210 Susanne Bauer
could be studied and the decision of what to study: were variables available
for the entire time span of the retrospective studies and were they beyond
suspicion of bias? A systematic error that would result from differences in
recording between exposure groups or disease groups compared to those
unexposed or healthy. For a study on reproductive health, this restricted the
study to using records available for the entire population, i.e., birth certificates, marriage certificates, and death certificates.
Variables on these administrative documents ended up performative in the
sense of what could be studied: causes of death data, for instance for childhood cancer between the 1950s and 1990s, were considered too heterogeneous in case ascertainment, with diagnostics established only for the later
decades. Hence, the results of such a study would rather be reflecting the data
recording and diagnostic practices than the effects of radiation exposure, or
these two would be impossible to disentangle. Following the logic of epidemiological methods, these considerations determined what could be studied
and what was left unstudied.
Prioritizing a scientifically sound study on reproductive health, as commissioned by WHO, scientists studied, for instance, sex ratios, an outcome
shown to be related to radiation exposure. 63 Another study examined twinning rates which also could be studied based on administrative documents on
marriages and births. 64 These were method-wise safe and sound studies, yet
concerns over childhood leukemia and congenital malformations remained
unaddressed in these studies, despite their scientific relevance and public
concern. In that sense and in order to put sound science first, researchers
found themselves trading public health relevance against methodologically
clean studies.
Preoccupied with databases as memory infrastructures and metrics, scientists
work with processes and politics of documentation. They grapple with the
effects of exposure, in this case human-made exposure due to nuclear testing in the past but the effects of which linger in the present and are known
to reach into the future. This is due to long half-lives of many radionuclides
released in nuclear explosions, including cesium, strontium, and plutonium.
While some exposures referred to (such as iodine in milk) were most pronounced in the months and years after atmospheric nuclear tests and thus an
issue mostly for the generations young at time of atmospheric nuclear testing,
the long-term effects of exposure, the internal exposures due to incorporated
long-lived nuclides such as strontium accumulating in bones as well as the
residual radionuclides in the ground that can be refracted and enter the food
web, will remain for decades. 65
Biomedical research into radiation effects studied the damage to people in affected territories. This, however, operated on epistemic conditions
shaped by Cold War research infrastructures of risk assessment and risk factor epidemiology. They brought with them their own ways of knowing the
health effects of radiation based on a certain set of study designs and concerns that often differed from those of local researchers and the concerns
could be studied and the decision of what to study: were variables available
for the entire time span of the retrospective studies and were they beyond
suspicion of bias? A systematic error that would result from differences in
recording between exposure groups or disease groups compared to those
unexposed or healthy. For a study on reproductive health, this restricted the
study to using records available for the entire population, i.e., birth certificates, marriage certificates, and death certificates.
Variables on these administrative documents ended up performative in the
sense of what could be studied: causes of death data, for instance for childhood cancer between the 1950s and 1990s, were considered too heterogeneous in case ascertainment, with diagnostics established only for the later
decades. Hence, the results of such a study would rather be reflecting the data
recording and diagnostic practices than the effects of radiation exposure, or
these two would be impossible to disentangle. Following the logic of epidemiological methods, these considerations determined what could be studied
and what was left unstudied.
Prioritizing a scientifically sound study on reproductive health, as commissioned by WHO, scientists studied, for instance, sex ratios, an outcome
shown to be related to radiation exposure. 63 Another study examined twinning rates which also could be studied based on administrative documents on
marriages and births. 64 These were method-wise safe and sound studies, yet
concerns over childhood leukemia and congenital malformations remained
unaddressed in these studies, despite their scientific relevance and public
concern. In that sense and in order to put sound science first, researchers
found themselves trading public health relevance against methodologically
clean studies.
Preoccupied with databases as memory infrastructures and metrics, scientists
work with processes and politics of documentation. They grapple with the
effects of exposure, in this case human-made exposure due to nuclear testing in the past but the effects of which linger in the present and are known
to reach into the future. This is due to long half-lives of many radionuclides
released in nuclear explosions, including cesium, strontium, and plutonium.
While some exposures referred to (such as iodine in milk) were most pronounced in the months and years after atmospheric nuclear tests and thus an
issue mostly for the generations young at time of atmospheric nuclear testing,
the long-term effects of exposure, the internal exposures due to incorporated
long-lived nuclides such as strontium accumulating in bones as well as the
residual radionuclides in the ground that can be refracted and enter the food
web, will remain for decades. 65
Biomedical research into radiation effects studied the damage to people in affected territories. This, however, operated on epistemic conditions
shaped by Cold War research infrastructures of risk assessment and risk factor epidemiology. They brought with them their own ways of knowing the
health effects of radiation based on a certain set of study designs and concerns that often differed from those of local researchers and the concerns