 |
 |
Dear all, Our next speaker for the Quant SIG Seminar Series (Monday, 14th November) will be Dr. James Hall, Department of Education, University of Oxford, who will be speaking on the topic: Testing non-nested regression models and distinguishing additive from multiplicative statistical interactions This promises to be another excellent talk. As usual the Quant SIG will meet in Seminar Room J from 12:15pm-2pm, Department of Education, 28 Norham Gardens, Oxford OX2 7PY. If you do not have access to the building, please contact Patrick Alexander ([log in to unmask]) to arrange access. Please find the abstract for this week's talk below. Best wishes, Patrick Alexander ** Speaker: Dr. James Hall Abstract: The difference-testing of 2+ statistical regression relationships relies upon comparing a “full” statistical model to an otherwise identical but “restricted” or “nested” model in which the 2+ statistical relationships are fixed to equality. The specification of this nested model provides a null- hypothesis model (where the relationships are equal to one another). There are however incidents where difference-testing is required but the specification of a nested model is either difficult or impossible. One such incidence is when alterative measures of the same concept need to have their parallel effects compared as they impact some outcome. Here we consider the less well-known statistical tests that compare non-nested models and consider a ‘real-life’ example in which these have been used: comparing two competing hypotheses for how a set of measures may work- together/interact. ** Dr. Patrick Alexander Department of Education, University of Oxford St. Hugh's College You may leave the list at any time by sending the command SIGNOFF allstat to [log in to unmask], leaving the subject line blank.