RESEARCH DOSSIER
Creative Evaluation, Iteration, and Revision
Abstract. Creative evaluation, iteration, and revision are the interdependent processes through which candidate ideas become selected, transformed, and tested creative artifacts. This critical narrative research synthesis distinguishes originality and usefulness appraisal from selection, forecasting, constructive criticism, and material revision. It analyzes classic consensual assessment, laboratory manipulations of revision standards, research on idea-selection bias, longitudinal design-process traces, organizational forecasting, creativity training, and contemporary neurocognitive and human–artificial studies. The reviewed evidence supports the importance of context-dependent criteria, distinct evaluative capacities, feedback conditions, and artifact-mediated cycles; it also identifies persistent limitations in longitudinal measurement, external validity, judge agreement, and causal identification. The research frames evaluation as a measurable series of judgments and interventions that can strengthen or prematurely constrain creative production.
Keywords: creative evaluation; idea selection; creative iteration; revision; creative forecasting; originality; usefulness; consensual assessment; creative feedback; design cognition; creative metacognition; human–artificial creativity
The Scientific Object: Evaluation, Iteration, and Revision
Creative evaluation is the assessment of candidate ideas or artifacts against criteria relevant to originality, appropriateness, feasibility, quality, meaning, and anticipated consequences. Iteration is the repeated cycle of producing a version, observing its properties, comparing those properties with changing criteria, and constructing a subsequent version. Revision denotes a material or conceptual change to a candidate, including the alteration of constraints, representation, composition, argument, or implementation. Together these processes explain how creative work proceeds beyond the mere production of possibilities. A research account must specify what is evaluated, which standards are used, who applies them, and whether the judgment changes subsequent work. The influential analysis by Runco & Jaeger (2012) treats originality and effectiveness as central dimensions of creativity. These dimensions guide measurement, yet they are not automatically interchangeable across scientific, artistic, technical, educational, and organizational domains.
The unit of analysis shifts over the creative process. In a divergent-thinking task it may be a short verbal response; in architectural design it may be a sequence of sketches; in scientific work it may be a hypothesis that survives critical experiments; in a collaborative production it may be a succession of collectively edited prototypes. Evaluation can happen during generation, after generation, in response to an external critique, or retrospectively when a finished artifact is judged by an audience. Selection is the commitment to one or more alternatives, whereas forecasting predicts consequences or future reception. Rietzschel et al. (2026) identifies evaluation and selection as distinct research tasks and documents the diversity of methods used to measure them. Consequently, a high originality rating, a selection decision, a successful revision, and a later achievement are different empirical outcomes.
The central explanatory question is how an evolving artifact and the creator's criteria alter each other across time. New knowledge about a proposal changes the standards by which it is judged; adopting a standard can prompt a new version that exposes previously unnoticed constraints. The process therefore incorporates feedback from materials, collaborators, audiences, and changing task environments. A rigorous model describes an ordered series of observable states and actions rather than treating creative achievement as a fixed individual trait. The chapters distinguish established experimental results from process interpretations and proposals for future study. They locate creative evaluation within the broader psychology of aesthetics and creativity while maintaining a specific focus on appraisal, decision, and transformation.
Definitions, Constructs, and Boundaries
Evaluation refers to graded judgments about the qualities of an option. A person may consider an idea original but impractical, technically adequate but uninteresting, or useful for one audience but misaligned with another. Selection is the acceptance, rejection, or ranking of options for further work, and its relationship to quality depends on the decision rule and available information. Revision is an action upon the current proposal that changes one or more properties; a revision may improve one dimension while weakening another. Iteration describes the temporal organization of such cycles. The number of revisions, degree of change, and quality of the final artifact must be measured separately because more changes do not necessarily imply greater creative achievement. The distinction between evaluative judgment and selection performance follows Rietzschel et al. (2026), while the originality–effectiveness framework follows Runco & Jaeger (2012).
Creative forecasting occupies a special position. The forecaster attempts to anticipate how a future audience, user, market, or empirical test will respond to an idea whose consequences are partly unknown. Forecasting accuracy requires comparing a prediction recorded before exposure with a later observable outcome, preferably under conditions that limit selective reporting. The person who produces an idea may know its technical details but have privileged motives or attachments that bias prediction; conversely, an external evaluator may underestimate the practical possibilities of an unfamiliar design. Berg (2016) makes forecasting an object of organizational field and laboratory investigation. Confidence in a forecast is a distinct construct and should be calibrated against success rates rather than interpreted as objective validity.
Several neighboring research objects have overlapping vocabulary but separate endpoints. Divergent thinking concerns the production of multiple ideas, insight concerns reorganization in problem solving, neurocreativity research examines neural processes, and consensual assessment concerns the measurement of artifact creativity. The present monograph asks how available ideas are valued, selected, altered, revalued, and stabilized. Its evidence includes, but is not confined to, the studies of generating alternatives. Creative evaluation may accompany aesthetic pleasure, organizational screening, moral reflection, or experimental testing, yet the scientific outcome must remain explicit. This boundary also prevents interpreting peer approval as proof of originality or a favorable self-rating as evidence of successful real-world implementation.
Historical Development of the Evaluation Problem
Research on creativity developed from broad accounts of talent and artistic production toward operational studies of thinking, judgment, and observable work. The twentieth-century expansion of divergent-thinking tests increased attention to ideational fluency and originality, and later investigators examined the subsequent fate of ideas. Experimental social psychology contributed methods for controlling assessment conditions and tracing how contextual factors shape creative performance. Amabile (1982) introduced an influential consensual assessment approach in which informed judges assess creative products, helping researchers obtain domain-sensitive measures of outcomes. This methodological step made evaluation itself visible while leaving open a crucial issue: agreement about a completed product does not reveal the internal sequence by which it was revised.
Studies of designing offered another historical line. In detailed observations of architectural sketching, Goldschmidt (1991) described visual reasoning in which representations help elicit new interpretations and guide subsequent transformations. Drawing supplies information to its maker; the designer perceives the materialized mark and reconceives the problem, often without a discrete generation stage being followed by a single final judgment. This tradition emphasized the artifact as an evolving participant in reasoning. Its empirical strength lies in close access to episodes of transformation, although a concentrated case or professional setting warrants cautious generalization. The historical lesson is that creative judgment can be embedded in the act of making rather than appended only after a solution appears.
