mixed
(Package: afex) :
p-values for fixed effects of mixed-model via lme4::lmer()
Calculates p-values for all fixed effects in a mixed model. This is done by first fitting (with lmer) the full model and then versions thereof in which a single effect is removed and comparing the reduced model to the full model. The default is to calculate type 3 like p-values using the Kenward-Roger approximation for degrees-of-freedom (using KRmodcomp; for LMMs only). Other methods for obtaining p-values are parametric bootstrap (method = "PB") or likelihood ratio tests (method = "LRT"), both of which are available for both LMMs and GLMMs. print, summary, and anova methods for the returned object of class "mixed" are available (the last two return the same data.frame). lmer_alt is simply a wrapper for mixed that only returns the "merMod" object and correctly uses the || notation to remove correlation among factors, but otherwise behaves like g/lmer (as for mixed, it calls glmer as soon as a family argument is present).
sk2011.1
(Package: afex) :
Data from Singmann & Klauer (2011, Experiment 1)
Singmann and Klauer (2011) were interested in whether or not conditional reasoning can be explained by a single process or whether multiple processes are necessary to explain it. To provide evidence for multiple processes we aimed to establish a double dissociation of two variables: instruction type and problem type. Instruction type was manipulated between-subjects, one group of participants received deductive instructions (i.e., to treat the premises as given and only draw necessary conclusions) and a second group of participants received probabilistic instructions (i.e., to reason as in an everyday situation; we called this "inductive instruction" in the manuscript). Problem type consisted of two different orthogonally crossed variables that were manipulated within-subjects, validity of the problem (formally valid or formally invalid) and plausibility of the problem (inferences which were consisted with the background knowledge versus problems that were inconsistent with the background knowledge). The critical comparison across the two conditions was among problems which were valid and implausible with problems that were invalid and plausible. For example, the next problem was invalid and plausible:
sk2011.2
(Package: afex) :
Data from Singmann & Klauer (2011, Experiment 2)
Singmann and Klauer (2011) were interested in whether or not conditional reasoning can be explained by a single process or whether multiple processes are necessary to explain it. To provide evidence for multiple processes we aimed to establish a double dissociation of two variables: instruction type and problem type. Instruction type was manipulated between-subjects, one group of participants received deductive instructions (i.e., to treat the premises as given and only draw necessary conclusions) and a second group of participants received probabilistic instructions (i.e., to reason as in an everyday situation; we called this "inductive instruction" in the manuscript). Problem type consisted of two different orthogonally crossed variables that were manipulated within-subjects, validity of the problem (formally valid or formally invalid) and type of the problem. Problem type consistent of three levels: prological problems (i.e., problems in which background knowledge suggested to accept valid but reject invalid conclusions), neutral problems (i.e., in which background knowledge suggested to reject all problems), and counterlogical problems (i.e., problems in which background knowledge suggested to reject valid but accept invalid conclusions).
md_16.1
(Package: afex) :
Data 16.1 / 10.9 from Maxwell & Delaney
Hypothetical Reaction Time Data for 2 x 3 Perceptual Experiment: Example data for chapter 12 of Maaxwell and Delaney (2004, Table 12.1, p. 574) in long format. Has two within.subjects factors: angle and noise.
Methods defined for objects returned from the ANOVA functions aov_car et al. of class afex_aov containing both the ANOVA fitted via car::Anova and base R's aov.