GALMj version ≥ 2.4.0

Introduction

GAMLj module can be installed and used in R as a standard R packages. Please check the R help for details. Here we provide some appliactions of the package in R.

Installation

At the moment, GAMLj is not in CRAN yet, so you need to install it via devtools

devtools::install_github("gamlj/gamlj")

General Linear Model

Description

General Linear Model. Estimates models using ‘lm()’ function and provides options to facilitate estimation of
interactions, simple slopes, simple effects, post-hoc tests, contrast analysis, effect size indexes and visualization of the results.

Usage

 
 gamljGlm( 
   data, 
   formula, 
   dep = NULL, 
   factors = NULL, 
   covs = NULL, 
   modelTerms = NULL, 
   fixedIntercept = TRUE, 
   showParamsCI = TRUE, 
   paramCIWidth = 95, 
   contrasts = NULL, 
   showRealNames = TRUE, 
   showContrastCode = FALSE, 
   plotHAxis = NULL, 
   plotSepLines = NULL, 
   plotSepPlots = NULL, 
   plotRaw = FALSE, 
   plotDvScale = FALSE, 
   plotError = "none", 
   ciWidth = 95, 
   postHoc = NULL, 
   eDesc = FALSE, 
   eCovs = FALSE, 
   simpleVariable = NULL, 
   simpleModerator = NULL, 
   simple3way = NULL, 
   simpleScale = "mean_sd", 
   cvalue = 1, 
   percvalue = 25, 
   simpleScaleLabels = "labels", 
   postHocCorr = list("bonf"), 
   scaling = NULL, 
   effectSize = list("beta", "partEta"), 
   homoTest = FALSE, 
   qq = FALSE, 
   normTest = FALSE, 
   normPlot = FALSE, 
   residPlot = FALSE, 
   interceptInfo = FALSE, 
   effectSizeInfo = FALSE, 
   dep_scale = "none" 
 ) 
 

Arguments

data

the data as a data frame

formula

(optional) an object of class formula (or one that can be coerced to that class): a symbolic description of the model to be fitted.

dep

a string naming the dependent variable; the variable must be numeric. Not needed if formula is used.

factors

a vector of strings naming the fixed factors. Not needed if formula is used.

covs

a vector of strings naming the covariates (continuous variables). Not needed if formula is used.

modelTerms

a list of character vectors describing fixed effects terms. Not needed if formula is used.

fixedIntercept

TRUE (default) or FALSE, estimates fixed intercept. Not needed if formula is used.

showParamsCI

TRUE (default) or FALSE , parameters CI in tables

paramCIWidth

a number between 50 and 99.9 (default: 95) specifying the confidence interval width for the parameter estimates

contrasts

a named vector of the form c(var1=‘type’, var2=‘type2’) specifying the type of contrast to use, one of ‘deviation’, ‘simple’, ‘dummy’, ‘difference’, ‘helmert’, ‘repeated’ or ‘polynomial’. If NULL, ‘simple’ is used. Can also be passed as a list of list of the form list(list(var=‘var1’,type=‘type1’)).

showRealNames

TRUE or FALSE (default), shows raw names of the contrasts variables

showContrastCode

TRUE or FALSE (default), shows contrast coefficients tables

plotHAxis

a string naming the variable placed on the horizontal axis of the plot

plotSepLines

a string naming the variable represented as separate lines in the plot

plotSepPlots

a string naming the variable defining the levels for multiple plots

plotRaw

TRUE or FALSE (default), plot raw data along the predicted values

plotDvScale

TRUE or FALSE (default), set the Y-axis range equal to the range of the observed values.

plotError

‘none’ (default), ‘ci’, or ‘se’. Use no error bars, use confidence intervals, or use standard errors on the plots, respectively.

ciWidth

a number between 50 and 99.9 (default: 95) specifying the confidence interval width for the plots.

postHoc

a rhs formula with the terms specifying the table to apply the comparisons (of the form ‘~x+x:z’

eDesc

TRUE or FALSE (default), provide lsmeans statistics

eCovs

TRUE or FALSE (default), provide lsmeans statistics conditioned to different values of the continuous variables in the model. Which levels of the continuous variable should be used is set by the simpleScale option.

simpleVariable

The variable for which the simple effects (slopes) are computed

simpleModerator

the variable that provides the levels at which the simple effects are computed

simple3way

a moderator of the two-way interaction which is probed

simpleScale

‘mean_sd’ (default), ‘custom’ , or ‘percent’. Use to condition the covariates (if any)

cvalue

how many st.deviations around the means used to condition simple effects and plots. Used if simpleScale=‘mean_sd’

percvalue

offsett (number of percentiles) around the median used to condition simple effects and plots. Used if simpleScale=‘percent’

simpleScaleLabels

how the levels of a continuous moderator should appear in tables and plots: labels, values and values_labels.