By the early twenty-first century, a series of experiments explicitly manipulated the standards applied to candidate ideas. Lonergan et al. (2004) examined how appraisal and revision criteria interact with initial originality, and Rietzschel et al. (2010) investigated the difficulty of choosing strong ideas after individual ideation. Later meta-analytic and methodological work demonstrated that evaluative skill and generative performance cannot be assumed to coincide. In Guo et al. (2022) the pooled association between divergent thinking and evaluative skill was small, while Rietzschel et al. (2026) organized otherwise heterogeneous performance tasks. Historical progress thus consists in sharper separation of generation, evaluation, decision, and implementation, together with renewed attention to the sequences that connect them. The dual-pathway model of Nijstad et al. (2010) further describes flexibility and persistence as pathways to creative ideation; those generative mechanisms help explain the supply of options that evaluation and revision subsequently transform, while remaining distinct from evidence about selection accuracy. Earlier research by Runco & Basadur (1993) distinguished ideational and evaluative skills in a multiphase approach to creative problem solving. A field-based proposal by Basadur (1995) also investigated variation in preferred ideation–evaluation ratios across organizational work types; its reported preliminary observations should not be interpreted as causal evidence that one ratio maximizes innovation.
Research Questions and the Logic of Evidence
The first research question concerns criterion construction: how do creators decide what counts as worthwhile when originality and suitability pull in different directions? The second concerns selective attention: which properties of an idea receive sufficient scrutiny to affect a decision? The third concerns revision mechanisms: under what conditions does criticism improve an artifact rather than merely reproduce conventional solutions? The fourth concerns temporal organization: how does an early evaluation influence later exploration, commitment, and abandonment? The fifth concerns expertise and context: when do disciplinary knowledge, audience expectations, and social power alter evaluative performance? These questions require observations at more than one point in the creative process. Rietzschel et al. (2026) underscores the importance of distinguishing measurement tasks before interpreting the evidence.
Different study designs answer different parts of these questions. Randomized manipulation of evaluation standards can isolate a short-term causal effect under laboratory conditions. Repeated revisions of a traceable artifact can reveal when quality changes, but observational associations between critique and improvement remain vulnerable to selection and time-varying confounding. Expert ratings can quantify product reception within a population of judges, but their validity depends on who the judges are and which judgment is requested. Organizational field outcomes may be ecologically meaningful while mixing creative quality with resources, visibility, institutional support, and timing. The combination of experimental, observational, psychometric, and qualitative methods is therefore necessary for scientific synthesis.
This monograph uses a critical narrative review of identifiable original studies, peer-reviewed theory and secondary quantitative syntheses. It reports the design and inferential scope of key studies, highlights disagreement, and proposes testable extensions when existing evidence does not determine an answer. It was not conducted as a preregistered systematic review; no exhaustive search claim, pooled estimate, or original participant dataset is asserted. Numerical effects are cited only when reported in identifiable source records. The conceptual framework is intended to connect discrete judgments with sequences of revision while leaving substantive claims dependent on the cited empirical evidence.
Criteria, Trade-offs, and Multiattribute Appraisal
Creative judgments ordinarily integrate multiple attributes that may not improve together. Originality concerns departure from familiar alternatives; usefulness, adequacy, and appropriateness refer to a proposal's relation to a specified purpose and constraint set. A solution may be highly surprising yet unable to operate under current conditions, or immediately usable but conceptually routine. Runco & Jaeger (2012) establishes the definitional importance of originality and effectiveness, while Lonergan et al. (2004) shows why the standards used in appraisal matter to subsequent revisions. Criteria can be represented as separate ratings or as an explicitly stated decision rule. A single overall creativity score obscures whether an evaluator rewards novelty, feasibility, aesthetic appeal, or familiarity.
The evaluative process has a temporal component: an initial plausibility judgment can determine which proposals obtain resources for testing. Such screening changes the distribution of observable outcomes. Rejected ideas seldom receive a comparable chance to demonstrate future usefulness, and selected ideas may become valuable through additional development rather than through their initial properties alone. The observed success of selected proposals therefore cannot be interpreted as direct proof that the original screening procedure identified the best options. To estimate selection quality, researchers must preserve an adequately sampled set of rejected ideas, independent quality ratings, and, where feasible, controlled development opportunities. These design implications emerge from the selection difficulties investigated by Rietzschel et al. (2010) and the methodological review of Rietzschel et al. (2026).
Criteria also carry values and institutional assumptions. A teacher may emphasize learning progress, a theater director may emphasize audience experience, and an engineer may prioritize safety and manufacturability. Each has a legitimate criterion configuration if it is transparent about the intended outcome; cross-domain rankings require new evidence. Evaluation research gains explanatory precision when it measures the weights attached to attributes, the person's understanding of those attributes, and changes in weight over time. The corresponding behavioral evidence includes ratings, forced-choice decisions, explanations, dwell time, and subsequent edits. A theory of creative judgment must account for disagreement attributable to preference, measurement error, expertise, and genuine variation in context.
From Generating Options to Selecting a Candidate
Producing a large pool of alternatives creates an opportunity to choose well; it also imposes a demanding comparison problem. Evaluators may rely on immediate appeal, ease of imagining implementation, similarity to familiar practices, or anticipated acceptance. Rietzschel et al. (2010) studied idea selection following individual idea generation and found that generating many ideas does not guarantee that the most creative candidates will be selected. The study highlights a mismatch between the ability to produce creative ideas and the ability to identify them for continuation. Since selection is a choice among available options, quality must be judged in relation to the entire candidate set, not merely the selected candidate in isolation.
A selection decision can be studied as a ranking, a binary continuation choice, a portfolio allocation, or a decision with an explicit resource budget. These procedures are not statistically equivalent. Ranking forces relative comparison; independent ratings allow more nuanced differentiation; resource allocation adds costs and opportunity constraints; continuation decisions may be strongly affected by the number of options allowed to survive. Rietzschel et al. (2026) reviewed the measurement landscape and argues for more coherent operationalization. Experiments should define how candidate quality was established independently, whether the selector previously generated the options, and whether a decision was made before or after information about practical constraints was supplied.
The process also exhibits path dependence. Once a creator chooses a candidate, subsequent effort can increase familiarity with it and raise the psychological cost of abandoning it. Additional expertise can improve anticipation of practical difficulties but can also make common solutions more readily accessible. The competing mechanisms suggest a research design with staged evaluations: record initial rankings, randomize feedback or further information, permit bounded revisions, and then compare changed rankings with later outcomes. Such a design separates a judgment about the starting idea from an assessment of the revised artifact. It also tests whether quality increases through learning about the candidate or merely through repeated exposure.
