University of Connecticut

Events Calendar

Psychological Sciences Colloquium Speaker: Jose Cortina

Monday, December 7, 2015
9:00am – 10:30am

Storrs Campus
Laurel Hall 305

"On the cutting edge or falling over the side: Rethinking the way that we evaluate research"

In this presentation, I will describe some of my work showing that published work in various psychological disciplines often contains misrepresentations of models, of their genesis, and of their tests. Among the questions that I will address are these. Why do authors report incorrect degrees of freedom for their structural equation models? Why don't they explain how they formed item parcels? Why don't they explain how they handled single-indicator latent variables? Why can't their SEM results be reproduced in re-analysis of their input matrices? Why is no justification given for their choice of fit indices? In moderated mediation models, why isn't the location of the moderator specified? Why don't the path diagrams match the hypotheses? Why are second stage models tested as first stage models? In MASEM models, why don't authors report all of the input matrix values and where they came from? Why do they combine corrected and uncorrected values in an input matrix? Why don't they take into account the moderators identified in the meta analyses from which the input values were taken? The answer to all of these questions is, in a sense, because the review process doesn't screen out the bad stuff. And this leads to the most important question of all: Why Not?


Vicki Magley,

I/O Psychology (primary), Psychology Department, UConn Master Calendar

Control Panel