Abstract
How neutral are decision-making technologies such as lie detection in security contexts, medical diagnosis tools in healthcare, and welfare benefits administration tools in social work? The growth of bureaucracy is a core characteristic of 20th century American politics. Administrative decision-making technologies seem to constrain bureaucratic discretion: they seem relatively rule-bound, objective, and thus bias-free. Yet existing evidence does not actually establish the neutrality of those technologies. This matters, because these technologies affect most Americans’ life chances in terms of life, liberty, and the means to pursue happiness. As denizens of the information age, we already know that increases in the number of such technologies and their use are steady and large: technology pervades modern life. Specifically, as a series of recent whistleblower leaks have established, we are living in a post-9/11 era in which means for mass data collection (the Big Data revolution) have changed the parameters of technology-mediated government. But mainstream public discourse is just beginning to incorporate this knowledge, and scholars do not know how neutral administrative technologies in general really are.
My answer to the question of whether decision-making technologies really are neutral is based on diverse forms of evidence that I evaluate using multiple methods. These methods include experimental and quasi-experimental quantitative methods, both of which rely heavily on diverse sources survey data, and qualitative historical analysis including synthesis of interviews and previously unreleased government documents from multiple federal agencies that I obtained under the Freedom of Information Act. By triangulating distinct sources of data and methods of analysis including original survey, survey experimental, and psychophysiology study evidence, I generate reliable and valid estimates of some forms of bias – and often of surprisingly robust neutrality – in technology-mediated administrative decisions. By combining the insights these analyses generate with qualitative historical insights, I build a fresh account of the political development of polygraph programs and the surveillance state.