Evaluative Skill and Its Association With Idea Generation
Evaluative skill refers to the capacity to discriminate among ideas according to specified dimensions such as originality or creativity, often using agreement with expert judgments or other target criteria as the benchmark. A person may generate unusually original responses yet be poorly calibrated when ranking the relative merit of those responses. Conversely, someone with modest ideational fluency may be adept at recognizing the promise of options generated by others. Guo et al. (2022) synthesized 96 effect sizes from 20 studies, comprising 3,019 participants, and reported a pooled correlation of r = .13, with a 95% confidence interval from .07 to .20, between divergent thinking and evaluative skill. The association is positive but small, supporting partial separability rather than independence or equivalence.
The meta-analysis also identified moderation by divergent-thinking test type, evaluation task, and the origin of the ideas under evaluation. In particular, assessing one's own ideas is not necessarily psychometrically equivalent to assessing ideas supplied by somebody else. The reported heterogeneity complicates interpretation of an isolated creativity-test score. When comparing groups or training methods, investigators should use multiple evaluation formats and report test content, reference standard, reliability, and susceptibility to familiarity effects. A correlation between scores also cannot demonstrate that strengthening generation will improve selection, or the reverse. Longitudinal or intervention studies with distinct outcome measures are required to investigate such causal transfer. Guo et al. (2022) provides evidence for construct relatedness, not a general causal mechanism.
Even the benchmark against which an individual is scored deserves scrutiny. Agreement with one panel may measure social conformity to local tastes as well as sensitivity to useful originality. A panel that is homogeneous in professional training may undervalue culturally different forms. Independent judge sampling, interrater reliability, explicit criteria, and validation against later outcomes help clarify what is being measured. A well-designed study may treat evaluative sensitivity as a multidimensional profile: discrimination of novelty, discrimination of appropriateness, calibration of confidence, and ability to identify revision potential. That profile allows informative analysis of divergent evaluators rather than reducing all disagreement to error.
Creative Forecasting and the Uncertain Future of Ideas
Forecasting asks an evaluator to predict how a proposal will perform in a future context. This differs from judging its immediate appeal because the outcome depends on execution, audience response, competitive alternatives, and changing constraints. Berg (2016) examined creative forecasting among 339 circus professionals and compared predictions with evaluations from 13,248 audience members, adding a laboratory experiment. In that setting, creators were more accurate than managers when predicting responses to other people's ideas, while their advantage did not extend to forecasting their own ideas. These findings illustrate the importance of role, psychological distance, and the difference between being able to create and being able to anticipate reception.
The inferential scope is specific. The field study draws strength from professional practice and external audiences; the underlying causal reasons cannot be established by role comparisons alone. The accompanying laboratory experiment addresses mechanisms under manipulated conditions, but a theater-related success criterion may not transfer directly to scientific discovery, software design, or public policy. A broader measurement program should register each prediction before outcomes are revealed, use accuracy or calibration statistics suited to the outcome, and compare predictors against transparent base rates. Confidence and vividness should be measured as separate variables. Berg (2016) supplies a rigorous example of an empirically assessable forecasting problem rather than a universal ranking of creators and managers.
Creative forecasting also generates self-fulfilling and self-defeating dynamics: organizational decisions about funding, visibility, and training can influence the very outcomes they are expected to predict. A proposed intervention could randomize resource allocation among matched ideas after recording preallocation forecasts, then assess how initial promise and later support jointly affect outcomes. Without such controls, an apparently prescient selector may simply occupy a position that determines which proposals are allowed to prosper. The practical lesson for inquiry is to locate judgment within an observable decision system, with appropriate attention to the unequal opportunities that follow from selection.
Iteration as a Temporal Research Process
An iterative creative process consists of successive, linked versions of a developing solution. The relevant observations are the state of the artifact at each point, the evaluator's perceived shortcomings or opportunities, the operation performed, and the judgment that follows. Iteration can involve incremental refinement, conceptual reconception, removal of elements, deliberate exploration of competing designs, or a return to an earlier version. Treating a final product as the only dependent variable hides these transitions. Goldschmidt (1991) described a process in which successive sketches provoke new interpretations, illustrating why representation and thought evolve in interaction. A trace of edits is useful evidence, but a count of edits alone cannot establish improvement or creativity.
Sequential research should distinguish opportunity to revise from actual revision and revision from independent improvement. For example, a participant may receive criticism but leave a proposal unchanged because the critique is irrelevant; another may produce extensive changes without affecting suitability; a third may simplify the design dramatically, producing a stronger outcome with few edits. Independent and blinded rating of each version can track outcome change, while timestamped logs locate turning points. Researchers should predefine the unit of revision and preserve rejected versions to avoid hindsight bias. Changes in task goals are especially important because a quality score may rise only because the criterion is revised to suit the emerging artifact. The standards examined experimentally by Lonergan et al. (2004) help explain how evaluation can guide transformation.
An adequate temporal model also considers delay and incubation as contextual processes without confusing them with iteration. Reflection between sessions, exposure to new examples, collaborator feedback, and testing can change an evaluator's information. Yet the duration between versions has no inherent causal interpretation: longer intervals may reflect complexity, fatigue, deliberate search, or organizational blockage. Event-based measurement should record time, intermediate evidence and decisions, then analyze how such factors relate to subsequent revisions. When the same people generate and evaluate their work, within-person longitudinal models are particularly useful because stable differences in expertise or standards may otherwise masquerade as effects of specific episodes.
Appraisal Standards and Experimental Revision Effects
A landmark experiment by Lonergan et al. (2004) asked 148 undergraduates to appraise and revise advertising-campaign ideas varying in initial originality while working under different standards. The study distinguished generative or innovative criteria from criteria emphasizing operational efficiency. Better plans emerged when innovative criteria were applied to initially less original ideas and operating criteria were applied to more original ideas. This interaction supports a conditional account: the value of a revision standard depends on the deficiencies and developmental opportunities of the starting idea. It challenges the assumption that an instruction to maximize novelty or feasibility is equally beneficial at every moment.
The study is important because it manipulated standards rather than merely measuring a preferred cognitive style. Its task and population nevertheless restrict inference: student-produced advertising plans may respond differently from expert engineering prototypes or creative writing developed over months. A revision manipulation can change both attention and motivation, and outcomes depend on how independent judges define plan quality. A direct replication should preregister which characteristics of starting ideas make a standard appropriate, randomize standard assignment, document the full edit sequence, and compare multiple domains. This makes the experiment a foundation for targeted process questions rather than evidence that one generic rule always maximizes creativity. Lonergan et al. (2004) is therefore best read as evidence for an interaction among initial quality, standards, and subsequent revision.