postHocCorr

one or more of ‘none’, ‘bonf’, or ‘holm’; provide no, Bonferroni, and Holm Post Hoc corrections respectively.

scaling

a named vector of the form c(var1=‘type1’,var2=‘type2’). Types are ‘centered’ to the mean, ‘standardized’, log-transformed ‘log’ or ‘none’. ‘none’ leaves the variable as it is. It can also be passed as a list of lists.

effectSize

a list of effect sizes to print out. They can be: ‘eta’ for eta-squared, ‘partEta’ for partial eta-squared, ‘omega’ for partial omega-squared, ‘epsilon’ for partial epsilon-squared, and ‘beta’ for standardized coefficients (betas). Default is beta and partEta.

homoTest

TRUE or FALSE (default), perform homogeneity tests

qq

TRUE or FALSE (default), provide a Q-Q plot of residuals

normTest

TRUE or FALSE (default), provide a test for normality of residuals

normPlot

TRUE or FALSE (default), provide a histogram of residuals superimposed by a normal distribution

residPlot

TRUE or FALSE (default), provide a scatterplot of the residuals against predicted

interceptInfo

TRUE or FALSE (default), provide ìnformation about the intercept (F test, effect size indexes)

effectSizeInfo

TRUE or FALSE (default), provide ìnformation about the effect size indexes

dep_scale

Re-scale the dependent variable. It can be ‘centered’ to the mean, ‘standardized’, log-transformed ‘log’ or ‘none’. ‘none’ (default) leaves the variable as it is.

Value

Examples

 
 data('ToothGrowth') 
 gamlj::gamljGlm(formula = len ~ supp,  data = ToothGrowth) 
 

Generalized Mixed Models

Description

Generalized Mixed Linear Model. Estimates models using ‘lme4::glmer()’ function and provides options to facilitate estimation of
interactions, simple slopes, simple effects, post-hoc tests, contrast analysis, effect size indexes and visualization of the results.

Usage

 
 gamljGlmMixed( 
   data, 
   formula = NULL, 
   dep = NULL, 
   factors = NULL, 
   covs = NULL, 
   modelTerms = NULL, 
   fixedIntercept = TRUE, 
   showParamsCI = FALSE, 
   showExpbCI = TRUE, 
   paramCIWidth = 95, 
   contrasts = NULL, 
   showRealNames = TRUE, 
   showContrastCode = FALSE, 
   plotHAxis = NULL, 
   plotSepLines = NULL, 
   plotSepPlots = NULL, 
   plotRaw = FALSE, 
   plotDvScale = FALSE, 
   plotError = "none", 
   ciWidth = 95, 
   postHoc = NULL, 
   eDesc = FALSE, 
   eCovs = FALSE, 
   simpleVariable = NULL, 
   simpleModerator = NULL, 
   simple3way = NULL, 
   simpleScale = "mean_sd", 
   cvalue = 1, 
   percvalue = 25, 
   simpleScaleLabels = "labels", 
   postHocCorr = list("bonf"), 
   scaling = NULL, 
   cluster = NULL, 
   randomTerms = list(list()), 
   correlatedEffects = "corr", 
   plotRandomEffects = FALSE, 
   plotLinearPred = FALSE, 
   nAGQ = 1, 
   effectSize = list("expb"), 
   modelSelection = "logistic", 
   custom_family = "gaussian", 
   custom_link = "identity", 
   cimethod = "wald" 
 ) 
 

Arguments

data

the data as a data frame

formula

(optional) the formula to use. The syntax is the same used in [lme4::lmer()]. See the examples

dep

a string naming the dependent variable from data, variable must be numeric

factors

a vector of strings naming the fixed factors from data

covs

a vector of strings naming the covariates from data

modelTerms

a list of character vectors describing fixed effects terms

fixedIntercept

TRUE (default) or FALSE, estimates fixed intercept

showParamsCI

TRUE (default) or FALSE , parameters CI in table

showExpbCI

TRUE (default) or FALSE , exp(B) CI in table

paramCIWidth

a number between 50 and 99.9 (default: 95) specifying the confidence interval width for the parameter estimates

contrasts

a named vector of the form c(var1=‘type’, var2=‘type2’) specifying the type of contrast to use, one of ‘deviation’, ‘simple’, ‘dummy’, ‘difference’, ‘helmert’, ‘repeated’ or ‘polynomial’. If NULL, ‘simple’ is used. Can also be passed as a list of list of the form list(list(var=‘var1’,type=‘type1’)).

