Correcting Citation Bias for Post-War Periodicals: Sengupta’s Weightage Formula

When researchers rank scientific journals, the simplest approach is to count citations. The journal with the most citations sits at the top, and the rest follow in descending order. This seems fair and objective. But there is a hidden problem buried inside this method, and it has to do with time. A journal that has existed for sixty years has had far more opportunities to be cited than a journal launched in 1965. Comparing their raw citation totals is like comparing the lifetime earnings of a sixty-year-old with those of a twenty-five-year-old and concluding the younger person is less capable. The bibliometrician I. N. Sengupta recognised this flaw and proposed a weightage formula to correct it. This post explains why the correction is needed, how the formula works, and what it does to journal rankings.

Why weightage is necessary 🔗

Traditional ranking lists are built on a single number: the total citations a journal has received from a set of source journals. Every citation is treated identically, regardless of when it was made. Sengupta pointed out that this equal treatment of all citations, irrespective of chronology, produces ranking positions that may not reflect a journal’s real demand or usefulness to researchers.

The issue is one of unequal time windows. An old, well-established journal has been accumulating citations for decades. A newer journal has not. So the raw citation count of a new journal is not really measuring its quality or its appeal to researchers. It is measuring something much more boring: how long the journal has been around to collect citations.

The distinction between pre-war and post-war journals 🔗

Sengupta drew a sharp line between two groups of journals. Pre-war journals are those established before the Second World War. By the time they are studied, they have a long, continuous record of publication and a deep reservoir of accumulated citations. Post-war journals are those that began publishing after the war, many of them only after the 1960s.

This timing matters enormously. According to Sengupta, the quantum of literature published by these newer journals is limited simply by the short period they have existed, in many cases less than twenty years. Their citation counts cover a variable and shorter span of time than those of the long-established journals. When you place both groups in the same ranking list and compare raw totals, the post-war journals are pushed down the list not because they are inferior, but because they are young.

The challenge with equal weightage for newly established journals 🔗

Consider what happens in practice. A post-war journal might be one of the most read and most respected publications in its field. Researchers might cite it heavily relative to how long it has existed. But in a raw citation count spanning, say, fifty years of source data, the new journal can only contribute citations from the fifteen or twenty years it has been alive. The old journals fill the rest of that window. The result is a ranking that systematically penalises new entrants.

This is a form of citation bias. The metric is supposed to measure usefulness or importance, but it is contaminated by an irrelevant variable, namely age. A fair ranking system should compare journals on equal footing. Equal weightage fails to do this because it ignores the fact that the journals being compared have had unequal lengths of time to gather citations.

Sengupta’s weightage formula 🔗

Sengupta’s solution was to standardise the time window. Instead of comparing journals over the actual, unequal periods they have existed, the formula projects every post-war journal’s citation performance onto a common reference period. He chose a benchmark of roughly twenty years, since this is the typical citation span of the established journals in his studies.

The logic of the correction 🔗

The core idea is proportional scaling. If a young journal has earned a certain number of citations over a short period, the formula estimates how many citations it would likely have earned had it existed for the full standard period. In simple terms, the correction multiplies the journal’s actual citations by the ratio of the standard period to the journal’s actual age.

The relationship can be expressed in this general form:

Corrected citations = Actual citations × (Standard period ÷ Actual period of existence)

Here the actual citations are the raw counts the journal received, the standard period is the benchmark window (around twenty years), and the actual period of existence is the number of years the journal has been publishing at the time of the study. A journal that has existed for only ten years and is being measured against a twenty-year standard would have its citations doubled, because it has had only half the time to collect them.

A worked example 🔗

Sengupta applied this method in his studies of journal rankings in fields such as biochemistry. Take a post-war journal that began publication in 1959 and received around a thousand citations in his data set. In the original ranking based purely on raw citations, it occupied a middle position. But that thousand citations had been collected over a shorter span than the decades available to the journals ranked above it. When the weightage formula scaled those citations up to the standard period, the journal’s corrected figure rose substantially, lifting it to a higher and more accurate position in the list.

The principle is straightforward once you see it. The formula does not reward a journal for being young. It simply removes the disadvantage of being young, so that what remains in the comparison is the journal’s actual citation density: how strongly it is cited per unit of time it has existed.