Selection and revision also create an important decision paradox. Rejecting an idea too early prevents discovery of what it might become; committing to it uncritically can consume scarce effort. The scientific variable is the opportunity value of further development relative to competing alternatives and the quality of new information expected from a test. A rational-seeming choice need not be psychologically realized as a formal utility calculation. Researchers can approximate the problem with experimentally assigned development budgets, recorded forecasts, and independent assessment of both initial ideas and revised artifacts. The core question is whether a decision procedure identifies which ideas benefit most from continued work.
Sketching, Prototyping, and the Material Feedback Loop
Creative judgment often takes place through engagement with an external representation. A drawing exposes spatial relations, a draft exposes the pace and implications of language, and a working prototype exposes interactions that could not be evaluated fully in imagination. Goldschmidt (1991) documented the shifting figurative and conceptual interpretations involved in architectural sketching. An artifact is therefore both the object being evaluated and a source of new evidence for the evaluator. The act of materialization can reveal inconsistency, stimulate alternative interpretations, and make constraints available for inspection. Such findings connect creativity to research on distributed and embodied cognition while remaining grounded in observable acts of design.
Process tracing must distinguish creator testimony from independently visible change. The designer's commentary may provide evidence about perceived goals but may omit tacit decisions or rationalize a sequence retrospectively. Video, version-control records, sketches, timestamps, tool interactions, eye movements, and coded verbal protocols each sample different aspects of activity. Triangulation improves the ability to reconstruct how an earlier representation influenced a later modification. Yet close observation alone does not identify whether feedback caused an improvement: more accomplished designers may both revise strategically and produce better final artifacts. Experimental task perturbation, randomized information, or matched comparisons of alternative feedback mechanisms can strengthen causal interpretation.
The concept of a prototype is domain-relative. A rough visual sketch may capture proportions while suppressing materials; an executable code model may expose performance while omitting social usability; an early narrative draft may indicate structure but conceal effects that appear only to an external reader. Evaluation should therefore specify which properties are represented reliably by each artifact version. Failures that appear under one test may be artifacts of the representation rather than faults in the underlying idea. An evidence-based revision cycle maintains traceability between the original problem, the evaluator's current model, the prototype's measured behavior, and the decision about what to change next.
Feedback, Criticism, and the Conditions of Productive Revision
Feedback communicates information about performance, criteria, or potential improvements. Its effects depend on content, delivery, authority, perceived controllability, and the creator's opportunity to act. In a laboratory study with 210 participants completing a role-playing task, Zhou (1998) examined feedback valence, feedback style, and task autonomy. Creative performance was highest in the combination of positive, informational feedback and high autonomy. The finding supports studying feedback as an interaction between message properties and the surrounding work environment. The task measured creative performance rather than a fully observed long-term revision trajectory, so its results should not be reported as direct evidence that every positive critique improves every draft.
A constructive critique identifies a feature and explains its relevance to a goal, giving the recipient usable evidence or a question that can be tested. Global approval and global disapproval may influence motivation yet offer little guidance about the next edit. A creator also needs enough autonomy to interpret feedback rather than merely comply with a command. Evaluation becomes useful when it can be translated into a decision: preserve an element, vary it, test it, remove it, or redefine a constraint. Lonergan et al. (2004) suggests that the same instruction can help or hinder depending on the initial proposal. Consequently, instructional design should match the critique to the current stage and observed properties of the artifact.
Feedback research requires careful counterfactual design. Praise may be given to inherently stronger ideas; heavily criticized products may be those with the most defects and consequently the most room for improvement. Comparing recipients without controlling starting quality confounds the effect of the feedback with the reasons it was given. Experiments should manipulate informative content separately from tone, measure autonomy and prior confidence, blind final judges to condition, and follow actual revisions. Qualitative evidence about whether the recipient understood the critique can explain null effects, but it cannot by itself substitute for independent outcome evaluation. Ethical handling of criticism is especially important in classrooms and workplaces where formal hierarchy affects willingness to dissent.
Expertise, Domain Knowledge, and Evaluation Standards
Domain expertise supplies the concepts, precedents, constraints, and technical expectations used to evaluate an emerging idea. It can make defects visible that a novice overlooks, and it can support efficient estimation of implementation costs. Expertise also affects what counts as original: a proposal may seem unusual to a novice because the novice has not encountered the relevant tradition. Amabile (1982) developed consensual assessment around the judgments of people familiar with a domain, providing a practical route to evaluating creative products without requiring one universal checklist. Such judgments are informative within defined conditions, but the reliability and composition of the judging population must be established rather than assumed.
Greater familiarity may make a proposal easier to understand, while genuinely unconventional work may initially challenge established standards. The resulting tension is empirical. Researchers should compare judges across levels and kinds of expertise, document their access to domain information, and distinguish ratings of originality, technical correctness, utility, and personal preference. When a work crosses fields, multiple expert communities may supply different legitimate assessments. A software interface, for example, can be technically innovative while inaccessible to intended users. Resolving the discrepancy requires explicit attention to task objectives, user experience, and later outcomes, not an appeal to expert status alone. Rietzschel et al. (2026) emphasizes that measurement tasks must identify the standard against which performance is assessed.
Expertise can also change the sequence of revision. A specialist may identify a high-leverage modification early, while a novice explores many superficial variants. Alternatively, the novice's unfamiliarity with conventions may encourage an overlooked possibility. Strong designs for studying these questions match participants on baseline output quality, manipulate access to domain feedback, and preserve intermediate versions. A final-quality comparison without process data cannot separate superior initial ideation from more effective revision. The scientific aim is to specify which forms of knowledge are active at each decision point and how they alter the quality of subsequent changes.
Educational Environments and the Development of Evaluative Capacity
Creative education often emphasizes producing numerous ideas, but a learner also needs experience explaining why one proposal is promising, how to test it, and how to improve a revision. These are potentially teachable capacities whose transfer requires empirical demonstration. Guo et al. (2022) found only a small association between divergent thinking and evaluative skill, cautioning against treating improved ideational fluency as automatic evidence of stronger judgment. Teaching evaluation can involve contrasting cases, explicit criteria, peer assessment, and documented revision, each of which isolates a different component. The meaningful endpoint is not merely student satisfaction; it is improved discrimination, revised artifact quality, and application to subsequent problems.
