Includes data analysis functions (e.g., to calculate effect sizes and 95% Confidence Intervals (CI) on Standardised Effect Sizes (d) for ABBA cross-over repeated-measures experimental designs), data presentation functions (e.g., density curve overlaid on histogram), and the data sets analyzed in different research papers in software engineering (e.g., related to software defect prediction or multi-site experiment concerning the extent to which structured abstracts were clearer and more complete than conventional abstracts) to streamline reproducible research in software engineering.
The R package reproducer is aimed to support reproducible research in software engineering. See the package homepage for details and examples.
One may install the stable version from CRAN:
install.packages('reproducer', dependencies = TRUE)You can use devtools to install the development version from my web site:
install.packages("devtools", dependencies = T, repos = "http://cran.r-project.org/")library(devtools)devtools::install_url("http://madeyski.e-informatyka.pl/download/R/reproducer_0.1.10.tar.gz")library(reproducer)The motivation is to support reproducible research in software engineering via sharing data sets and code behind the published or just submitted papers.
calculateSmallSampleSizeAdjustment,
constructEffectSizes,
transformRtoZr,
transformZrtoR,
transformHgtoR,
calculateHg,
transformRtoHg,
transformZrtoHgapprox,
transformZrtoHg,
PrepareForMetaAnalysisGtoR,
ExtractMAStatistics,
aggregateIndividualDocumentStatistics,
reproduceTablesOfPaperMetaAnalysisForFamiliesOfExperiments.KitchenhamMadeyskiBrereton.MetaAnalysisReportedResults, KitchenhamMadeyskiBrereton.ABBAMetaAnalysisReportedResults, KitchenhamMadeyskiBrereton.ReportedEffectSizes, KitchenhamMadeyskiBrereton.ABBAReportedEffectSizes KitchenhamMadeyskiBrereton.ExpData, and KitchenhamMadeyskiBrereton.DocDataMadeyskiKitchenham.EUBASdata and functions getEffectSizesABBA, effectSizeCIgetTheoreticalEffectSizeVariancesABBAgetSimulationData, plotOutcomesForIndividualsInEachSequenceGroup, getEffectSizesABBA, effectSizeCIeffectSizeCI to calculate 95% Confidence Intervals (CI) on Standardised Effect Sizes (d) for cross-over repeated-measures designsreproduceSimulationResultsBasedOn500Reps1000Obs function (we agreed to write joint paper with Dr Curtin describing corrections to his equations to calculate effect size variances for continuous outcomes of cross-over clinical trials)getSimulationDataplotOutcomesForIndividualsInEachSequenceGroupgetEffectSizesABBAgetEffectSizesABBAIgnoringPeriodEffectreproduceSimulationResultsBasedOn500Reps1000ObspercentageInaccuracyOfLargeSampleVarianceApproximationproportionOfSignificantTValuesUsingCorrectAnalysisproportionOfSignificantTValuesUsingIncorrectAnalysisKitchenhamMadeyski.SimulatedCrossoverDataSets backed by functions (varianceSimulation, getSimulatedCrossoverDataSets) to reproduce the data set.cloudOfWordsKitchenhamMadeyskiBudgen16.FINNISHKitchenhamMadeyskiBudgen16.PolishSubjectsKitchenhamMadeyskiBudgen16.SubjectDataKitchenhamMadeyskiBudgen16.PolishDataKitchenhamMadeyskiBudgen16.DiffInDiffDataKitchenhamMadeyskiBudgen16.COCOMOdensityCurveOnHistogramboxplotHVboxplotAndDensityCurveOnHistogramprintXTablecloudOfWordsreproduceForestPlotRandomEffectsreproduceMixedEffectsAnalysisWithEstimatedVarianceAndExperimentalDesignModeratorreproduceMixedEffectsAnalysisWithExperimentalDesignModeratorreproduceMixedEffectsForestPlotWithExperimentalDesignModeratorreproduceTableWithEffectSizesBasedOnMeanDifferencesreproduceTableWithPossibleModeratingFactorsreproduceTableWithSourceDataByCiolkowskiCiolkowski09ESEM.MetaAnalysis.PBRvsCBRorARMadeyskiKitchenham.MetaAnalysis.PBRvsCBRorARMadeyski15EISEJ.StudProjects$STUD data setMadeyski15SQJ.NDCMadeyski15EISEJ.OpenProjectsMadeyski15EISEJ.PropProjectsMadeyski15EISEJ.StudProjects
and functions (for importing data, visualization and descriptive analyses):readExcelSheetdensityCurveOnHistogramboxplotHVboxplotAndDensityCurveOnHistogramSee the package homepage (http://madeyski.e-informatyka.pl/reproducible-research/) for documentation and examples.