Is Your Paper Ready for Computers & Education? A Pre-Submission Readiness Check

These are the current, public submission limits and metrics that bear on readiness. Confirm them on the journal's own guide for authors before you submit, since Elsevier updates the limits and the metric schedule.

Source: Computers & Education guide for authors and aims and scope (Elsevier, ScienceDirect), Clarivate JCR 2024, and Scopus CiteScore (accessed June 2026). The acceptance rate is a community estimate, not an official figure; verify the live limits before submitting.

The headline that matters for readiness: the desk filter is fast, and it is about education first. The guide states that "papers that do not address the criteria outlined in the author guidelines will be returned without review," and the criteria center on a connection to learning and teaching. Treat the learning outcome and the scope framing as gating, not as polish.

Computers & Education triages on education and the learning outcome before any reviewer reads the methods. Each named pattern below maps to a specific editorial triage decision, and editors consistently return for these before peer review begins.

The tool-evaluation study with no learning outcome. The most common fast return. The system is real and the evaluation is competent, but the dependent variables are satisfaction, engagement, perceived learning, or usage, with no test of actual learning or teaching. The guide is explicit that delivery-platform performance is insufficient; a paper whose strongest result is "84 percent of students agreed the tool was helpful" is not ready, because it never measures an educational consequence.

No learning-theory framing. A study can have a learning outcome and still be returned if nothing connects the design to a theory of how people learn. When the rationale is "we built this feature because it is novel" rather than "cognitive load theory predicts this segmentation will reduce extraneous load," reviewers read the contribution as engineering, not education, and editors triage accordingly.

A CS or HCI system paper with a thin education angle. The science is sound, the system is genuinely new, but the education is a wrapper. The fix is rarely possible at the desk: if the real contribution is the architecture, the model, or the interaction technique, a CS or HCI venue is the honest home, and Computers & Education: Artificial Intelligence is the closer sibling for AI-method-led work with a learning claim.

The underpowered single-context quasi-experiment. One class, no control, a small sample, and a positive result reported as if it generalizes. Editors and reviewers at a selective educational-technology journal read this as a pilot. The claim has to match the design, and the design has to support generalization, or the paper reads as preliminary.

Architecture-first AI framing. A new model or prompting pipeline with benchmark numbers and an assertion that it will help learning, but no learning measure and no theory. As large-language-model papers flood the field, this is the fastest-growing return pattern: the educational consequence is asserted, never demonstrated.

Walk each manuscript component before you submit. The order below mirrors what a Computers & Education editor reads first.

Title and abstract. The 250-word abstract must name the educational problem and the learning result before the technology. If a generalist ed-tech editor cannot see what students learned or how teaching changed from the abstract alone, the paper is not ready. Lead with the learning contribution, not the system.

Highlights. Three to five bullets of up to 85 characters. These must describe the educational contribution, the learning effect, or the theory advance, not the system's features. Highlights that read as a product changelog signal a tool-evaluation paper.

Introduction and theory. This is where the breadth case is won. State the learning or teaching question, then the gap, then the theory that frames your design. A technology-first introduction that postpones the education to the discussion is the most common reframe we recommend.

Methods and statistical analysis. The design must match the claim: a real comparison or control, a sample powered for the effect, and an analysis plan transparent enough to reproduce. A single-context quasi-experiment can be publishable, but only if the generalization argument is explicit and honest.

Learning measures. Name the validated construct: pre-post knowledge test, transfer task, authentic assessment, or an educational-psychology measure. Satisfaction and engagement can be secondary, never the only dependent variable.

Results and figures. Lead with the figure that carries the learning effect, not the system screenshot. A results section whose centerpiece is a usage dashboard reads as a technology report.

Discussion and generalization. Connect the result back to the theory and argue, concretely, why it would hold beyond this classroom, this tool, and this cohort. Editors look for the educational implication, not a feature roadmap.

Ethics, data availability, and references. Name the approving committee and reference number, point the data-availability statement to a repository or state the specific access conditions, and keep the references current in learning sciences, not only in computing.

If you want a manuscript-specific signal across all of these components before you submit, run a free readiness scan.

If the readiness check says the paper is sound but not a Computers & Education fit, route it deliberately rather than dropping a tier and blasting it out.

For a paper returned on scope, the question is which half failed. If the learning outcome and theory are present but the venue is a poor fit, a lateral move to BJET or the AI sibling is the right call. If the learning outcome is genuinely absent, no re-target fixes that until you add a learning measure, so close that gap first.

In our pre-submission review work with Computers & Education manuscripts, four readiness gaps separate papers that clear the education-first desk triage from those that come back returned-without-review. Three of the four are fixable before you submit, and recognizing which one applies to your paper is the difference between a clean submission and a wasted desk cycle.

The learning-outcome gap: a tool-evaluation study wearing an education label. This is the readiness failure we see most often in Computers & Education submissions. The system is built, the evaluation is clean, but the dependent variables are satisfaction, engagement, usage time, or perceived learning, never a validated learning measure. The tell is consistent: the abstract's strongest sentence is about how much students used or liked the tool, not what they learned.

The fix is sometimes new analysis of data you already have (a pre-post test you ran but buried) and sometimes a new study, but recognizing that the journal will not accept a usage or satisfaction outcome as the headline is the single most important readiness check. Across the Computers & Education manuscripts we review, adding or surfacing a real learning measure changes more desk outcomes than any other intervention.

The theory gap: a sound study with no learning-theory framing. Computers & Education is a learning-sciences journal, and editors expect the design to follow from a theory of how people learn. We repeatedly see methodologically competent manuscripts that are not ready because the rationale is "this feature is novel" rather than a hypothesis grounded in cognitive load, self-regulated learning, social constructivism, or a comparable framework. The fix is rarely new data.

It is reframing the introduction so the theory drives the design and the predictions, and connecting the discussion back to that theory. A study that says "we added adaptive hints and engagement rose" is not ready; one that says "scaffolding theory predicts adaptive hints will support transfer, and we tested that" is.

The scope gap: a computer-science or HCI paper with a thin education angle. This is the one readiness gap that reframing alone often cannot close. We routinely flag manuscripts whose real contribution is an architecture, a model, or an interaction technique, with the education present as a one-paragraph motivation and a small user study.

The honest call is usually a different venue: a CS or HCI conference or journal for system-led work, or Computers & Education: Artificial Intelligence for AI-method work that genuinely carries a learning claim. Submitting a systems paper to a learning journal because the impact metric is higher is the fastest way to a return-without-review.

The rigor gap: an underpowered, single-context claim. The fourth pattern is a real learning outcome and even some theory, undermined by a design that cannot support the claim: one classroom, no control group, a small sample, and a result reported as if it generalizes. At a selective educational-technology journal, reviewers read this as a pilot.

The fix is either a stronger design with a comparison and adequate power, or an honest narrowing of the claim to what the data support, plus an explicit generalization argument. Every reputable ed-tech venue applies the same standard, so closing this gap protects the paper wherever it goes next.

The practical takeaway: the learning-outcome, theory, and rigor gaps are readiness fixes you make before submitting. The scope gap, a systems paper in education clothing, is a signal to change the target venue, not to keep arguing an engineering contribution to a learning-sciences editor. Our internal analysis of these submissions points to the same conclusion every time: at Computers & Education, the learning outcome and the education-first framing decide more desk outcomes than how impressive the technology is.

Before you commit, a Computers & Education scope and readiness check tests your manuscript against these exact gaps, so you find them before a desk editor does.