A critical review of five decades of training research by Sio & Lortie-Forgues (2024) evaluates the impact of creativity instruction on performance across varied interventions and outcomes. Such synthesis supports treating creativity training as an evidence question rather than assuming a uniform intervention effect. A course focused on revision should include a control group that receives comparable practice and attention without the specific evaluative procedure, blinded assessment of products, delayed follow-up, and tasks in which learners must choose among unfamiliar options. Measuring both self-generated and externally generated ideas would clarify whether learners acquire general evaluative sensitivity or only greater familiarity with their own work. A focused educational review by van Broekhoven (2025) argues that students' evaluation and selection of creative proposals require explicit attention in learning environments. The review also cautions that student-centered pedagogy by itself does not establish improved creative selection.
Instructional feedback must also protect the exploratory phase. Students need criteria they can understand without being pushed to select the first conventionally acceptable solution. One experimental strategy is to compare early and delayed feedback, coupled with freedom to retain multiple provisional ideas. Another is to separate originality-focused and usefulness-focused revision rounds and examine their interaction with starting quality, building on Lonergan et al. (2004). Developmental conclusions should be bounded by the ages, school contexts, and tasks sampled; a classroom improvement under a rubric cannot be presumed to predict creative achievement decades later. Well-designed evaluation instruction makes the relationship between criteria, evidence, and revision observable.
Organizations, Group Deliberation, and Selection Institutions
Organizations distribute creative evaluation among originators, managers, users, peers, and funding bodies. These participants encounter different information and face different incentives. Berg (2016) demonstrates that forecasting success depends in part on whether evaluators operate in creative or managerial roles, and that predicting one's own ideas differs from predicting others' work. In an organization, an idea may be recognized as promising yet never reach implementation because of budgeting, authority, or coordination constraints. Evaluating the quality of the initial idea, the selection decision, and the later organizational outcome therefore requires separate measurements and records of allocated resources.
Group deliberation creates opportunities for error correction and for premature convergence. Participants can pool specialized knowledge, but they may also reinforce the first persuasive judgment, avoid disagreement, or use evaluation as an expression of status. An experimentally useful design compares independent preliminary ratings, structured discussion, and final decisions, recording who spoke and what information changed. A group consensus score alone cannot show whether collective evaluation improved accuracy. Blind independent assessments before discussion help distinguish information gain from conformity; follow-up measurement of revised artifacts tests whether agreement translated into better work. These methodological implications are consistent with the distinction between evaluation and selection developed in Rietzschel et al. (2026). A theoretical review by Rietzschel et al. (2019) analyzes how originality–feasibility tensions can lead groups to overlook high-novelty options. The chapter offers a framework and synthesis rather than a new trial that independently estimates the causal size of group selection bias.
Creative review systems also shape the pool of future ideas. Applicants anticipate what will be rewarded and alter their submissions accordingly; earlier selections influence the expertise and assumptions of subsequent judges. Such reciprocal effects are especially important in grant panels, editorial boards, product development, and cultural institutions. Longitudinal analyses should examine rejected and accepted proposals where feasible, track who controls continuation, and distinguish quality from the unequal opportunity to accumulate attention. Evaluation practices have distributive consequences, making transparency and procedural fairness relevant to the psychology of creativity. Scientific assessment should describe these institutional mechanisms without inferring that every rejected idea would have succeeded under different conditions.
Motivation, Affect, and Personal Commitment to a Draft
Creative evaluation is carried out by people who may feel pride, uncertainty, anxiety, attachment, or excitement toward the work under consideration. These states can redirect attention to particular qualities and affect willingness to abandon a preferred design. Motivation may sustain demanding revision while also encouraging defensive interpretation of criticism. A study of feedback context by Zhou (1998) supports the relevance of autonomy and informational delivery to creative performance. Nevertheless, the effects of particular emotions on later revision must be tested within their own populations and tasks. It would be misleading to infer from one laboratory feedback experiment a universal law that positive emotion promotes creativity or that criticism diminishes it.
Psychological ownership of an idea complicates evaluation because the actor who proposed it is also asked to assess it. Berg (2016) found limits to creators' forecasting advantage when judging their own ideas; this distinction makes the relation between personal commitment and judgment accuracy scientifically consequential. Investigators can compare within-person ratings of own and other-generated work while equating content quality and controlling for knowledge of authorship. The effects of external feedback may depend on whether it is experienced as diagnostic information or an attack on competence. Recording confidence before feedback, willingness to revise, and actual decisions offers more interpretable evidence than self-report alone.
An evaluation procedure should allow a creator to preserve productive commitment while remaining responsive to evidence. That practical proposition can be studied by manipulating the timing and specificity of feedback, the cost of switching, and whether multiple candidates remain available. Outcomes should include number and character of revisions, independent quality judgments, and later use. Affective responses may mediate some effects, but mediation analysis cannot replace randomization of the relevant intervention and careful measurement of confounding factors. The purpose of a psychological account is to explain when motivational investment helps sustained work and when it prevents a creator from recognizing an alternative with greater potential.
Neurocognitive Accounts of Generation and Evaluation
Research on creative cognition investigates how internally generated candidates are represented, monitored, and valued. A network perspective reviewed by Beaty et al. (2016) associates creative performance with interactions among brain systems involved in spontaneous thought, executive control, and attention. These findings challenge simple proposals that a single brain region or a single uninhibited mode of thought produces creative achievement. Evaluation requires maintaining criteria, comparing candidate properties, and sometimes suppressing highly accessible alternatives. Neural systems supporting these operations may be recruited in overlapping, task-dependent configurations. Functional imaging correlations do not establish a necessary neural mechanism or demonstrate that a region uniquely computes creativity.
An evaluative episode can be separated analytically into representing the candidate, inferring its originality and adequacy, attributing subjective value, and choosing an action. In many experiments these operations occur within seconds and share behavioral indicators, so neural dissociation requires carefully designed tasks. The same task may combine semantic search, affective response, working memory, and response selection. Moreno-Rodriguez et al. (2025) provides a targeted investigation of how originality and adequacy information relate to subjective valuation during idea evaluation, using a combination of behavioral ratings and functional magnetic resonance imaging. Its results bear directly on evaluative representations but do not track months of product revision or institutional success.
To connect neurocognitive research with material iteration, future studies can record the successive versions of a creative task together with ratings and computational descriptions of changing candidate properties. Such work would permit analysis of whether neural value-related signals predict decisions to persist, revise, or abandon a proposal. The relevant inference remains probabilistic and task-specific: brain activity can be evidence about processing under a defined experimental contrast, while performance and artifact trajectories remain the principal evidence for creative outcomes. Beaty et al. (2016) supplies a theoretical network account, not a universal physiological signature by which any artwork can be declared creative.
