercises lower the cost of integrative work and raise the cost of reinvention.

Prong 1: Require a domain map and a nearest-neighbor test. A domain map is a brief statement of where the proposed construct sits relative to established constructs and that measures represent the closest alternatives. A nearest-neighbor test then treats those alternatives as mandatory comparators. Because the closest alternatives are not always obvious from construct labels, Rosenbusch et al.’s (Reference Rosenbusch, Wanders and Pit2020) Semantic Scale Network can complement domain maps by using latent semantic analysis of item content to surface semantically similar existing measures as candidate comparators. This tool should be treated as screening support (not proof of uniqueness) and used to inspect item content across the nearest neighbors rather than relying on numeric cutoffs, given that the corpus will never be exhaustive. Requiring authors to identify and justify nearest-neighbor comparators is not a new expectation in social science. Strong review traditions often begin by mapping the construct space, naming overlapping labels, and clarifying what is in scope versus out of scope. Tepper’s (Reference Tepper2007) review of abusive supervision is a good example. It foregrounds the fragmentation problem, explicitly situates abusive supervision alongside adjacent constructs such as petty tyranny, supervisor aggression, supervisor undermining, and workplace bullying, and uses that mapping to justify what should be included in the evidence base and what should be excluded. This is the same logic Bowling and colleagues are pushing in their focal example. Progress comes from explicitly locating a construct in its neighborhood and then showing what changes when the closest alternatives are treated as required comparators.

For work–family, this means that a proposed boundary construct should be tested against established boundary and segmentation measures, and a proposed balance construct should be tested against the most widely used balance measures and the conflict and enrichment indicators that are often embedded in balance definitions (Wayne et al., Reference Wayne, Butts, Casper and Allen2017). Domain maps do not need to be exhaustive. They need to be candid about proximity, and they should make clear whether the contribution is a new mechanism, a new level of analysis, a new temporal process, or simply a new label.

The nearest-neighbor expectation should also apply to new temporal framings. Episodic perspectives on work–family conflict and strain (French & Allen, Reference French and Allen2020) make it possible to test mechanisms that are difficult to observe with cross-sectional designs, and they can yield more actionable implications for timing and target of interventions. Those advantages will cumulate faster if episodic variables are explicitly anchored to their nearest traditional counterparts and if papers show what changes once the nearest neighbors are in the model. In my view, the question is rarely whether a new label is interesting. The more useful question is whether the new framing changes our conclusions once the nearest neighbors are included in the model.

Prong 2: Make stewardship a routine editorial expectation. In my associate editor work in journals that publish work–family and occupational health research, I often see a consistent pattern in new-construct arguments. Authors present a new label, cite a narrow conceptual niche, and then demonstrate incremental prediction over a baseline model that does not include the closest alternatives. Stewardship requires stronger defaults. When I handle manuscripts that introduce a new work–family construct or measure, three requests typically clarify whether the contribution is real: (a) specify the nearest neighbors explicitly and justify the set of comparators, including, when feasible, item-content screening (e.g., the Semantic Scale Network; Rosenbusch et al., Reference Rosenbusch, Wanders and Pit2020); (b) test the proposed construct against those comparators using models that address shared variance and measurement quality; and (c) state what practical decision or theoretical inference changes if the new construct is retained versus subsumed under an existing construct family.

Editorial routines also matter for applied credibility. Work–family intervention research is growing, and syntheses now map interventions onto theory and summarize effects on outcomes such as conflict, enrichment, and balance (von Allmen et al., Reference von Allmen, Hirschi, Burmeister and Shockley2024). This expansion increases the need for construct stability, especially when intervention studies are pooled. As Livingston and colleagues (Reference Livingston, Pichler, Kossek, Thompson and Bodner2022) note, interventions can produce alpha change (true change in the target construct), beta change (change in the measurement scale), or gamma change (redefinition of the construct). The gamma change point is particularly relevant to construct proliferation: If our labels drift, we cannot easily tell whether interventions changed people, changed measures, or changed what the field means by the construct. Stewardship makes intervention evaluation more credible because it pins the outcome to a clearer conceptual and measurement anchor.

Prong 3: Use training and shared tools to make integration easier than reinvention. Stewardship is not only a reviewer problem, but it is also a socialization problem, and graduate training is where norms about what counts as a contribution get set. Work–family is a good training domain because the construct space is large but still coherent and because the field already has an unusually clear signal of “best work” through the Rosabeth Moss Kanter Award for Excellence in Work–Family Research. Building on that award pipeline, the Center for Families at Purdue University recently released the Kanter Zotero Library, a curated Zotero collection that compiles the peer-reviewed articles nominated for the Kanter Award across 2000–2025 (Center for Families at Purdue University, 2025). The value here is not just convenience. It is a vetted, high-impact corpus that is searchable, citable, and interdisciplinary, which makes construct auditing feasible for students and for research teams.

In practice, this can be built into both doctoral and master’s training. A basic assignment is to select a construct family (for example, balance or boundary management), sample articles (such as those from the Kanter library), extract operational definitions and measures, and produce a one-page domain map that flags overlap, missing discriminant evidence, and the most informative measurement options for future work. In a seminar format, that map can become a short in-class Q-sort activity, where students sort construct definitions or items into clusters, label the clusters, and then compare their sorting decisions to the published labels. That exercise makes overlap visible quickly and reveals when purported distinctions are not psychologically meaningful even to trained readers, which is a useful warning sign when evaluating construct claims. For doctoral students, the next step is to translate the audit into a study design plan that includes a nearest-neighbor comparison and a measurement linking strategy. For master’s students, the same exercise can be tied to evidence-based practice by asking what measure set would be defensible for program evaluation or organizational needs assessment. The broader point is simple. When we give trainees shared, curated tools and teach them how to use them, integration becomes the default workflow and “new construct” proposals become harder to justify without real incremental value.