showRealNames

TRUE or FALSE (default), provide raw names of the contrasts variables

showContrastCode

TRUE or FALSE (default), provide contrast coefficients tables

plotHAxis

a string naming the variable placed on the horizontal axis of the plot

plotSepLines

a string naming the variable represented as separate lines on the plot

plotSepPlots

the variable for whose levels multiple plots are computed

plotRaw

TRUE or FALSE (default), provide descriptive statistics

plotDvScale

TRUE or FALSE (default), scale the plot Y-Axis to the max and the min of the dependent variable observed scores.

plotError

‘none’, ‘ci’ (default), or ‘se’. Use no error bars, use confidence intervals, or use standard errors on the plots, respectively

ciWidth

a number between 50 and 99.9 (default: 95) specifying the confidence interval width

postHoc

a list of terms to perform post-hoc tests on

eDesc

TRUE or FALSE (default), provide lsmeans statistics

eCovs

TRUE or FALSE (default), provide lsmeans statistics

simpleVariable

The variable for which the simple effects (slopes) are computed

simpleModerator

the variable that provides the levels at which the simple effects computed

simple3way

a moderator of the two-way interaction which is probed

simpleScale

‘mean_sd’ (default), ‘custom’ , or ‘custom_percent’. Use to condition the covariates (if any)

cvalue

offset value for conditioning

percvalue

offset value for conditioning

simpleScaleLabels

decide the labeling of simple effects in tables and plots. labels indicates that only labels are used, such as Mean and Mean + 1 SD. values uses the actual values as labels. values_labels uses both.

postHocCorr

one or more of ‘none’, ‘bonf’, or ‘holm’; provide no, Bonferroni, and Holm Post Hoc corrections respectively

scaling

a named vector of the form c(var1=‘type1’,var2=‘type2’). Types are ‘centered’ to the mean, ‘clusterbasedcentered’ to the mean of each cluster, ‘standardized’, ‘clusterbasedstandardized’ standardized within each cluster, log-transformed ‘log’, or ‘none’. ‘none’ leaves the variable as it is. It can also be passed as a list of lists.

cluster

a vector of strings naming the clustering variables from data

randomTerms

a list of lists specifying the models random effects.

correlatedEffects

‘nocorr’, ‘corr’ (default), or ‘block’. When random effects are passed as list of length 1, it decides whether the effects should be correlated, non correlated. If ‘randomTerms’ is a list of lists of length > 1, the option is automatially set to ‘block’. The option is ignored if the model is passed using formula.

plotRandomEffects

TRUE or FALSE (default), add random effects predicted values in the plot

plotLinearPred

TRUE or FALSE (default), plot the predicted values in the linear predictor scale. If set, plotRaw and plotDvScale are ignored.

nAGQ

for the adaptive Gauss-Hermite log-likelihood approximation, nAGQ argument controls the number of nodes in the quadrature formula. A positive integer

effectSize

.

modelSelection

Select the generalized linear model: linear,poisson,logistic,nb, the latest being Negative Binomial

custom_family

Distribution family for the custom model, accepts gaussian, binomial, gamma and inverse_gaussian .

custom_link

Distribution family for the custom model, accepts identity, log and inverse, onemu2 (for 1/mu^2).

cimethod

.

Value

Examples

 
 data(subjects_by_stimuli) 
 gamlj::gamljMixed( 
        formula = y ~ 1 + cond+( 1|subj ), 
        data = subjects_by_stimuli) 
 gamlj::gamljMixed( 
        data = subjects_by_stimuli, 
        dep = "y", 
        factors = "cond", 
        modelTerms = "cond", 
        cluster = "subj", 
        randomTerms=list(list(c("cond","subj")))) 
         
 

Generalized Linear Models

Description

Generalized Linear Models. Estimates models using ‘glm()’ function and similar functions. It provides options to facilitate estimation of
interactions, simple slopes, simple effects, post-hoc tests, contrast analysis, effect size indexes and visualization of the results.