Subjective Value and the Neuropsychology of Idea Choice
The valuation of an idea is related to, but distinguishable from, judgments of its originality and suitability. Moreno-Rodriguez et al. (2025) asked participants to generate and assess creative ideas and modeled their subjective value as a relative weighting of originality and adequacy. The reported functional imaging analyses linked these attributes to activity in networks associated with default-mode and executive-control processing and linked subjective idea value to human reward-related systems. This provides an experimentally grounded starting point for understanding why different people prefer different candidates. It does not entail that subjective value is identical to social creativity, nor that an activated region demonstrates a particular mental experience.
Valuation depends upon preferences, expectations, and available alternatives. A candidate may be judged highly original yet receive little subjective value because the evaluator expects poor fit, or it may be attractive despite modest novelty when it solves an urgent problem. Behavioral measurement should separately record originality, adequacy, liking, confidence, and final choice. Computational models can test whether an estimated weighting of attributes predicts held-out choices rather than merely reproducing ratings used to fit the model. Moreno-Rodriguez et al. (2025) is relevant because it assesses multiple components within one paradigm. Generalization beyond brief idea-generation tasks remains an open question requiring different domains and external outcome measures.
Interpretation of neural findings must remain proportional to their designs. Sample sizes typical of functional imaging research can limit precision of individual-difference estimates, and researcher-selected regressors may partly determine which psychological quantities appear separable. Independent replication, transparent preprocessing, task reliability, alternative model comparison, and prediction in new participants are essential. A causal test of a neural mechanism would require justified interventions or convergent perturbational evidence beyond regional correlation. When discussing Artificial Era creativity, computational analogies to scoring and reward functions may be useful analytical tools, but they do not license direct equivalence between engineered optimization signals and human neural value systems.
Human–Artificial Co-Creation and Evaluative Dependence
Generative systems can now supply candidates, critiques, and revisions within human creative workflows. The pertinent psychological question is how exposure to machine-produced alternatives changes human evaluation, selection, and subsequent editing. Doshi & Hauser (2024) used an online randomized experiment in which writers could receive generative artificial intelligence ideas for short stories. AI-supported stories were evaluated more favorably on several individual creative-quality dimensions, while the resulting set of stories showed greater similarity across writers. The findings reveal a tension between improvements in individual outputs and diversity across a population of artifacts. They concern a specified task, model, participant group, and evaluation procedure; the result should not be generalized mechanically to every creative domain.
Human–artificial iteration can involve human criterion-setting, machine candidate generation, human selection, further machine transformations, and critical human redrafting. Scientific assessment needs an explicit record of which actor or tool performed each operation. If a system proposes many similar alternatives, the human may experience high local quality while seeing a narrower region of the broader idea space. An experimental study can compare human-only drafting, automated suggestions, automated critiques, and combined workflows while holding time and resources as constant as possible. It should measure human agency, independent artifact quality, stylistic diversity, subsequent revision, and retention of evaluative skill. Doshi & Hauser (2024) provides evidence about output similarity, not a complete theory of co-creative practice.
Artificial Era research also makes authorship and responsibility salient. A creative artifact can emerge from an iterative relation among human judgments, computational outputs, organizational constraints, and public standards. Attribution should describe observable contributions rather than infer a system's subjective experience or untested autonomous creative intent. A human evaluating a machine output may rely on apparent fluency and overlook factual, cultural, or practical defects; equally, a machine suggestion may expose a productive alternative that would otherwise remain unexplored. These are empirically testable possibilities. Their evaluation requires provenance of prompts and outputs, blind ratings where possible, and tests of downstream performance. The philosophical analysis of Homo and Artificial provides a broader historical context; scientific claims about the psychological effects of particular tools must stand on independent evidence.
Diversity, Originality, and Collective Consequences
Evaluating a single artifact and evaluating a portfolio of artifacts are different scientific tasks. An individual may improve a story by adopting a strong suggestion; if many individuals receive similar suggestions, the group may produce a narrower range of work. Doshi & Hauser (2024) demonstrated this distinction within an online story-writing paradigm. Individual creativity ratings and cross-person similarity both matter, but they answer different questions. A strong design should therefore measure local outcome quality, diversity across outputs, and the longer-term distribution of creative opportunities. It should also consider whether outputs converge because of shared prompts, shared training exposure, socially transmitted conventions, or user choices.
The statistical measurement of diversity requires choices about representations and distances. Lexical overlap, semantic embeddings, thematic coding, expert judgments of conceptual similarity, and formal pattern analysis may disagree, especially in artistic work. Researchers should validate that a diversity metric captures conceptually important differences in the domain and is robust to irrelevant changes in length or vocabulary. Diversity may also reflect constraints that are desirable for safety and interoperability, so any claim of improvement must identify the objectives. A portfolio of novel but unusable designs differs from a balanced collection of viable and distinctive ones. The originality and effectiveness framework articulated by Runco & Jaeger (2012) remains useful, provided the level of analysis is explicit.
An extended human–artificial revision study would follow successive drafts and compare when automated suggestions are introduced. The researcher could independently estimate content similarity, final quality, originality, evaluator calibration, and the frequency with which suggestions are rejected or transformed. Randomization of suggestion diversity, alongside controlled task exposure, would help isolate causal mechanisms. Without intermediate versions, investigators cannot tell whether convergence arises during initial idea generation, selective uptake, or later revision. The scientific significance of Artificial Era creativity therefore includes the changed ecology of evaluation: large quantities of candidate material enter human decision processes under conditions that alter the balance between individual improvement and collective variation.
Methods for Studying Evaluation and Sequential Revision
Creative evaluation research has a measurement problem at three distinct levels: the quality of a candidate, the accuracy of the judgment about that candidate, and the effectiveness of the change made after judgment. Rietzschel et al. (2026) reviews evaluation and selection tasks and notes their methodological fragmentation. A valid experiment should specify a criterion before participants respond, sample candidates with an appropriate range of quality, and distinguish an ordinal ranking from a dichotomous selection. When researchers assess revision, they should retain the complete sequence of versions and obtain evaluations from judges blinded to the order of versions and experimental condition. Such procedures help distinguish improvement from expectations that later versions must be better.
Relevant data sources include item-level ratings, forced choices, response latency, confidence, eye movements, screen recordings, design sketches, version-control histories, peer critiques, and subsequent implementation outcomes. None is a complete measure of creative judgment. Verbal reports may capture conscious standards while leaving automatic preferences unreported; process logs document actions while not revealing the reasons for them; neural signals provide correlates of defined computational contrasts but do not certify novelty. Combining these measures permits competing explanations to be tested. The design should report the sampling of ideas, people, judges, tasks, and contexts, because each influences the effective population to which results can generalize. Goldschmidt (1991) exemplifies detailed trace evidence, while Moreno-Rodriguez et al. (2025) exemplifies a different behavioral-neural methodology.
