この節の作者: Ravi Selker, Jonathon Love, Damian Dropmann

Repeated Measures ANOVA

(anovaRM)

Description

The Repeated Measures ANOVA is used to explore the relationship between a continuous dependent variable and one or more categorical explanatory variables, where one or more of the explanatory variables are 'within subjects' (where multiple measurements are from the same subject). Additionally, this analysis allows the inclusion of covariates, allowing for repeated measures ANCOVAs as well.

Usage

anovaRM(
  data,
  rm = list(list(label = "RM Factor 1", levels = list("Level 1", "Level 2"))),
  rmCells = NULL,
  bs = NULL,
  cov = NULL,
  effectSize = NULL,
  depLabel = "Dependent",
  rmTerms = NULL,
  bsTerms = NULL,
  ss = "3",
  spherTests = FALSE,
  spherCorr = list("none"),
  leveneTest = FALSE,
  contrasts = NULL,
  postHoc = NULL,
  postHocCorr = list("tukey"),
  emMeans = list(list()),
  emmPlots = TRUE,
  emmTables = FALSE,
  emmWeights = TRUE,
  ciWidthEmm = 95,
  emmPlotData = FALSE,
  emmPlotError = "ci",
  groupSumm = FALSE
)

Arguments

data	the data as a data frame
rm	a list of lists, where each list describes the label (as a string) and the levels (as vector of strings) of a particular repeated measures factor
rmCells	a list of lists, where each list decribes a repeated measure (as a string) from data defined as measure and the particular combination of levels from rm that it belongs to (as a vector of strings) defined as cell
bs	a vector of strings naming the between subjects factors from data
cov	a vector of strings naming the covariates from data. Variables must be numeric
effectSize	one or more of 'eta', 'partEta', or 'omega'; use eta², partial eta², and omega² effect sizes, respectively
depLabel	a string (default: 'Dependent') describing the label used for the dependent variable throughout the analysis
rmTerms	a list of character vectors describing the repeated measures terms to go into the model
bsTerms	a list of character vectors describing the between subjects terms to go into the model
ss	'2' or '3' (default), the sum of squares to use
spherTests	TRUE or FALSE (default), perform sphericity tests
spherCorr	one or more of 'none' (default), 'GG', or 'HF'; use no p-value correction, the Greenhouse-Geisser p-value correction, and the Huynh-Feldt p-value correction for shericity, respectively
leveneTest	TRUE or FALSE (default), test for homogeneity of variances (i.e., Levene's test)
contrasts	in development
postHoc	a list of character vectors describing the post-hoc tests that need to be computed
postHocCorr	one or more of 'none', 'tukey' (default), 'scheffe', 'bonf', or 'holm'; use no, Tukey, Scheffe, Bonferroni and Holm posthoc corrections, respectively
emMeans	a list of lists specifying the variables for which the estimated marginal means need to be calculate. Supports up to three variables per term.
emmPlots	TRUE (default) or FALSE, provide estimated marginal means plots
emmTables	TRUE or FALSE (default), provide estimated marginal means tables
emmWeights	TRUE (default) or FALSE, weigh each cell equally or weigh them according to the cell frequency
ciWidthEmm	a number between 50 and 99.9 (default: 95) specifying the confidence interval width for the estimated marginal means
emmPlotData	TRUE or FALSE (default), plot the data on top of the marginal means
emmPlotError	'none', 'ci' (default), or 'se'. Use no error bars, use confidence intervals, or use standard errors on the marginal mean plots, respectively
groupSumm	TRUE or FALSE (default), report a summary of the different groups

Details

This analysis requires that the data be in 'wide format', where each row represents a subject (as opposed to long format, where each measurement of the dependent variable is represented as a row).

A non-parametric equivalent of the repeated measures ANOVA also exists; the Friedman test. However, it has the limitation of only being able to test a single factor.

Output

A results object containing:

results$rmTable	a table
results$bsTable	a table
results$assump$spherTable	a table
results$assump$leveneTable	a table
results$contrasts	an array of tables
results$postHoc	an array of tables
results$emm	an array of the estimated marginal means plots + tables
results$groupSummary	a summary of the groups

Tables can be converted to data frames with asDF or as.data.frame(). For example:

results$rmTable$asDF

as.data.frame(results$rmTable)

Examples

data('bugs', package = 'jmv')

anovaRM(
    data = bugs,
    rm = list(
        list(
            label = 'Frightening',
            levels = c('Low', 'High'))),
    rmCells = list(
        list(
            measure = 'LDLF',
            cell = 'Low'),
        list(
            measure = 'LDHF',
            cell = 'High')),
    rmTerms = list(
        'Frightening'))

#
#  REPEATED MEASURES ANOVA
#
#  Within Subjects Effects
#  -----------------------------------------------------------------------
#                  Sum of Squares    df    Mean Square    F       p
#  -----------------------------------------------------------------------
#    Frightening              126     1         126.11    44.2    < .001
#    Residual                 257    90           2.85
#  -----------------------------------------------------------------------
#    Note. Type 3 Sums of Squares
#
#
#
#  Between Subjects Effects
#  -----------------------------------------------------------------
#                Sum of Squares    df    Mean Square    F    p
#  -----------------------------------------------------------------
#    Residual               954    90           10.6
#  -----------------------------------------------------------------
#    Note. Type 3 Sums of Squares
#