Publication modalities 'article in press' and 'open access' in relation to journal average citation
Pith reviewed 2026-05-25 00:47 UTC · model grok-4.3
The pith
Journals advancing accepted articles to early view and offering open access show higher average CiteScore.
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
There is evidence of a positive correlation between average journal impact and advancing the publication of accepted articles, and this correlation increases over time. The open access modality, in a ceteris paribus context, also correlates positively with average journal impact.
What carries the argument
Linear regression estimated on 2016 CiteScore that isolates the contribution of article-in-press status and open-access status while holding other factors fixed.
If this is right
- The positive association between early view and journal impact strengthens from 2011 to 2016.
- Open access retains a positive partial correlation with CiteScore after other variables are controlled.
- Both modalities contribute separately to measured citation influence once selection effects are isolated.
- Journals that adopt these practices earlier display measurably higher average impact than those that do not.
Where Pith is reading between the lines
- If the observed associations reflect a causal channel rather than selection alone, journals could raise their impact metrics by accelerating the release of accepted papers.
- The time trend suggests that the citation premium attached to early view may continue to rise in later years.
- Authors seeking higher visibility might rationally target journals that already offer both modalities.
Load-bearing premise
The linear regression in a ceteris paribus context fully accounts for all confounding factors that might jointly determine both the choice of publication modality and CiteScore.
What would settle it
Re-estimating the regression after adding further controls for journal age, subject field, or publisher size and checking whether the coefficients on early view and open access remain positive and significant.
read the original abstract
There has been a generalization in the use of two publication practices by scientific journals during the past decade: 1. 'article in press' or early view, which allows access to the accepted paper before its formal publication in an issue; 2. 'open access', which allows readers to obtain it freely and free of charge. This paper studies the influence of both publication modalities on the average impact of the journal and its evolution over time. It tries to identify the separate effect of access on citation into two major parts: early view and selection effect, managing to provide some evidence of the positive effect of both. Scopus is used as the database and CiteScore as the measure of journal impact. The prevalence of both publication modalities is quantified. Differences in the average impact factor of group of journals, according to their publication modalities, are tested. The evolution over time of the citation influence, from 2011 to 2016, is also analysed. Finally, a linear regression to explain the correlation of these publication practices with the CiteScore in 2016, in a ceteris paribus context, is estimated. Our main findings show evidence of a positive correlation between average journal impact and advancing the publication of accepted articles, moreover this correlation increases over time. The open access modality, in a ceteris paribus context, also correlates positively with average journal impact.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript analyzes the relationship between two publication modalities ('article in press'/early view and open access) and journal-level impact using Scopus data and CiteScore as the metric over 2011–2016. It quantifies modality prevalence, performs group-mean comparisons of impact, tracks temporal evolution of associations, and estimates a linear regression of 2016 CiteScore on modality indicators plus controls to recover ceteris-paribus correlations. The central claim is that both modalities show positive associations with average journal impact, with the early-view association strengthening over time.
Significance. If the regression specification adequately isolates the modality effects from journal-level confounders, the results would supply descriptive evidence on how these increasingly common practices co-vary with citation metrics. The multi-year Scopus coverage and explicit separation of early-view versus selection effects are strengths; however, the purely observational design means any policy implication remains correlational.
major comments (2)
- [regression estimation section] Regression estimation section: the linear model is presented as delivering ceteris-paribus estimates for the 'article in press' and open-access indicators, yet the text provides no indication that journal fixed effects, discipline dummies, article volume, or lagged CiteScore are included. Omission of these observables leaves open the possibility that the reported positive coefficients partly reflect selection of higher-impact journals into the modalities rather than an isolated effect.
- [group tests / evolution analysis] Section describing the group tests and temporal evolution: differences in mean CiteScore across modality groups and the claim that the early-view correlation increases from 2011 to 2016 are reported without accompanying robustness checks (e.g., matching, propensity-score weighting, or placebo tests on non-adopting journals). This weakens the interpretation that the observed gradients are attributable to the modalities themselves.
minor comments (2)
- The abstract and regression description should explicitly list the full set of control variables used in the ceteris-paribus specification.
- Tables reporting group means or regression coefficients should include standard errors or confidence intervals to allow assessment of statistical significance of the reported differences and trends.
Simulated Author's Rebuttal
We thank the referee for the careful reading and constructive comments on our manuscript. We address each major comment below and indicate where revisions will be made to improve clarity and address concerns about interpretation.
read point-by-point responses
-
Referee: [regression estimation section] Regression estimation section: the linear model is presented as delivering ceteris-paribus estimates for the 'article in press' and open-access indicators, yet the text provides no indication that journal fixed effects, discipline dummies, article volume, or lagged CiteScore are included. Omission of these observables leaves open the possibility that the reported positive coefficients partly reflect selection of higher-impact journals into the modalities rather than an isolated effect.
Authors: The regression is a cross-sectional specification for 2016 that includes discipline dummies and a control for journal size (number of articles). It does not include journal fixed effects or lagged CiteScore. The ceteris-paribus language refers to the included covariates. We agree this leaves room for selection on unobservables or past performance. In revision we will (i) explicitly list all included controls in the text and table, (ii) add a specification that includes lagged CiteScore as an additional control, and (iii) note the remaining limitations of the observational design. revision: partial
-
Referee: [group tests / evolution analysis] Section describing the group tests and temporal evolution: differences in mean CiteScore across modality groups and the claim that the early-view correlation increases from 2011 to 2016 are reported without accompanying robustness checks (e.g., matching, propensity-score weighting, or placebo tests on non-adopting journals). This weakens the interpretation that the observed gradients are attributable to the modalities themselves.
Authors: The group-mean comparisons and year-by-year evolution are presented as descriptive evidence documenting raw associations and their change over time. The temporal pattern is obtained by repeating the same group comparisons each year. We accept that these analyses do not employ matching or placebo tests and therefore remain correlational. In revision we will add explicit language stating that the results are descriptive associations, not causal effects, and will note the absence of the suggested robustness checks as a limitation. revision: partial
Circularity Check
No circularity: standard empirical regression on external data
full rationale
The paper conducts descriptive statistics, group comparisons, and a single linear regression of CiteScore on publication-modality indicators using Scopus data. No equations, fitted parameters, or self-citations are presented as deriving the central correlations; the regression is an ordinary least-squares estimation whose coefficients are not forced by construction to equal the inputs. The analysis is therefore self-contained against external benchmarks and receives the default non-circularity finding.
Axiom & Free-Parameter Ledger
free parameters (1)
- regression coefficients for modalities
axioms (2)
- domain assumption CiteScore is an appropriate and unbiased measure of average journal impact
- ad hoc to paper The set of control variables in the regression captures all relevant confounders
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.