Karl Theodor von Inama-Sternegg, “Neue Beiträge zur allgemeinen Methodenlehre der Statistik”

This article-length methodological review essay, first published in Statistische Monatsschrift in 1890 and later gathered in 1903, turns the smallpox-vaccination controversy into a theory of statistical proof. Inama-Sternegg follows J. Körösi’s claim that vaccination’s protective force cannot be settled by physiological experiment alone, but by observation of mass facts and mass effects. The article’s central thesis is that statistics can reach causes, not through absolute demonstration in every case, but through disciplined, probabilistic comparisons suited to collective life.

The structure is fourfold. It begins by showing how a concrete medical controversy leads into general statistical criticism; then it treats the denominator problem of “living totals”; next it examines statistical negative proof for alleged vaccine-transmitted diseases; finally it distinguishes experiment from observation. The first major conceptual move is against the claim that vaccination statistics is useless because no one knows the total living vaccinated and unvaccinated populations. Inama-Sternegg concedes that some coefficients require such totals, but argues that many valid demographic measures relate an event to its nearest causally relevant collective, not to the whole population.

Es wird eben immer eine speziell beobachtete Erscheinung auf jene nächst höhere Kollektiverscheinung bezogen, von der sie einen Teil darstellt und von der mit Grund angenommen werden kann, daß ein kausaler Zusammenhang besteht, dessen Maß gefunden werden soll.

English translation: A specifically observed phenomenon is always related to that next higher collective phenomenon of which it forms a part, and of which it may with good reason be assumed that a causal connection exists whose measure is to be found.

This prepares Körösi’s method of “relative intensity.” Since vaccinated and unvaccinated populations differ by age, poverty, and frailty, a crude comparison of all living persons would not itself prove vaccine protection. The better test compares their general morbidity and mortality with their specifically smallpox morbidity, mortality, and lethality. By recording vaccination status for all deaths, not only smallpox deaths, the statistician can ask whether the unvaccinated are merely weaker overall or disproportionately exposed to smallpox. On the Budapest and Hungarian provincial-city material for 1886, Inama-Sternegg presents this as empirically favorable to vaccination and methodologically decisive.

Es ist damit aber auch der Beweis geliefert, daß die Statistik wirklich im Stande ist, zur Lösung der Impffrage in entscheidender Weise beizutragen, indem sie die Differenzen der Morbidität, Mortalität und Letalität bei Geimpften und Nichtgeimpften aufzeigt und die spezifische Blattern-Erkrankungs- und Sterbensgefahr beider Kategorien zu messen gestattet.

English translation: With this, however, the proof is also furnished that statistics is really capable of contributing decisively to the solution of the vaccination question, by showing the differences in morbidity, mortality, and lethality between the vaccinated and the unvaccinated, and by permitting the measurement of the specific risk of smallpox illness and death in both categories.

The third section treats a harder, negative claim: that vaccination transmits other diseases. Statistics cannot prove absolute impossibility; individual transmissions may occur. It can, however, compare disease groups allegedly affected by vaccination and test whether the claimed causal pattern appears in mass data. Inama-Sternegg’s broader methodological point is that causal inference from effects is hypothetical but not arbitrary. It becomes scientific when probable antecedents are isolated through analogy, prior conjunctions, and comparative series.

Bei der ersten Methode (Zerfällung der Gesamtheiten) scheidet man alle fremden Faktoren aus; bei der letzteren (Intensitätsberechnung) berechnet man den Einfluß eines unausgeschiedenen oder unausscheidbaren Faktors durch die gegebenen übrigen Werte.

English translation: In the first method (decomposition of totals) one excludes all extraneous factors; in the latter (intensity calculation) one computes the influence of an unexcluded or inextricable factor by means of the other given values.

The final section explains why this procedure is not simply failed experiment. Older distinctions between experiment as artificial arrangement and observation as passive finding are, for Inama-Sternegg, secondary. The decisive distinction is directional: experiment proceeds from cause to effect, while statistical observation usually begins with effects and works backward through initial and final states, parallel movements, and residual contrasts.

Nichtsdestoweniger muß ein solcher prinzipieller Unterschied angenommen werden, welcher darauf beruht, daß es in der wissenschaftlichen Forschung in der Tat zwei ganz verschiedene Arten der Beweisführung für einen kausalen Zusammenhang gibt, indem entweder von der Ursache zur Wirkung, progressiv, oder von der Wirkung zur Ursache, regressiv, fortgeschritten wird.

English translation: Nevertheless such a fundamental distinction must be assumed, which rests on the fact that in scientific research there are indeed two quite different modes of proving a causal connection, in that one proceeds either from cause to effect, progressively, or from effect to cause, regressively.

The essay’s relevance lies in its early defense of statistical causal inference in population phenomena. It makes the indirectness of social and medical mass observation into a method: choose the right collective, compare intensities rather than crude totals, treat negative proof probabilistically, and understand statistical knowledge as regressive but indispensable.