Usage

 
 gamljGzlm( 
   data, 
   dep, 
   factors = NULL, 
   covs = NULL, 
   modelTerms = NULL, 
   fixedIntercept = TRUE, 
   showParamsCI = FALSE, 
   showExpbCI = TRUE, 
   paramCIWidth = 95, 
   contrasts = NULL, 
   showRealNames = TRUE, 
   showContrastCode = FALSE, 
   plotHAxis = NULL, 
   plotSepLines = NULL, 
   plotSepPlots = NULL, 
   plotRaw = FALSE, 
   plotDvScale = FALSE, 
   plotError = "none", 
   ciWidth = 95, 
   postHoc = NULL, 
   eDesc = FALSE, 
   eCovs = FALSE, 
   simpleVariable = NULL, 
   simpleModerator = NULL, 
   simple3way = NULL, 
   simpleScale = "mean_sd", 
   cvalue = 1, 
   percvalue = 25, 
   simpleScaleLabels = "labels", 
   postHocCorr = list("bonf"), 
   scaling = NULL, 
   effectSize = list("expb"), 
   modelSelection = "linear", 
   custom_family = "gaussian", 
   custom_link = "identity", 
   formula 
 ) 
 

Arguments

data

the data as a data frame

dep

a string naming the dependent variable from data, variable can be numeric or factor

factors

a vector of strings naming the fixed factors from data

covs

a vector of strings naming the covariates from data

modelTerms

a list of character vectors describing fixed effects terms

fixedIntercept

TRUE (default) or FALSE, estimates fixed intercept

showParamsCI

TRUE or FALSE (default), parameters CI in table

showExpbCI

TRUE (default) or FALSE , exp(B) CI in table

paramCIWidth

a number between 50 and 99.9 (default: 95) specifying the confidence interval width for the parameter estimates

contrasts

a named vector of the form c(var1=‘type’, var2=‘type2’) specifying the type of contrast to use, one of ‘deviation’, ‘simple’, ‘dummy’, ‘difference’, ‘helmert’, ‘repeated’ or ‘polynomial’. If NULL, ‘simple’ is used. Can also be passed as a list of list of the form list(list(var=‘var1’,type=‘type1’)).

showRealNames

TRUE or FALSE (default), provide raw names of the contrasts variables

showContrastCode

TRUE or FALSE (default), provide contrast coefficients tables

plotHAxis

a string naming the variable placed on the horizontal axis of the plot

plotSepLines

a string naming the variable represented as separate lines on the plot

plotSepPlots

a string naming the variable to separate over to form multiple plots

plotRaw

TRUE or FALSE (default), provide descriptive statistics

plotDvScale

.

plotError

‘none’, ‘ci’ (default), or ‘se’. Use no error bars, use confidence intervals, or use standard errors on the plots, respectively

ciWidth

a number between 50 and 99.9 (default: 95) specifying the confidence interval width

postHoc

a list of terms to perform post-hoc tests on

eDesc

TRUE or FALSE (default), provide lsmeans statistics

eCovs

TRUE or FALSE (default), provide lsmeans statistics

simpleVariable

The variable for which the simple effects (slopes) are computed

simpleModerator

the variable that provides the levels at which the simple effects computed

simple3way

a moderator of the two-way interaction which is probed

simpleScale

‘mean_sd’ (default), ‘custom’ , or ‘custom_percent’. Use to condition the covariates (if any)

cvalue

offset value for conditioning

percvalue

offset value for conditioning

simpleScaleLabels

style for presenting condition values of a moderator. It can be ‘labels’ (default), ‘values’ or ‘labels_values’ for both.

postHocCorr

one or more of ‘none’, ‘bonf’, or ‘holm’; provide no, Bonferroni, and Holm Post Hoc corrections respectively

scaling

a named vector of the form c(var1=‘type1’,var2=‘type2’). Types are ‘centered’ to the mean, ‘standardized’, log-transformed ‘log’ or ‘none’. ‘none’ leaves the variable as it is. It can also be passed as a list of lists.

effectSize

.

modelSelection

Select the generalized linear model: linear,poisson,logistic,multinomial

custom_family

Distribution family for the custom model, accepts gaussian, binomial, gamma and inverse_gaussian .

custom_link

Distribution family for the custom model, accepts identity, log and inverse, onemu2 (for 1/mu^2).

formula

(optional) the formula to use, see the examples

Value

Examples

 
 data<-emmeans::neuralgia 
  gamlj::gamljGzlm( 
            formula = Pain ~ Duration, 
            data = data, 
             modelSelection = "logistic") 
 

Mixed Model

Description

Mixed Linear Model. Estimates models using ‘me4::lmer()’ function and provides options to facilitate estimation of
interactions, simple slopes, simple effects, post-hoc tests, contrast analysis, effect size indexes and visualization of the results.

