TAKEPLACE: i AM FIXED

Sunday, July 3, 2011

i AM FIXED

Fixed-effect meta-regression assumes that the true effect size

θ

is normally distributed with $\mathcal{N}(\theta,\sigma_\theta)$ where $\sigma_\theta^2$ is the within study variance of the effect size. A fixed effect meta-regression model thus allows for within study variability, but no between study variability because all studies have the identical expected fixed effect size

θ

, i.e. $\varepsilon=0$ . ***Note that for the "fixed-effect" no plural is used (in contrast to "random-effects") as only ONE true effect across all datasets is assumed.***

$y_j=\beta_0+\beta_1 x_{1j}+\beta_2 x_{2j}+\cdots+\eta_j \,$

Here $\sigma^2_{\eta_j}$ is the variance of the effect size in study

j

. Fixed effect meta-regression ignores between study variation. As a result, parameter estimates are biased if between study variation can not be ignored. Furthermore, generalizations to the population are not possible.

No comments:

Post a Comment

TakePlace was made with you in mind It's about ideas and innovation, creative perspectives, and solutions. We're advocates for change, because change needs friends. We explore complex systems, adaptive agents, modeling, simulation, artificial life, and complex (information, biological, and social) networks. We especially focus on the Web as a complex information network in which we leave abundant traces of our social and semantic activities: what we do, what we are interested in, whom we talk to, what knowledge we acquire and contribute. Our research spans from modeling the dynamic processes that occur on the Web (how information networks grow and evolve, how individual and collective traffic patterns emerge, how attention bursts are generated and shaped by social and search tools) to designing tools that mine the Web to build better search, navigation, management, and recommendation tools (where ‘better’ means more intelligent, autonomous, robust, personalized, contextual, scalable, adaptive, and so on).