## submitted). Mixed-effects modeling with crossed random effects for subjects and items (2007)

Citations: | 207 - 3 self |

### BibTeX

@MISC{Baayen07submitted).mixed-effects,

author = {R. H. Baayen and Max Planck and D. J. Davidson and D. M. Bates},

title = {submitted). Mixed-effects modeling with crossed random effects for subjects and items},

year = {2007}

}

### OpenURL

### Abstract

and items

### Citations

726 |
Statistical parametric maps in functional imaging: a general linear approach
- Friston, Holmes, et al.
- 1995
(Show Context)
Citation Context ...muli as random effects has not been widely addressed in the imaging and physiological community. In imaging studies that compare experimental conditions, for example, statistical parameter maps (spm; =-=Friston et al. 1995-=-) are calculated based on successively recorded time series for the different experimental conditions. A hypothesized hemodynamic response function is convolved with a function that encodes the experi... |

541 | Bayesian Inference in Statistical Analysis - Box, Tiao - 1992 |

482 |
Statistical Principles in Experimental Design
- Winer
- 1962
(Show Context)
Citation Context ...th participants and items analyses. Clark’s oft-cited paper presented a technical solution to this modeling problem, based on statistical theory and computational methods available at the time (e.g., =-=Winer, 1971-=-). This solution involved computing a quasi-F statistic which, in the simplest-to-use form, could be approximated by the use of a combined minimum-F statistic derived from separate participants (F1) a... |

276 |
lme4: Linear mixed-effects models using S4 classes. R package version
- Bates, Sarkar
- 2007
(Show Context)
Citation Context ...terms involving subject or item. 3 Data analysis To our knowledge, the only software currently available for fitting mixed-effects models with crossed random effects is the lme4 package (Bates, 2005; =-=Bates & Sarkar, 2005-=-) in R, an open-source language and environment for statistical 7scomputing (R development core team, 2005), available at http://cran.rproject.org. In statistical computing, R is the leading platform ... |

247 | An introduction to mcmc for machine learning - Andrieu, Freitas, et al. |

191 |
On methods in the analysis of profile data
- Geisser, Greenhouse
- 1959
(Show Context)
Citation Context ...techniques are used is to account for correlation between successive measurements in the imaging time series. These corrections are similar to corrections familiar to psychologists for nonsphericity (=-=Greenhouse & Geisser, 1958-=-). Similar analysis concerns are present within electrophysiology. In the past, journal policy in psychophysiological research has dealt with the problems posed by repeated measures experimental desig... |

182 | The language-as-"xed-e!ect fallacy: a critique of language statistics in psychological research - Clark - 1973 |

168 | Analyzing Linguistic Data: A practical introduction to statistics
- Baayen
- 2008
(Show Context)
Citation Context ...2006) using the lme4 package of Bates and Sarkar (2005) (see also Bates, 2005). The code for the simulations is available in the languageR package in the cran archives (http://cran.r-project.org, see =-=Baayen, 2007-=-). 4.1 A Split-Plot Design One constructed dataset discussed by Raaijmakers, Schrijnemakers and Gremmen (1999) has a split-plot design. The data comprise 64 observations with 8 subjects and 8 items. I... |

167 | Using SAS proc mixed to fit multilevel models, hierarchical models, and individual growth models - Singer - 1998 |

121 | 2002a. Classical and Bayesian inference in neuroimaging: theory - Friston, Penny, et al. |

88 | Early vocabulary growth: Relation to language input and gender - Huttenlocher, Haight, et al. - 1991 |

85 | Conditions under which mean square ratios in repeated measure designs have exact F-distributions - Huynh, Feldt - 1980 |

84 | Regression analyses of repeated measures data in cognitive research - Lorch, Myers - 1990 |

73 | 2002b. Classical and Bayesian inference in neuroimaging: applications - Friston, Glaser, et al. |

66 | How to deal with ”the languageas-fixed-effect fallacy”: Common misconceptions and alternative solutions - Raaijmakers, Schrijnemakers, et al. - 1999 |

57 |
Fitting linear mixed models
- Bates
- 2005
(Show Context)
Citation Context ... interaction terms involving subject or item. 3 Data analysis To our knowledge, the only software currently available for fitting mixed-effects models with crossed random effects is the lme4 package (=-=Bates, 2005-=-; Bates & Sarkar, 2005) in R, an open-source language and environment for statistical 7scomputing (R development core team, 2005), available at http://cran.rproject.org. In statistical computing, R is... |