Usage

 
 gamljMixed( 
   data, 
   dep = NULL, 
   factors = NULL, 
   covs = NULL, 
   modelTerms = NULL, 
   fixedIntercept = TRUE, 
   showParamsCI = TRUE, 
   paramCIWidth = 95, 
   ciRE = FALSE, 
   contrasts = NULL, 
   showRealNames = TRUE, 
   showContrastCode = FALSE, 
   plotHAxis = NULL, 
   plotSepLines = NULL, 
   plotSepPlots = NULL, 
   plotRaw = FALSE, 
   plotDvScale = FALSE, 
   plotError = "none", 
   ciWidth = 95, 
   postHoc = NULL, 
   eDesc = FALSE, 
   eCovs = FALSE, 
   simpleVariable = NULL, 
   simpleModerator = NULL, 
   simple3way = NULL, 
   simpleScale = "mean_sd", 
   cvalue = 1, 
   percvalue = 25, 
   simpleScaleLabels = "labels", 
   postHocCorr = list("bonf"), 
   scaling = NULL, 
   dep_scale = "none", 
   cluster = NULL, 
   randomTerms = list(list()), 
   correlatedEffects = "corr", 
   reml = TRUE, 
   lrtRandomEffects = FALSE, 
   plotRandomEffects = FALSE, 
   cimethod = "wald", 
   dfmethod = "Satterthwaite", 
   qq = FALSE, 
   normTest = FALSE, 
   normPlot = FALSE, 
   residPlot = FALSE, 
   clusterBoxplot = FALSE, 
   randHist = FALSE, 
   formula 
 ) 
 

Arguments

data

the data as a data frame

dep

a string naming the dependent variable from data, variable must be numeric

factors

a vector of strings naming the fixed factors from data

covs

a vector of strings naming the covariates from data

modelTerms

a list of character vectors describing fixed effects terms

fixedIntercept

TRUE (default) or FALSE, estimates fixed intercept

showParamsCI

TRUE (default) or FALSE , parameters CI in table

paramCIWidth

a number between 50 and 99.9 (default: 95) specifying the confidence interval width for the parameter estimates

ciRE

TRUE or FALSE (default), produce random effects confidence intervals based on profile method

contrasts

a named vector of the form c(var1=‘type’, var2=‘type2’) specifying the type of contrast to use, one of ‘deviation’, ‘simple’, ‘dummy’, ‘difference’, ‘helmert’, ‘repeated’ or ‘polynomial’. If NULL, ‘simple’ is used. Can also be passed as a list of list of the form list(list(var=‘var1’,type=‘type1’)).

showRealNames

TRUE or FALSE (default), provide raw names of the contrasts variables

showContrastCode

TRUE or FALSE (default), provide contrast coefficients tables

plotHAxis

a string naming the variable placed on the horizontal axis of the plot

plotSepLines

a string naming the variable represented as separate lines on the plot

plotSepPlots

the variable for whose levels multiple plots are computed

plotRaw

TRUE or FALSE (default), provide descriptive statistics

plotDvScale

TRUE or FALSE (default), scale the plot Y-Axis to the max and the min of the dependent variable observed scores.

plotError

‘none’, ‘ci’ (default), or ‘se’. Use no error bars, use confidence intervals, or use standard errors on the plots, respectively

ciWidth

a number between 50 and 99.9 (default: 95) specifying the confidence interval width

postHoc

a list of terms to perform post-hoc tests on

eDesc

TRUE or FALSE (default), provide lsmeans statistics

eCovs

TRUE or FALSE (default), provide lsmeans statistics

simpleVariable

The variable for which the simple effects (slopes) are computed

simpleModerator

the variable that provides the levels at which the simple effects computed

simple3way

a moderator of the two-way interaction which is probed

simpleScale

‘mean_sd’ (default), ‘custom’ , or ‘custom_percent’. Use to condition the covariates (if any)

cvalue

offset value for conditioning

percvalue

offset value for conditioning

simpleScaleLabels

decide the labeling of simple effects in tables and plots. labels indicates that only labels are used, such as Mean and Mean + 1 SD. values uses the actual values as labels. values_labels uses both.

postHocCorr

one or more of ‘none’, ‘bonf’, or ‘holm’; provide no, Bonferroni, and Holm Post Hoc corrections respectively

scaling

a named vector of the form c(var1=‘type1’,var2=‘type2’). Types are ‘centered’ to the mean, ‘clusterbasedcentered’ to the mean of each cluster, ‘standardized’, ‘clusterbasedstandardized’ standardized within each cluster, log-transformed ‘log’, or ‘none’. ‘none’ leaves the variable as it is. It can also be passed as a list of lists.

