Multilevel models, covariates, and controlled factors in experimental speech research: Unified analyses of highly structured data
暂无分享,去创建一个
Experimental speech research often makes use of complex experimental designs, but even when multiple experimental factors are manipulated, measured outcomes may be influenced by non-controlled and incompletely controlled factors. Multilevel models (of which mixed-effect models are a special case) enable unified analysis of the relationships between, on the one hand, trial-level data and, on the other, experimental factors and potentially important non-controlled variables. Fitted multilevel models allow us to draw inferences simultaneously about group-level experimental effects and covariates (the typical focus of experimental work) as well as individual subject and item properties (both of which can be important in applied research). The utility of multilevel models will be illustrated with analyses of data from a number of studies. We present models of phonological structure, gender differences, and within-gender subject variability in the acoustics of spoken English consonants; simultaneous modeling of...
[1] J. Kroll,et al. Losing Access to the Native Language While Immersed in a Second Language: Evidence for the Role of Inhibition in Second-Language Learning , 2009, Psychological science.
[2] D. Barr,et al. Random effects structure for confirmatory hypothesis testing: Keep it maximal. , 2013, Journal of memory and language.
[3] Andrew Gelman,et al. Data Analysis Using Regression and Multilevel/Hierarchical Models , 2006 .