Book review by Luciano Magaldi Sardella of “The Credibility Crisis in Science by Thomas Plümper and Eric Neumayer: “…The scholarly landscape into which this book enters has been shaped by growing recognition of publication bias – the documented phenomenon that studies reporting positive or statistically significant findings are substantially more likely to be published than those reporting null results or failures to replicate. Research by Franco, Malhotra, and Simonovits found that among social science experiments that were peer-reviewed and approved before data collection, 91 per cent of published studies reported positive results, while only 72 per cent of the total conducted experiments did so. This systematic skewing of the published record, combined with growing awareness of questionable research practices (QRPs), has catalysed calls for reform across multiple disciplines. Plümper and Neumayer’s work addresses this ferment by providing a comprehensive taxonomy of how empirical findings become corrupted.
The authors’ central conceptual innovation lies in their tripartite taxonomy of empirical manipulation. They distinguish carefully between three categories often conflated in public discourse. First, outright data fabrication – the wholesale invention of findings – exemplified by the case of Diederik Stapel, the Dutch social psychologist who, between 2000 and 2011, fabricated data for at least 55 publications in leading journals including Science. Second, data manipulation – the alteration of raw data to produce desired outcomes. Third, and most significantly for their argument, “tweaking” – the intentional selection of research designs, model specifications, and analytical protocols based on the results they yield, often without explicit disclosure of these choices…(More)”.