dep_scale

Re-scale the dependent variable. It could be ‘centered’ to the mean, ‘clusterbasedcentered’ to the mean of each cluster, ‘standardized’, ‘clusterbasedstandardized’ standardized within each cluster, log-transformed ‘log’, or ‘none’ (default).

cluster

a vector of strings naming the clustering variables from data

randomTerms

a list of lists specifying the models random effects.

correlatedEffects

‘nocorr’, ‘corr’ (default), or ‘block’. When random effects are passed as list of length 1, it decides whether the effects should be correlated, non correlated. If ‘randomTerms’ is a list of lists of length > 1, the option is automatially set to ‘block’. The option is ignored if the model is passed using formula.

reml

TRUE (default) or FALSE, should the Restricted ML be used rather than ML

lrtRandomEffects

TRUE or FALSE (default), LRT for the random effects

plotRandomEffects

TRUE or FALSE (default), add random effects predicted values in the plot

cimethod

.

dfmethod

.

qq

TRUE or FALSE (default), provide a Q-Q plot of residuals

normTest

TRUE or FALSE (default), provide a test for normality of residuals

normPlot

TRUE or FALSE (default), provide a histogram of residuals superimposed by a normal distribution

residPlot

TRUE or FALSE (default), provide a scatterplot of the residuals against predicted

clusterBoxplot

TRUE or FALSE (default), provide a boxplot of random effects by the first defined clustering variable

randHist

TRUE or FALSE (default), provide histogram of random Coefficients

formula

(optional) the formula to use, see the examples#’

Value

Examples

 
 data(subjects_by_stimuli) 
 gamlj::gamljMixed( 
        formula = y ~ 1 + cond+( 1|subj ), 
        data = subjects_by_stimuli) 
 gamlj::gamljMixed( 
        data = subjects_by_stimuli, 
        dep = "y", 
        factors = "cond", 
        modelTerms = "cond", 
        cluster = "subj", 
        randomTerms=list(list(c("cond","subj")))) 
         
 

Assumptions checking plots

Description

This function returns a list of plots as a ggplot objects produced by the assumptions checking options, such as ‘residPlot’, ‘normPlot’ (for continuous dependent variable models), ‘randHist’ and ‘clusterBoxplot’ (for mixed models).

Usage

 
 gamlj_assumptionsPlots(object) 
 

Arguments

object

a gamlj results object of the class ’gamlj*Results“

Value

Author(s)

Marcello Gallucci

Extract data

Description

This function returns a dataset with the variables in the GAMLj model
transformed according to GAMLj options. It is usefull to run additional
models with other R packages with the same setup used by GAMLj

Usage

 
 gamlj_data(object) 
 

Arguments

object

a gamlj results object of the class ’gamlj*Results’

Value

Author(s)

Marcello Gallucci

Examples

 
 data('qsport') 
 gmod<-gamlj::gamljGlm(formula = performance ~ hours, 
                 data = qsport, 
                 scaling = c(hours='standardized')) 
 gdata<-gamlj_data(gmod) 
 lm(performance ~ hours,data=gdata) 
 

Extract the R model

Description

Return the estimated model object of the class lm, lmer, glm depending on the model.

Usage

 
 gamlj_model(object) 
 

Arguments

object

a gamlj results object of the class ’gamlj*Results’

Value

Author(s)

Marcello Gallucci

Examples

 
 data('qsport') 
 obj<-gamlj::gamljGlm( 
    formula = performance ~ hours, 
    data = qsport) 
  model<-gamlj_model(obj) 
   
 

Predicted values from GAMLj models

Description

Deprecated. Please use ‘predict()’

Usage

 
 gamlj_predict(gobj, re.form = NULL, type = "response") 
 

Arguments

gobj

a gamlj results object of the class ’gamlj*Results’

re.form

if not NULL, specifies the random effect to be included in the computation of the predicted values. Used only for the mixed models.

type

the type of prediction required. The default is on the scale of the response variables (‘response’); Thus for binomial models the default is to compute the predicted probabilities. ‘link’ gives the scale of the linear predictors;
is on the scale of the linear predictors; The ‘terms’ option returns a matrix giving the fitted values of each term in the model formula on the linear predictor scale.
Cf. stats::predict(), stats::predict.lm()

Value

Author(s)

