There is also the question of deciding what constitutes a radiation effect. Lung cancer? Genetic mutations?

These decays release radioactive alpha particles, which miners inhale. Determining
causality via accepted scientific practice demands isolating the effects
of radon exposure—deciding whether illness in uranium miners comes only
from radon exposure, or also from other contaminants. There is also the question
of deciding what constitutes a radiation effect. Lung cancer? Genetic
mutations? Epidemiologists and geneticists respond differently. When do
“effects” occur? Is lung cancer thirty years after the victim’s last exposure an
“effect”? Labor lawyers and mining corporations offer different answers.
Regardless of perspective, all these questions ultimately required knowing
how much radiation mineworkers absorb. Before the 1980s, personal
dosimetry—giving each worker a film badge or a dosimeter pen—only
detected the external exposures produced by gamma rays emitted by radioactive
rocks. Such instruments did not detect the alpha radiation emitted by
inhaled radon daughters. In many places, mine managers also feared personal
dosimetry would scare workers by alerting them to an otherwise invisible
danger. Ambient dosimetry could accommodate the heavier instruments
required to “capture” radon daughters. Less personally intrusive, it involved
installing instruments throughout the mine and averaging out their readings.
But averages did not account for the experience of men assigned to “hot
spots”: spots far from air intakes, where reduced ventilation meant elevated
radon-daughter levels and higher temperatures—the kind of place where, for
example, white foremen stationed black workers in South African mines.
The scientific (and apparently presentist and delocalized) question of
causality—“does radon cause cancer?”—is thus also, always, a historical and
geographical question. It has no single, abstract answer above and beyond
the politics of expert controversy, labor organization, capitalist production,
or colonial difference and history. That answers depend on the friction
between these, however, is only visible at the technopolitical margins of
nuclearity.12