A deliberately conservative correction 🔗

One careful detail in Sengupta’s reasoning is worth noting. He chose to make the correction err on the cautious side. He deliberately ignored the very oldest accumulated citations of the long-established journals, reasoning that the quantum of such early citations was small and that authors tend to give somewhat more attention to newly established periodicals out of novelty. The combined effect is that the correction is designed to yield a slightly lower rather than a higher value for the new journals. In other words, the weightage applied is conservative and cannot be accused of artificially inflating young journals.

Impact of the formula 🔗

The most visible effect of applying the weightage formula is a reshuffling of the ranking list. Journals do not move randomly. Post-war journals tend to climb, while the relative positions of the older journals adjust accordingly. The new ranking reflects usefulness and demand more honestly than a list built on raw citation counts.

More accurate journal rankings in research 🔗

This matters because journal rankings are not just academic curiosities. They influence real decisions. Libraries use ranking lists to decide which journals to subscribe to, especially when budgets are tight and every subscription must be justified. Researchers use them to decide where to submit their best work and which journals to read first. Funding bodies and institutions sometimes look at where work is published. If the underlying ranking is biased against newer journals, all these downstream decisions inherit the bias.

Sengupta’s correction is part of a broader effort in bibliometrics and scientometrics to make citation-based measures fairer and more meaningful. The field has long been aware that raw citation counts are a blunt instrument. The same researcher proposed additional bibliometric parameters, such as the scientific interest, compactness, and scientific value of journals, precisely to refine traditional ranking lists and identify the accurate positions of journals in order of their usefulness.

How the idea fits into modern citation analysis 🔗

The concern Sengupta addressed in the 1970s and 1980s has not disappeared. Modern bibliometricians continue to wrestle with the problem of comparing publications and journals of different ages. Contemporary citation indicators often use techniques such as field and time normalisation, where citation counts are adjusted to account for the year of publication and the subject field, so that newer items are not unfairly disadvantaged when compared against older, more heavily cited ones.

The vocabulary has changed and the mathematics has grown more sophisticated, but the underlying insight is the same one Sengupta articulated. A citation count is meaningless until you know the time window over which it was collected. Comparing journals or papers without accounting for their age is comparing things that are not truly comparable. The weightage formula was an early and influential expression of this principle, applied specifically to the gap between pre-war and post-war periodicals.

Strengths and limitations 🔗

The strength of the formula is its simplicity. It requires only two pieces of information beyond the citation count: the standard period and the age of the journal. This makes it easy to apply by hand, which was essential in an era before automated bibliometric software. The conservative design also protects it from the criticism that it merely promotes fashionable new journals.

The limitations are equally real. The formula assumes that a journal would have continued to collect citations at the same rate had it existed longer, which is a simplification. Citation rates change over a journal’s life, often starting slow, peaking, and then declining. The choice of a twenty-year standard is also a judgement call tied to the disciplines and data Sengupta studied, and may not transfer perfectly to every field. These are not fatal flaws, but they are reasons the formula is best understood as a corrective adjustment rather than a precise prediction.

The bigger lesson for measuring research 🔗

Sengupta’s weightage formula is a small piece of a large question: how do we measure the value of scholarly work fairly? Every citation-based metric carries assumptions, and those assumptions can quietly distort the conclusions we draw. The lesson here is to always ask what a number is really measuring. A raw citation count looks objective, but it secretly encodes the age of the journal. Once you separate the signal, namely usefulness, from the noise, namely time, you get a ranking you can actually trust.

This habit of questioning the metric rather than accepting it at face value is at the heart of good scientometric practice. Numbers do not interpret themselves. The researcher has to understand where they came from and what hidden factors shaped them before drawing any conclusion about which journal, paper, or author is truly more important.

What do you think? If a journal’s citation count depends so heavily on its age, should age-correction be applied automatically to every ranking list, or are there situations where raw citation counts still tell us something useful? And given that modern databases can track citations year by year, do you think a simple twenty-year standard period is still the right benchmark, or should the correction adapt to each field’s typical citation lifespan?

References

1. https://link.springer.com/article/10.1007/BF02026414
2. https://library.stevens.edu/c.php?g=233364&p=1550018
3. https://link.springer.com/article/10.1007/BF02016772
4. https://arxiv.org/pdf/1311.4731