Marcello Gallucci

Examples

 
 data('qsport') 
 gmod<-gamlj::gamljGlm( 
    formula = performance ~ hours, 
    data = qsport) 
  preds<-gamlj_predict(gmod) 
   
 

Residuals values of a GAMLj model

Description

Deprecated. Please use ‘residuals()’

Usage

 
 gamlj_residuals(gobj, type = "response") 
 

Arguments

gobj

a gamlj results object of the class ’gamlj*Results’

type

the type of the residuals for generalized models. The alternatives are: ‘deviance’ (default), ‘pearson’, ‘working’, ‘response’, and ‘partial’. Can be abbreviated.,
cf. stats::residuals.lm()

Details

stats::predict(), stats::residuals.lm(), stats::residuals.glm()

Value

Author(s)

Marcello Gallucci

Examples

 
 data('qsport') 
 gmod<-gamlj::gamljGlm( 
    formula = performance ~ hours, 
    data = qsport) 
  preds<-gamlj_residuals(gmod) 
   
 

Deprecated simple effects

Description

Deprecated, please use ‘simpleffects’ #’

Usage

 
 gamlj_simpleEffects( 
   object, 
   variable = NULL, 
   moderator = NULL, 
   threeway = NULL, 
   ... 
 ) 
 

Arguments

object

a gamlj results object of the class ’gamlj*Results’

variable

the independent variable name

moderator

the moderator variable name

threeway

the name of the additional moderator for three way interactions

any other options accepted by the gamlj_* function

Value

Author(s)

Marcello Gallucci

GAMLj plot

Description

This function re-estimates a GAMLj model adding a new plot. If no options is passed, extracts the
plots present in the ’gamlj*Results’ object. If one plot is present, it is returned as a ggplot2 object, if more than one is present, a list of plots is returned. FALSE is returned if no plot is present or defined.

Usage

 
 ## S3 method for class 'gamlj' 
 plot(x, formula = NULL, ...) 
 

Arguments

x

a gamlj results object of the class ‘gamlj’

formula

a right hand side formula specifying the effect to plot, of the form ‘~x’, ’~x*z’ or ‘~xzw’.
It has prevalence on other options defining a plot.

all options accepted by a gamlj model function. Relevant for new plots are ‘plotHAxis’, ‘plotSepLines’ and ‘plotSepPlots’

Value

Author(s)

Marcello Gallucci

Examples

 
 data(qsport) 
 gmod<-gamlj::gamljGlm( 
   formula = performance ~ hours, 
   data = qsport) 
 plot(gmod,plotHAxis = 'hours') 
 plot(gmod,formula=~hours) 
 

Post-hoc test on GAMLj results

Description

This is a convenience function to re-estimates a GAMLj model adding posthoc tests. If no options is passed, extracts the
post-hoc tests tables already in the model results (if any). If new post-hoc are defined, the post-hoc tests tables
are returned.

Usage

 
 posthoc(object, ...) 
 ## S3 method for class 'gamlj' 
 posthoc(object, formula = NULL, ...) 
 

Arguments

object

a gamlj results object of the class ‘gamlj’

all options accepted by a gamlj model function. Relevant for new tests are
‘postHoc’ (a list of list of terms), ‘postHocCorr’, a list of correction to apply: one or more of ‘none’, ‘bonf’, or ‘holm’; provide no, Bonferroni, and Holm Post Hoc corrections respectively.

formula

a right hand side formula specifying the factors or factors combinations to test, of the form ‘~x+z’, ‘~x:z’ or ’~x*z’.
It has prevalence on other options defining a post-hoc test via character options.

Value

Author(s)

Marcello Gallucci

Examples

 
 data(fivegroups) 
 fivegroups$Group<-factor(fivegroups$Group) 
 gmod<-gamlj::gamljGlm( 
   formula = Score ~Group, 
   data = fivegroups) 
 posthoc(gmod,formula =~Group) 
 

Predicted values from GAMLj models

Description

Returns predicted values from the estimated model

Usage

 
 ## S3 method for class 'gamljGlmResults' 
 predict(object, ...) 
 ## S3 method for class 'gamljGzlmResults' 
 predict(object, type = "response", ...) 
 ## S3 method for class 'gamljMixedResults' 
 predict(object, re.form = NULL, type = "response", ...) 
 ## S3 method for class 'gamljGlmMixedResults' 
 predict(object, re.form = NULL, type = "response", ...) 
 

