The evaluation of scientific journal quality worldwide has been largely based on the quartile attained in citation indicators such as SCImago Journal Rank (SJR), Journal Citation Reports (JCR), CiteScore (CS) or the H5 index, despite substantial differences in their methodological foundations [¹-³]. This is particularly evident in Colombia, where Publindex treats SJR/JCR/H5 as equivalent in the sense that a journal’s category is determined by the best quartile it attains across indicators. For instance, a journal ranked Q1 in SJR but Q2 in JCR will be classified as Publindex category A1 under the principle of favorability [⁴].

However, it is well known that each of the citation indicators mentioned above captures somewhat different constructs: SJR reflects citation impact weighted by the prestige of the citing journal; JCR estimates a journal’s average annual citation frequency (focusing on popularity); CS relates to the average citations per document over a four-year window (in contrast to the others, which use three- and two-year windows, respectively); and H5 is a hybrid of productivity (volume) and impact (citations) over a five-year window [⁵]. This raises the question of the extent to which these indicators are correlated, equivalent, or interchangeable when producing journal rankings.

For this purpose, a sample of 19 Ibero-American journals in physical therapy, sports therapy, and rehabilitation with presence in SJR and CiteScore and with a profile in Google Scholar Metrics (GSM) was selected in August 2025 (Table 1).

The Shapiro-Wilk test was applied to assess the normality of the indicators in the sample. Given the observed non-normality, Spearman’s coefficient (ρ_s) was used to measure monotonic correlation and Pearson’s coefficient (ρ_p) as a measure of linear correlation, to verify concordance in the direction of the relationships and strengthen the robustness of the results. In addition, Kendall’s coefficient of concordance (W_k) was computed to assess overall agreement among journal positions across citation indicators; the overlap of the top five positions (Top-5 overlap, T5_ovl) to identify agreement among the highest-ranked journals; and the maximum rank delta (∆_{rank, max}) to quantify the greatest displacement of journals between lists. These metrics were used to complement the bivariate evaluation by adding dimensions of ordering and discrepancy that are not reflected in correlation coefficients, thereby providing a more comprehensive view of concordance among the indicators.

A strong positive correlation was observed between SJR and CS and a moderate correlation between CS and H5. This indicates that journals with high SJR values also tend to have high CS values; a similar, though less pronounced, pattern occurs between CS and H5. By contrast, SJR and H5 did not show a clear concordance pattern; that is, a journal with a high H5 does not necessarily have a high SJR (Table 2).

Regarding overlap among the highest positions across indicators, SJR and CS share four journals in the top 5, whereas only two of the top 5 journals by H5 are also in the SJR top 5. This suggests that the journals leading in one indicator do not always occupy equivalent positions in another, particularly between the latter two.

When analyzing the maximum discrepancies in journal positions between indicators, the displacement reached 13 positions between SJR and H5; 11 between H5 and CiteScore; and 8 between SJR and CiteScore. These variations show that, although there are correlation trends among the three indicators, journal ordering can differ substantially depending on the metric chosen. In particular, E-Balonmano.com: Revista de Ciencias del Deporte showed the largest discrepancy between SJR and H5 (4th in SJR vs 17th in H5); SportTK showed the largest variation between SJR and CiteScore (13th vs 5th); and Revista Internacional de Medicina y Ciencias de la Actividad Física y del Deporte changed the most between CiteScore and H5 (2nd vs 12th).

Kendall’s coefficient of concordance (Wk = 0.692, p = 0.005) indicates a moderate-to-high and significant overall agreement among the citation indicators SJR, CiteScore, and H5, suggesting that despite individual discrepancies there is a general tendency to rank journals similarly across the three lists. Notably, although the bivariate correlation between SJR and H5 was not significant, Wk was significant, suggesting that even if SJR and H5 do not show a strong association on their own, there are shared ordering patterns across the three rankings when considered jointly.

The strong similarity between SJR and CiteScore may be because both are computed from the same database: Scopus. In contrast, Google Scholar Metrics’ H5 retrieves citations from a much broader network of documents, including theses, books, conference proceedings, and other sources not necessarily indexed in Scopus, which can yield higher, and often redundant, citation counts. Therefore, the data source and its quality control are decisive when interpreting bibliometric metrics [⁶].

It can be concluded that SJR, CiteScore, and H5 should not be interpreted as fully interchangeable indicators, even if they are partially comparable, for the following reasons: i) SJR and CiteScore rely on similar indexing logics and quality-control mechanisms, which favors greater concordance in impact assesment; ii) H5 is based on a broader and more heterogeneous corpus of documents, incorporating citation dynamics that are not always captured by the other two Scopus-based indicators; iii) these structural differences explain the observed variations in journal rankings; and iv) in academic assessment it may be methodologically more robust to recognize the complementarity of indicators and avoid assuming automatic equivalence among them.

Considering the recent changes introduced in the Publindex 2026 call [⁷], future research may extend this line of inquiry by examining how the Composite Metrics Index (ICM) correlates or aligns with established indicators such as SJR, CS and JCR.