reproducer (version 0.5.3)

ConstructLevel1ExperimentRData: ConstructLevel1ExperimentRData

Description

This function returns the r value for a 2-group (2G) or 4-Group (4G) Crossover experiment for a group of 1 or more experiments for each sequence group and each outcome metric. For sets of 2 or more experiments, the experiments are assumed to be replicates and to report the same sets of Metrics and have the same Control treatment and use the same sequence Group identifiers, but are not necessarily the same Type. We return both the exact r value and the r value based on pooled variances for each sequence group and outcome metric.

Usage

ConstructLevel1ExperimentRData(
  Data,
  StudyID,
  ExperimentNames,
  Groups,
  Metrics,
  Type,
  Control
)

Value

R.Data.Table this is a tibble holding information identifying for each metric and sequence group the first time period and second time period variance, the pooled variance, the variance of the difference values and the exact r and pooled r.

Arguments

Data

This is a list parameter each entry in the list holds the data for each participant in a 2-group or 4-group crossover experiment in the 'wide' format. I.e., there is only one entry per participant. The data should have been generated from a long version of the data based on a variable labelled 'Period' which is used to define which participant data was collected in the first period of the experiment - see function ExtractLevel1ExperimentRData.

StudyID

This holds an identifier used to identify the origin of the experimental data in the output from this function.

ExperimentNames

This a list of identifiers used to define each experiment in the output from this function.

Groups

This is a list that defined the sequence group identifiers used in the dataset.

Metrics

This is a list of of character strings identifying each outcome metric reported in each of the experiments in the set of replicated experiments.

Type

this is a list of character strings specifying for each experiment whether the experiment is a 2-group or 4-group experiment

Control

this is a character string that defines the control treatment in the experiment.

Author

Barbara Kitchenham and Lech Madeyski

Examples

#
ShortExperimentNames <- c("E1", "E2", "E3", "E4")
FullExperimentNames <- c("EUBAS", "R1UCLM", "R2UCLM", "R3UCLM")
Metrics <- c("Comprehension", "Modification")
Groups <- Groups <- c("A", "B", "C", "D")
Type <- c(rep("4G", 4))
StudyID <- "S2"
Control <- "SC"
# Obtain experimental data from each file and put in wide format
dataset2 <- KitchenhamEtAl.CorrelationsAmongParticipants.Scanniello14TOSEM
ReshapedData <- ExtractExperimentData(dataset2,
  ExperimentNames = FullExperimentNames,
  idvar = "ParticipantID", timevar = "Period", ConvertToWide = TRUE
)
# Calculate the correlations for each sequence group and each metric in each experiment
ConstructLevel1ExperimentRData(
  Data = ReshapedData, StudyID = StudyID,
  ExperimentNames = ShortExperimentNames, Groups = Groups, Metrics = Metrics, Type = Type,
  Control = Control
)
# # A tibble: 32 x 15
# Study Exp   Group Metric Id        n ControlFirst    var1   var2    varp
#                       
#   1 S2    E1    A     Compr… S2E1A     6 FALSE        0.0183  0.0163 0.0173
# 2 S2    E1    B     Compr… S2E1B     6 TRUE         0.0201  0.0326 0.0263
# 3 S2    E1    C     Compr… S2E1C     6 FALSE        0.00370 0.0155 0.00962
# 4 S2    E1    D     Compr… S2E1D     6 TRUE         0.0173  0.0201 0.0187
# 5 S2    E1    A     Modif… S2E1A     6 FALSE        0.0527  0.0383 0.0455
# 6 S2    E1    B     Modif… S2E1B     6 TRUE         0.0185  0.0482 0.0333
# 7 S2    E1    C     Modif… S2E1C     6 FALSE        0.00655 0.0244 0.0155
# 8 S2    E1    D     Modif… S2E1D     6 TRUE         0.0222  0.0266 0.0244
# 9 S2    E2    A     Compr… S2E2A     6 FALSE        0.0194  0.0425 0.0309
# 10 S2    E2    B     Compr… S2E2B     6 TRUE         0.0198  0.0192 0.0195
# # … with 22 more rows, and 5 more variables: ControlVarProp ,
# #   VarProp , vardiff , r , r.p