The model used as the basis for regulation of human radiation exposures in the United States has been a source of controversy for decades because human health consequences have not been determined with statistically meaningful certainty for the dose levels allowed for radiation workers and the general public. This dissertation evaluates the evolution of the science and regulation of radiation health effects in the United States since the early 1900s using actor-network theory and the concept of co-production of science and social order. This approach elucidated the ordering instruments that operated at the nexus of the social and the natural in making institutions, identities, discourses, and representations, and the sociotechnical imaginaries animating the use of those instruments, that culminated in a regulatory system centered on the linear no-threshold dose-response model and the As Low As Reasonably Achievable philosophy. The science of radiation health effects evolved in parallel with the development of radiation-related technologies and the associated regulatory system. History shows the principle of using the least amount of radiation exposure needed to achieve the desired effect became established as a social convention to help avoid inadvertent harm long before there was a linear no-threshold dose-response model. Because of the practical need to accept some level of occupational radiation exposure, exposures from medical applications of radiation, and some de minimis exposure to the general public, the ALARA principle emerged as an important ordering instrument even before the linear no-threshold model had gained wide support. Even before ALARA became the law, it had taken hold in a manner that allowed the nuclear industry to rationalize its operations as representing acceptable levels of risk, even though it could not be proven that the established exposure limits truly precluded harm to the exposed individuals. Laboratory experiments and epidemiology indicated that a linear dose-response model appeared suitable as a "cautious assumption" by the 1950s. The linear no-threshold model proved useful to both the nuclear establishment and its detractors. In the hands of proponents of nuclear technologies, the model predicted that occupational exposures and exposures to the public represented small risks compared to naturally occurring levels of radiation and other risks that society deemed acceptable. Conversely, opponents of nuclear technologies used the model to advance their causes by predicting health impacts for undesirable numbers of people if large populations received small radiation exposures from sources such as fallout from nuclear weapon testing or effluents from nuclear reactor operations. In terms of sociotechnical imaginaries, the linear no-threshold model was compatible with both of the dominant imaginaries involved in the actor-network. In the technocratic imaginary of institutions such as the Atomic Energy Commission, the model served as a tool for qualified experts to make risk-informed decisions about applications of nuclear technologies. In the socially progressive imaginary of the citizen activist groups, the model empowered citizens to formulate arguments informed by science and rooted in the precautionary principle to challenge decisions and actions by the technocratic institutions. This enduring dynamic tension has led to the model retaining the status of "unproven but useful" even as the underlying science has remained contested. / Doctor of Philosophy / This dissertation provides a social science perspective on an enduring paradox of the nuclear industry: why is regulation of radiation exposure based on a model that everyone involved agrees is wrong? To answer that question, it was necessary to delve into the history of radiation science to establish how safety regulation began and evolved along with the understanding of radiation's health effects. History shows the philosophy of keeping radiation exposures as small as possible for any given application developed long ago when the health effects of radiation were very uncertain. This practice turned out to be essential as science started to indicate that there may not be a safe threshold dose below which radiation exposure had no potential for health consequences. By the 1950s, a combination of theory, experiments, health studies of the survivors of the World War II atomic bombings, and other evidence suggested that the risk of cancer was proportional to the amount of radiation a person received (i.e., linear). Although this "linear no-threshold" model was far from proven, both sides used it in debates over nuclear weapon testing and safety standards for nuclear reactors in the 1950s through the early 1970s. Since the model predicted small health risks for the levels of radiation experienced by radiation workers and the public, nuclear advocates used it to argue that the risks were smaller than many other risks that people accept every day. At the same time, opposing activists used the model to argue that small cancer likelihoods added up to a lot of cancers when large populations were exposed. This decades-long discourse effectively institutionalized the model. The model's "unproven but useful" status was strengthened in the early 1970s when the Atomic Energy Commission supplemented its numeric exposure limits by turning the longtime practice of dose minimization into a requirement. This "As Low As Reasonably Achievable" requirement plays a vital role in rationalizing why a non-zero exposure limit is safe enough despite the fact that the linear no-threshold model treats any amount of radiation as harmful.

Studies on relations between number of sessions and effect of psychotherapy have usually assumed a constant rate of change across different lengths of therapy, explained by a model called the dose-effect model. This assumption has been challenged by the good-enough level (GEL) model, which makes the prediction that the rate of change will vary as a function of total number of sessions. This study aimed to compare these models. We also assessed the relationship between reliable and clinically significant change (RCSI) and total dose of therapy. Participants were drawn from two datasets in the Swedish primary care (n = 640) and adult psychiatric care (n = 249). The participants made session-wise ratings on the Clinical Outcomes in Routine Evaluation-Outcome Measure (CORE-OM). Multilevel analyses indicated a better fit using the GEL-model, with some reservations concerning RCSI and patterns of change. The results may indicate a general lawful relationship that may have implications for future research, as well as psychotherapy practice and policy making.