For longitudinal analysis, repeated measurements should distinguish between-person differences from within-person change. A creator's overall skill cannot automatically explain why a particular edit improved a particular artifact. Researchers can model change in blinded quality ratings across successive versions, allowing for variation across creators, tasks, and judges. Experiments might randomize feedback timing or evaluation criteria, while field studies can use preregistered matched comparisons and sensitivity analyses for unobserved confounding. Reporting all experimental conditions, unsuccessful revisions, and negative results is vital. The outcome must be specified as a measurable change in artifacts or decisions rather than the impression that the creative process was rich or enjoyable.
Evidence Appraisal: What Studies Establish and What They Leave Open
The clearest experimental evidence about revision standards comes from Lonergan et al. (2004), where the effect of evaluative instructions depended on the originality of the starting idea. Rietzschel et al. (2010) experimentally investigated the discrepancy between producing ideas and selecting strong ones. Zhou (1998) identified an interaction among feedback features and autonomy in a laboratory creative-performance task. These studies make causal claims about manipulated conditions within defined contexts. They do not establish a universal mechanism governing all kinds of artistic work, nor can their distinct dependent variables be combined without accounting for task differences. The central empirical synthesis is that evaluation matters, that its standards are contingent, and that successful selection and improvement require more than generating many candidates.
Quantitative secondary evidence offers an important cross-study check. Guo et al. (2022) aggregated 96 effect sizes from 20 studies involving 3,019 participants; the small pooled correlation between divergent thinking and evaluative skill indicates meaningful separation between generative and evaluative performance. However, heterogeneity in task design and the origins of ideas limits simple score comparisons. Sio & Lortie-Forgues (2024) reviewed creativity training evidence across decades, but training effects cannot be attributed specifically to evaluation or revision unless the intervention isolated those components. Rietzschel et al. (2026) documents how methodological fragmentation in the evaluation and selection literature itself complicates synthesis. No meta-analytic effect for revision quality is reported here because an appropriate directly comparable pooled estimate was not verified.
Field and contemporary neural studies contribute different kinds of inference. Berg (2016) tested forecasting with professionals and audience outcomes and supplied a laboratory component; external validity is stronger than in a single student task, but market and performance success remain context dependent. Moreno-Rodriguez et al. (2025) linked behavioral valuation variables to neural signals, supporting a candidate account of subjective value rather than causal localization of creativity. Doshi & Hauser (2024) randomized access to generative AI suggestions and measured both individual performance and inter-output similarity within a bounded writing task. These results collectively motivate an integrated process model, although no single study simultaneously measures the full chain from proposal generation through repeated critique, revision, external uptake, and long-term achievement.
Replication, Construct Validity, and Generalizability
Replication must preserve the scientific question while testing the stability of its mechanism. A direct replication of the Lonergan et al. (2004) interaction would retain the starting-idea manipulation and evaluation-standard conditions and prospectively document the scoring of revised plans. A conceptual replication would extend the interaction to a different medium, such as engineering designs or narrative editing, while keeping appropriate originality and implementation criteria. Failure to reproduce an exact effect may arise from different initial ideas, judge populations, instruction interpretation, task incentives, or genuine context dependence. Such possibilities should be modeled and reported rather than selected after results are known.
Construct validity depends on a match between the proposed psychological capacity and its behavioral indicator. An originality-rating task can show sensitivity to novelty but may reveal little about whether a participant can predict practical success. An ability to choose one favored solution may instead reflect conformity to group standards. Rietzschel et al. (2026) clarifies the methodological need to separate creative evaluation and selection; Guo et al. (2022) demonstrates that their relationships with divergent thinking depend on task characteristics. Investigators should examine interrater agreement, test–retest reliability, measurement invariance across cultural and educational groups, and validation against prospective outcomes. Independent raters and transparent scoring criteria are essential, especially where the evaluation target is inherently contested.
Generalization also depends on the task population. A word-association task differs from a collaborative scientific investigation; undergraduate advertising ideas differ from professional architectural designs; an online story-writing experiment differs from long-term artistic education. Across these domains, originality and usefulness remain informative, but the procedures for deciding and revising are shaped by materials, time, expertise, and institutions. A cumulative research program needs multi-site studies, sufficiently heterogeneous task sets, preregistered analysis plans, shared data where ethically possible, and independent investigations of real-world consequences. Scientific disagreement about the meaning of creative quality may itself be a research outcome, not simply an obstacle to obtaining a single score.
Open Questions and a Program for Cumulative Research
One open question is when evaluation should begin. Early screening can conserve effort but may eliminate fragile ideas that require development to demonstrate their promise. Another concerns how creators learn to revise: is evaluative improvement domain-specific, or can principles transfer across technical, artistic, and interpersonal tasks? A third concerns who should evaluate. Expertise may improve judgments while particular institutional standards exclude unfamiliar forms. A fourth concerns optimal persistence: neither continuous revision nor immediate acceptance necessarily produces the best artifact. Each question can be expressed as a testable comparison involving defined candidate pools, time budgets, evaluation criteria, random assignments, and independently rated outcomes. Lonergan et al. (2004), Rietzschel et al. (2010), and Guo et al. (2022) provide complementary starting points.
Longitudinal evidence is particularly needed to connect momentary judgment with later achievement. Researchers should register successive versions, the source and content of critiques, changes in criteria, abandoned options, predictions of success, and subsequent audience or user evaluation. Cross-cultural samples can reveal whether a widely used measure rewards local conventions. Collaborative settings should record asymmetries of authority and access to further development. Human–artificial settings add new questions about suggestion diversity, perceived ownership, evaluation dependence, and the long-term maintenance of human evaluative skill. The result of Doshi & Hauser (2024) makes the distinction between individual quality and population-level diversity especially urgent, while Berg (2016) shows why forecasting should be compared with actual outcomes.
An integrative experiment could assign matched creators to conditions that vary the timing and type of evaluation, preserve all intermediate versions, and separately measure novelty, adequacy, predicted usefulness, and preference. A second phase could vary the availability of expert or automated suggestions and the number of revisions permitted. Independent judges would assess versions in randomized order, and later users would evaluate performance in an authentic context. Such a program could test whether criterion changes mediate improvement, whether selection removes candidates with high revision potential, and whether automated support changes diversity. These are proposed designs, not findings already established by the studies cited here. Their value is to convert broad intuitions about creative cycles into falsifiable predictions.