Arguments

object

a gamlj results object of the class ’gamlj*Results’

additional arguments for specific predict methods other than the ones specified here.

type

the type of prediction required. The default is on the scale of the response variables (‘response’); Thus for binomial models the default is to compute the predicted probabilities. ‘link’ gives the scale of the linear predictors;
is on the scale of the linear predictors; The ‘terms’ option returns a matrix giving the fitted values of each term in the model formula on the linear predictor scale.
Cf. stats::predict(), stats::predict.lm()

re.form

(formula, NULL, or NA) specify which random effects to condition on when predicting. If NULL, include all random effects; if NA or ~0, include no random effects. Used only for the mixed models.

Value

Author(s)

Marcello Gallucci

Examples

 
 data('qsport') 
 gmod<-gamlj::gamljGlm( 
    formula = performance ~ hours, 
    data = qsport) 
  preds<-predict(gmod) 
   
 

Residuals values from GAMLj models

Description

Returns residuals values from the estimated model

Usage

 
 ## S3 method for class 'gamljGlmResults' 
 residuals(object, ...) 
 ## S3 method for class 'gamljGzlmResults' 
 residuals(object, type = "deviance", ...) 
 ## S3 method for class 'gamljMixedResults' 
 residuals(object, ...) 
 ## S3 method for class 'gamljGlmMixedResults' 
 residuals(object, type = "deviance", ...) 
 

Arguments

object

a gamlj results object of the class ’gamlj*Results’

additional arguments for specific residuals methods.

type

the type of residuals for generalized models. The alternatives are: ‘deviance’ (default), ‘pearson’, ‘working’, ‘response’, and ‘partial’. Can be abbreviated. Cf. stats::residuals(), stats::residuals.lm(), stats::residuals.glm()

Value

Author(s)

Marcello Gallucci

Examples

 
 data('qsport') 
 gmod<-gamlj::gamljGlm( 
    formula = performance ~ hours, 
    data = qsport) 
  preds<-residuals(gmod) 
   
 

R-squared and pseudo-rsquared for a list of (generalized) linear (mixed) models

Description

This function calls the generic r.squared function for each of the models in the list and rbinds the outputs into one data frame

Usage

 
 rsquared.glmm(obj) 
 

Arguments

obj

single model or a list of fitted (generalized) linear (mixed) model objects

Value

Author(s)

: Jon Lefcheck

References

Lefcheck, Jonathan S. (2015) piecewiseSEM: Piecewise structural equation modeling in R for ecology, evolution, and systematics. Methods in Ecology and Evolution. 7(5): 573-579. DOI: 10.1111/2041-210X.12512

Simple Effects on GAMLj results

Description

This is a convenience function to re-estimates a GAMLj model adding simple effect analysis. If no options is passed, extracts the
simple effects tables already in the model results (if any). If new tests are defined, the simple effects tests tables
are returned.

Usage

 
 simpleEffects(object, ...) 
 ## S3 method for class 'gamlj' 
 simpleEffects(object, formula = NULL, ...) 
 

Arguments

object

a gamlj results object of the class ‘gamlj’

all options accepted by a gamlj model function. Relevant for new tests are
‘simpleVariable’ (the simple effect variable), ‘simpleModerator’ (the moderator), and ‘simple3way’ for the second moderator.

formula

a right hand side formula specifying the variables to test, of the form ‘~x:z’, ‘~x:z:w’ or ’~x*z’.
The formula is not exanded, so the first variable is the simple effect variable, the second is the moderator,
the third an optional additional moderator It has prevalence on other options defining a simple effects test via character options.

Value

Author(s)

Marcello Gallucci

Examples

 
 data(winkel) 
 wicksell$time<-factor(wicksell$time) 
 wicksell$group<-factor(wicksell$group)  
 gmod<-gamlj::gamljMixed( 
    formula = dv ~ 1 +group+ time:group+ time+( 1 | subj ), 
    data = wicksell) 
 simpleEffects(gmod,formula =~time:group) 
 

Update a GAMLj model

Description

Re-estimates a GAMLj model applying new options to the original model

Usage

 
 ## S3 method for class 'gamljGlmResults' 
 update(object, ...) 
 ## S3 method for class 'gamljGzlmResults' 
 update(object, ...) 
 ## S3 method for class 'gamljMixedResults' 
 update(object, ...) 
 ## S3 method for class 'gamljGlmMixedResults' 
 update(object, ...) 
 

Arguments

object

of class ’gamlj*Results’

any parameter to be passed to gamljGlm, gamljMixed, gamljGzlm, or gamljGlmMixed

Value

Author(s)

Marcello Gallucci

Examples

Comments?

Got comments, issues or spotted a bug? Please open an issue on GAMLj at github“ or send me an email