BEGIN:VCALENDAR VERSION:2.0 PRODID:researchseminars.org CALSCALE:GREGORIAN X-WR-CALNAME:researchseminars.org BEGIN:VEVENT SUMMARY:Daniel Spielman (Yale University) DTSTART:20231031T133000Z DTEND:20231031T143000Z DTSTAMP:20260423T035539Z UID:MathPic/120 DESCRIPTION:Title: Discrepancy Theory and Randomized Controlled Trials\nby Daniel Spiel man (Yale University) as part of Mathematical Picture Language Seminar\n\n Lecture held in Jefferson 253 and Zoom.\n\nAbstract\nDiscrepancy theory te lls us that it is possible to partition vectors into sets that look surpri singly similar to each other. By "surprisingly similar" we mean much more similar than the sets produced by a random partition. Randomized Controll ed Trials are used to test the effectiveness of interventions\, like medic al treatments and educational innovations. Randomization is used to ensur e that the test and control groups are probably similar. When we know not hing about the experimental subjects\, a random partition into test and co ntrol groups is the best choice. When we do have prior information about t he experimental subjects\, we can combine the strengths of randomization w ith the guarantees of discrepancy theory. This allows us to obtain more a ccurate estimates of the effectiveness of treatments\, or to conduct trial s with fewer experimental subjects. I will survey some fundamental results in discrepancy theory\, present a model for the analysis of RCTs\, and su mmarize results from my joint work with Chris Harshaw\, Fredrik Sävje\, a nd Peng Zhang.\n LOCATION:https://researchseminars.org/talk/MathPic/120/ END:VEVENT END:VCALENDAR