53 |
Multilevel models in educational and social research
- Goldstein
- 1987
(Show Context)
Citation Context ...tilevel models) are used in educational and developmental research to model data that is structured in a hierarchical form, such as the natural hierarchy formed by students nested within a classroom (=-=Goldstein, 1987-=-). More complex hierarchies such as school district:school:classroom:student can also be modeled. Studies in educational settings are often focused on learning over time, and techniques developed for ... |

49 |
The continuing problem of false positives in repeated measures ANOVA in psychophysiology: A multivariate solution
- Vasey, Thayer
- 1987
(Show Context)
Citation Context ...ems posed by repeated measures experimental designs by suggesting that researchers adopt statistical procedures that take into account the correlated data obtained from these designs (Jennings, 1987; =-=Vasey & Thayer, 1987-=-). Mixed effects models are less commonly applied in psychophysiological research, as the most common techniques are the traditional univariate anova with adjustments or multivariate anova (Dien & San... |

48 | Frequency effects and the representational status of regular inflections
- Alegre, Gordon
- 1999
(Show Context)
Citation Context ...ure combines precise and imprecise information on equal footing.) We refer to this procedure as by-participant regression. Some studies report both by-item and by-participant regression models (e.g., =-=Alegre & Gordon, 1999-=-). The by-participant regression is widely regarded as superior to the by-item regression. However, the by-participant regression does not take item-variability into account. To see this, compare an e... |

40 | Maximum likelihood computations with repeated measures: Application of the EM algorithm - Laird, Lange, et al. - 1987 |

26 | Generalizability, random effects, and population inference - Holmes, Friston - 1998 |

25 |
Statistical modeling of data on teaching styles
- Aitkin, Anderson, et al.
- 1981
(Show Context)
Citation Context ...me, termed the analysis of growth curves (Goldstein, 1987; Willet et al., 1998; Nutall et al. 1989). Examples of this include the assessment of different teaching techniques on students’ performance (=-=Aitkin et al., 1981-=-), and the comparison 27sof the effectiveness of different schools (Aitkin & Longford, 1986). Goldstein et al. (1993) used multilevel techniques to study the differences between schools and students w... |

25 | A multilevel analysis of school examination results. Oxford Review of Education - Goldstein, Rasbash, et al. - 1993 |

25 | Comparing personal trajectories and drawing causal inferences from longitudinal data - Raudenbush - 2001 |

24 | On multi-level modeling of data from repeated measures designs: A tutorial - Quené, Bergh - 2004 |

24 |
The design and analysis of longitudinal studies of development and psychopathology in context: Statistical models and methodological recommendations. Development and Psychopathology
- Willet, Singer, et al.
- 1998
(Show Context)
Citation Context ... time, and techniques developed for this type of data often attempt to characterize how individuals’ performance or knowledge changes over time, termed the analysis of growth curves (Goldstein, 1987; =-=Willet et al., 1998-=-; Nutall et al. 1989). Examples of this include the assessment of different teaching techniques on students’ performance (Aitkin et al., 1981), and the comparison 27sof the effectiveness of different ... |

18 |
Mixed-effects and fMRI studies
- Friston, Stephan, et al.
- 2005
(Show Context)
Citation Context ... Field theory to make inferences about activity patterns in very large data sets (voxels from fMRI scans). These techniques are formally comparable to the techniques that are advocated in this paper (=-=Friston et al. 2005-=-). Interestingly, however, the treatment of stimuli as random effects has not been widely addressed in the imaging and physiological community. In imaging studies that compare experimental conditions,... |

13 |
Statistical modelling in school effectiveness studies
- Aitkin, Longford
- 1986
(Show Context)
Citation Context ...et al. 1989). Examples of this include the assessment of different teaching techniques on students’ performance (Aitkin et al., 1981), and the comparison 27sof the effectiveness of different schools (=-=Aitkin & Longford, 1986-=-). Goldstein et al. (1993) used multilevel techniques to study the differences between schools and students when adjusting for pre-existing differences when students entered classes. For a methodologi... |

13 | A natural approach to studying vision - Felsen, Dan - 2005 |

11 |
Application of repeated measures ANOVA to high-density ERP datasets: a review and tutorial
- Dien, Santuzzi
- 2005
(Show Context)
Citation Context ...ayer, 1987). Mixed effects models are less commonly applied in psychophysiological research, as the most common techniques are the traditional univariate anova with adjustments or multivariate anova (=-=Dien & Santuzzi, 2004-=-), but some researchers have advocated them to deal with repeated measures data. For example, Bagiella et al. (2000) suggest that mixed effects models have advantages over more traditional techniques ... |

9 | Analysis of longitudinal data: The integration of theoretical model, temporal design, and statistical model - Collins, M - 2006 |

