Field Weighted Citation Impact (FWCI)

Field Weighted Citation Impact (FWCI) is a normalized citation metric that compares the number of citations a publication has received to the expected number of citations for publications of the same document type, publication year, and subject area. An FWCI of 1.0 means the publication is cited exactly at the world average; values above 1.0 indicate above average citation performance, and values below 1.0 indicate below average performance.

Raw citation counts are misleading when comparing publications across different fields, years, or document types. A paper in molecular biology with 50 citations may be performing below average for its field, while a paper in mathematics with 50 citations may be exceptional. FWCI corrects for these differences by normalizing citations against appropriate benchmarks. This makes it possible to compare the citation impact of publications across disciplines, institutions, and time periods on a level playing field. Universities, research institutions, and funding agencies increasingly use FWCI in research evaluation, institutional rankings, and benchmarking exercises. It is one of the primary metrics in the Scopus database and is featured prominently in SciVal, Elsevier's research analytics platform.

FWCI is calculated by dividing the actual number of citations a publication has received by the expected number of citations for publications with the same characteristics. The expected citation count is based on the global average for publications of the same document type (article, review, conference paper, etc.), published in the same year, and classified in the same subject area(s). Citations are counted over a fixed window, the year of publication plus the following three years, both for the publication and for the similar publications that set the expectation.

For a publication in a journal classified in more than one subject area (which is common in interdisciplinary research), the expected citation count combines the averages for those areas; Elsevier's SciVal uses their harmonic mean.

FWCI is calculated at the individual publication level, but it can be aggregated to the level of a researcher, department, institution, or country by averaging the FWCI values of all publications in the group. Self citations are counted in the standard calculation; SciVal offers an option to exclude them, and reports should say which version they use.

The metric has a natural lower bound of zero (for uncited publications) and no fixed upper bound. Highly cited publications can have FWCI values of 10, 50, or even higher. The distribution of FWCI values is typically right skewed, with most publications below 1.0 and a small number of highly cited publications pulling the mean upward.

FWCI is appropriate for comparing citation impact across disciplines, evaluating the research output of multidisciplinary institutions, benchmarking institutional performance against national or global averages, and tracking citation trends over time with appropriate normalization. It is particularly useful for interdisciplinary comparisons that raw citation counts cannot support. It should be used as one metric among several in research evaluation, not as the sole criterion for assessing research quality.

A common error is interpreting FWCI as a measure of research quality rather than citation impact. Citation impact and research quality are related but not identical; some highly important papers receive fewer citations than expected (for example, in niche fields), while some heavily cited papers are cited primarily for methodological reasons rather than for their contributions. Another mistake is comparing FWCI values calculated from different databases or using different normalization windows, which can produce inconsistent results. Aggregating FWCI across publications with different volumes can be misleading; an institution with one highly cited paper may have a high average FWCI that does not represent its overall output. Finally, using FWCI for very recent publications is unreliable because citation accumulation takes time, and early FWCI values may not reflect the eventual impact of the work.

A university benchmarks its chemistry department against three peer institutions (the figures are illustrative). The department's publications from 2018 to 2022 (450 articles) have a mean FWCI of 1.45, meaning they are cited 45% more than the world average for chemistry publications of the same type and year. Peer Institution A has a mean FWCI of 1.62 (520 articles), Peer B has 1.31 (380 articles), and Peer C has 1.48 (490 articles). The department ranks third of the four institutions, just behind Peer C. Breaking down by subfield, the department's organic chemistry publications have an FWCI of 2.10, while its physical chemistry publications have an FWCI of 0.95, suggesting uneven citation performance across subfields.