Conclusion
Creative evaluation, iteration, and revision form a measurable organization of judgments, decisions, and transformations through which possible ideas become developed artifacts. The literature distinguishes assessing a candidate from choosing it, predicting its future reception from rating its immediate appeal, and changing an artifact from showing that it has improved. The definitional framework of Runco & Jaeger (2012) and the methodological synthesis of Rietzschel et al. (2026) establish why these distinctions matter. Their consequences are evident in experimental findings that evaluative standards interact with starting originality, in the limited correspondence between divergent thinking and evaluative skill, and in field evidence that forecasting accuracy changes with the evaluator's role. The scientific object is a sequence of situated decisions and observable results.
The most secure conclusions are conditional. Evaluation is shaped by the criterion, the initial quality of the candidate, the information available, the evaluator's expertise and relationship to the work, and the opportunity to revise. Lonergan et al. (2004) shows that revision standards can work differently depending on the nature of the idea; Guo et al. (2022) supports partial separability between generation and evaluation; Berg (2016) documents role-linked forecasting differences; Zhou (1998) shows context-dependent effects of feedback and autonomy. Design tracing Goldschmidt (1991), neural valuation research Moreno-Rodriguez et al. (2025), and human–artificial writing experiments Doshi & Hauser (2024) expand the evidence across levels of analysis. Each contribution addresses a different outcome and should remain identifiable as such.
A cumulative psychology of creative revision will depend on careful longitudinal measurement, independent quality standards, transparent decisions, experimental manipulation of feedback and constraints, and assessment of both successful and rejected alternatives. Research can then determine when evaluation protects creative potential, when selection prematurely narrows it, and which revision procedures genuinely improve work. In the Artificial Era, the same rigor applies to human–artificial creative workflows, where diverse candidates, human evaluative capacity, and collective outcomes are reshaped by computational assistance. The enduring scientific task is to explain the pathways from a provisional idea to an independently assessable, contextually effective, and sometimes original contribution.
References
- Amabile, T. M. (1982). Social psychology of creativity: A consensual assessment technique. Journal of Personality and Social Psychology, 43(5), 997–1013. https://doi.org/10.1037/0022-3514.43.5.997
- Basadur, M. (1995). Optimal ideation-evaluation ratios. Creativity Research Journal, 8(1), 63–75. https://doi.org/10.1207/s15326934crj0801_5
- Beaty, R. E., Benedek, M., Silvia, P. J., & Schacter, D. L. (2016). Creative cognition and brain network dynamics. Trends in Cognitive Sciences, 20(2), 87–95. https://doi.org/10.1016/j.tics.2015.10.004
- Berg, J. M. (2016). Balancing on the creative highwire: Forecasting the success of novel ideas in organizations. Administrative Science Quarterly, 61(3), 433–468. https://doi.org/10.1177/0001839216642211
- Doshi, A. R., & Hauser, O. P. (2024). Generative AI enhances individual creativity but reduces the collective diversity of novel content. Science Advances, 10(28), eadn5290. https://doi.org/10.1126/sciadv.adn5290
- Goldschmidt, G. (1991). The dialectics of sketching. Creativity Research Journal, 4(2), 123–143. https://doi.org/10.1080/10400419109534381
- Guo, Y., Lin, S., Acar, S., Jin, S., Xu, X., Feng, Y., & Zeng, Y. (2022). Divergent thinking and evaluative skill: A meta-analysis. The Journal of Creative Behavior, 56(3), 432–448. https://doi.org/10.1002/jocb.539
- Lonergan, D. C., Scott, G. M., & Mumford, M. D. (2004). Evaluative aspects of creative thought: Effects of appraisal and revision standards. Creativity Research Journal, 16(2–3), 231–246. https://doi.org/10.1080/10400419.2004.9651455
- Moreno-Rodriguez, S., Béranger, B., Volle, E., & Lopez-Persem, A. (2025). The human reward system encodes the subjective value of ideas during creative thinking. Communications Biology, 8, Article 37. https://doi.org/10.1038/s42003-024-07427-4
- Nijstad, B. A., De Dreu, C. K. W., Rietzschel, E. F., & Baas, M. (2010). The dual pathway to creativity model: Creative ideation as a function of flexibility and persistence. European Review of Social Psychology, 21(1), 34–77. https://doi.org/10.1080/10463281003765323
- Rietzschel, E. F., Nijstad, B. A., & Stroebe, W. (2010). The selection of creative ideas after individual idea generation: Choosing between creativity and impact. British Journal of Psychology, 101(1), 47–68. https://doi.org/10.1348/000712609X414204
- Rietzschel, E. F., Nijstad, B. A., & Stroebe, W. (2019). Why great ideas are often overlooked: A review and theoretical analysis of research on idea evaluation and selection. In P. B. Paulus & B. A. Nijstad (Eds.), The Oxford handbook of group creativity and innovation (pp. 179–197). Oxford University Press. https://doi.org/10.1093/oxfordhb/9780190648077.013.11
- Rietzschel, E. F., Ritter, S. M., & Baas, M. (2026). A systematic review of creativity evaluation and creativity selection measurement tasks. Psychology of Aesthetics, Creativity, and the Arts, 20(1), 21–46. https://doi.org/10.1037/aca0000638
- Runco, M. A., & Basadur, M. (1993). Assessing ideational and evaluative skills and creative styles and attitudes. Creativity and Innovation Management, 2(3), 166–173. https://doi.org/10.1111/j.1467-8691.1993.tb00088.x
- Runco, M. A., & Jaeger, G. J. (2012). The standard definition of creativity. Creativity Research Journal, 24(1), 92–96. https://doi.org/10.1080/10400419.2012.650092
- Sio, U. N., & Lortie-Forgues, H. (2024). The impact of creativity training on creative performance: A meta-analytic review and critical evaluation of 5 decades of creativity training studies. Psychological Bulletin, 150(5), 554–585. https://doi.org/10.1037/bul0000432
- van Broekhoven, K. (2025). The evaluation and selection of creative ideas in educational settings: Current knowledge and future directions. Creativity Research Journal, 37(2), 307–313. https://doi.org/10.1080/10400419.2023.2253403
- Zhou, J. (1998). Feedback valence, feedback style, task autonomy, and achievement orientation: Interactive effects on creative performance. Journal of Applied Psychology, 83(2), 261–276. https://doi.org/10.1037/0021-9010.83.2.261