9 |
Towards a perceptual-span theory of distributed processing in reading: A reply to Rayner
- Kliegl
- 2007
(Show Context)
Citation Context ...l noise should be brought under statistical control, at the risk of failing to detect otherwise significant effects. (For an example of mixed-effects modeling applied to eye-movement data, see, e.g., =-=Kliegl, 2007-=-). Third, qualitative properties of preceding trials should be brought under statistical control. Here, one can think of whether the response to the preceding trial in a lexical decision task was corr... |

8 | Mixed-effects models in psychophysiology. Psychophysiology 37 - Bagiella, Sloan, et al. - 2000 |

8 |
The word frequency effect in lexical decision: Finding a frequencybased component
- Gardner, Rothkopf, et al.
- 1987
(Show Context)
Citation Context ...ustified on the grounds that items may vary in familiarity across participants. For instance, if items are words, than lexical familiarity is known to vary considerably across occupations (see, e.g., =-=Gardner, Rothkopf, Lapan, & Lafferty, 1987-=-). Technically, however, nesting amounts to the strong assumption that there need not be any commonality at all for a given item across participants. This strong assumption is justified only when the ... |

8 | Development Core Team (2005), R: A language and environment for statistical computing, R Foundation for Statistical Computing - unknown authors |

8 | Regular morphologically complex neologisms leave detectable traces in the mental lexicon - Vaan, Schreuder, et al. - 2007 |

7 | Multilevel modeling of hierarchical data in developmental studies - Boyle, Willms - 2001 |

7 | Generalizing to a language population - Coleman - 1964 |

6 | Hierarchical linear models: Problems and prospects - Kreft - 1995 |

5 | Regression Modeling Strategies - Harrel - 2001 |

5 |
Differential school effectiveness
- Nutall, Goldstein, et al.
- 1989
(Show Context)
Citation Context ... developed for this type of data often attempt to characterize how individuals’ performance or knowledge changes over time, termed the analysis of growth curves (Goldstein, 1987; Willet et al., 1998; =-=Nutall et al. 1989-=-). Examples of this include the assessment of different teaching techniques on students’ performance (Aitkin et al., 1981), and the comparison 27sof the effectiveness of different schools (Aitkin & Lo... |

4 |
Linear Mixed Model Implementation
- Bates
- 2007
(Show Context)
Citation Context ... (i.e. greater than the p-value that would be obtained through simulation). Although the mathematical details of model fitting with mixed effects models are beyond the scope of the present paper (see =-=Bates, 2007-=-, for an in13sσ 2 4 3 2 1 168 167 166 165 164 163 162 161 160 159 158 155 156 157 ML 1 2 3 4 158 159 161 162 6 8 16 14 12 10 163 164 165 1 2 3 4 18 ldZ 20 σ 1 22 24 8.6 8.4 8.2 8.0 7.8 7.6 lr2 1 2 3 4... |

4 | Linear Models with R. Boca - Faraway - 2005 |

1 |
More on the language-as-fixed effect: Monte-Carlo estimates of error rates for
- Forster, Dickinson
- 1976
(Show Context)
Citation Context ..., a quasi-F test, a by-participant analysis, a by-item analysis, and an analysis that accepted the effect of SOA to be significant only if both the F1 and the F2 test were significant (F1+F2, compare =-=Forster & Dickinson, 1976-=-). The only procedures with nominal Type I error rates are the quasi-F test and the mixed-effects model with mcmc sampling. For small data sets, as typically found in agriculture with a small number o... |

1 |
Editorial policy on analysis of variance with repeated measures
- Jennings
- 1987
(Show Context)
Citation Context ...t with the problems posed by repeated measures experimental designs by suggesting that researchers adopt statistical procedures that take into account the correlated data obtained from these designs (=-=Jennings, 1987-=-; Vasey & Thayer, 1987). Mixed effects models are less commonly applied in psychophysiological research, as the most common techniques are the traditional univariate anova with adjustments or multivar... |

1 | The adequacy of repeated-meassures regression for multilevel research - Misangyi, LePine, et al. - 2006 |

1 |
A further look at the ”language-as-a-fixed-effect fallacy
- Raaijmakers
- 2003
(Show Context)
Citation Context ...consistency, if human participants, faces, and speech are to be considered random variables, then objects, artifacts, and scenes might just as well be considered random variables (also pointed out by =-=Raaijmakers, 2003-=-). Any naturalistic stimulus which is a member of a population of stimuli which has not been exhaustively sampled should be considered a random variable for the purposes of an experiment. The present ... |