Improving Social Science

Social science is in a replication crisis. Luckily, new and powerful statistical tools are being developed. Associate professor Matt Loftis will introduce Multiverse Analyses and Open Science Standards as ways to restore trust in the social sciences. The event is for everyone studying or interested in social science. The talk will be in English.

The replication crisis stems from researchers being unable to reproduce each other’s work. This suggested that findings thought to be true in the scientific community may not be caused by generally valid truths. Instead, they might have been caused by random, happy coincidences or by which methods the researchers (sometimes strategically) happened to apply to their data.

In other words, there were good reasons to be skeptical about whatever “knowledge” the scientific community produced in the social sciences. Much has changed in the past decades.

For this talk, we’ll be exploring the intuitions of Open Science Standards and Multiverse Analyses. Open Science Standads regard principles of openness and transparency. Multiverse Analysis is a kind of sensitivity analysis to explore the ‘alternative realities’ of what your results would look like if you had chosen to analyze it differently.