Traditional

Outlier Detection

Mahalanobis Distance

Graphical Methods of Outlier Detection

One Sample Hypothesis Testing

Independent Groups
Hotelling's T2

Dependent Groups (Repeated Measures)

Hotelling's T2

Two Sample Hypothesis Testing

Independent Groups

MANOVA

Independent Groups

MANOVA

Multivariate Regression
Independent Groups

Multivariate Regression

Principal Components

Independent Groups

Principal Component Analysis

Data Management

Matrix Splitting
tlier Detection

esting
 easures)

Testing
Complete Cases

With Missing Values

2 Groups

>2 Groups and stores in list mode

626-630
 638-641

hotel1(x, null.value=0, tr=0) 642-645

manova(y~x) 648-649
model=lm(y~x) ; summary.aov(model) ; 648-649

prcomp(x, cor=F) 666-670

regpca(x, cor=T, SCORES=F, scree=T, xlab="Principal Component", 666-670

ylab="Proportion of Variance")

matsplit(m, coln) 646

mat2grp(m, coln) 646

Robust

Minimum Volume Elipsoid (MVE)

Minimum Covariance Determinant

(MCD)

Median Ball Algorithm (RMBA)

Rocke's Translated Biweight S-

Estimator (TBS)
Orthogonal Gnanadesikan-Kettenring
(OGK) Estimator

Wilcox Outlier Function Multiple Purpose Outlier Detection: MVE, MCD, RMBA, TBS, OGK

Projection Methods Out Pro (# Variables <9)

Out Pro (# Variables >9)

OP Estimators Skipped Mean

3D Graphs 3D Outlier Detection using Out Pro and Theil-Sen Regression Estimator

3D Graphs 3D Outlier Detection using Out Pro and Theil-Sen Regression Estimator

Independent Groups
Dependent Groups (Repeated Measures)

OP Estimators Bootstraped Skipped Mean (Tests against specified Null Value)

Bootstrap Test of Pairwise Differences

Independent Groups

OP Estimators Bootstraped Skipped Mean (Tests against specified Null Value)

Rank Based Methods Generalized Wilcoxon-Mann-Whitney (Marginal Medians)

Generalized Wilcoxon-Mann-Whitney (Donoho-Gasko Medians)

Generalized Wilcoxon-Mann-Whitney (MCD)

Independent Groups

Trimmed Means MANOVA with Trimmed Means (No P Value)

MANOVA with Trimmed Means (P Value)

Munzel-Brunner Method
Rank Based Methods Rank Based (Midranks): Heteroscedasticity with Tied Values

Choi-Marden Multivariate Rank Test

Independent Groups

Rousseeuw Multivariate Regression Takes into account dependence among outcome variables. Good with
Estimator Heteroscedasticity

Bootstraped Rousseeuw Multivariate Regression Estimator

Theil-Sen Multivariate Regression

Independent Groups

Outlier Removed PCA Eliminates Outliers using Outpro

Huebert PCA

Detailed Huebert PCA with Eigenvalues

Maximizing Robust General Variance

Used to choose number of components

cov.mve(m) 630-632

covmve(m) 630-632

cov.mcd(m) 630-632

covmcd(m) 630-632

rmba(m) 633

tbs(m) 633

ogk(m) 633

out(m, x, cov.fun=cov.mve, plotit=T, SEED=T, xlab= "X", ylab="Y", crit=NULL) 633

outpro(m, gval=NA, center=NA, plotit=T, op=T, MM=F, cop=3, xlab="VAR1", 634-637

ylab="VAR2")
outproad(m, gval=NA, center=NA, plotit=T, op=T, MM=F, cop=3, xlab="VAR1", 634-637
ylab="VAR2", rate=0.05)

smean(x, MM=F, MC=F) 638

out3d(x, outfun=outpro, xlab="VAR1", ylab="VAR2", zlab="VAR3", reg.plane=F, 634-637

regfun=tsreg)

out3d(x, outfun=outpro, xlab="VAR1", ylab="VAR2", zlab="VAR3", reg.plane=F, 634-637

regfun=tsreg)
smeancrv2(m, nullv=rep(0, ncol(m)), nboot=500, plotit=T, MC=F, xlab="VAR1", 642-645
ylab="VAR2")

rmdzeroOP(m, nullv=rep(0, ncol(m)), nboot=500, MC=F) 642-645

smean2(m1, m2, nullv=rep(0, ncol(m1)), nboot=500, plotit=T) 645-646

mulwmwv2(m1, m2, plotit=T, cop=3, alpha=0.05, nboot=1000, pop=4) 652-655

mulwmwv2(m1, m2, plotit=T, cop=1, alpha=0.05, nboot=1000, pop=4) 652-655

mulwmwv2(m1, m2, plotit=T, cop=2, alpha=0.05, nboot=1000, pop=4) 652-655

MULtr.anova(x, J=NULL, p=NULL, tr=0.2, alpha=0.05) 650-652

MULAOVp(x, J=NULL, p=NULL, tr=0.2) 650-652

mulrank(J, K, x) 656-658

cmanova(J, K, x) 658-660
mlrreg(x, y) 662-664

mlrreg.Stest(x, y, nboot=100, SEED=T) 662-664

mopreg(x, y, regfun=tsreg ,cop=3, KEEP=T, MC=F) 662-664

outpca(x, cor=T, SCORES=F, ADJ=F, scree=T, xlab="Principal Component", 671-674

ylab="Proportion of Variance")

robpcaS(x, SCORES=F) 671-674

robpca(x, scree=T, xlab="Principal Component", ylab="Proportion of Variance") 671-674

Ppca(x, p=ncol(x)-1, SCORES=F, SCALE=T, gcov=rmba, MC=F) 671-674

Ppca.summary(x, MC=F, SCALE=T) 671-674

Traditional

Measures of Location

mean

median

Functions for Variance

standard deviation

standard error

variance
Binomial Confidence Interval
mean(x) 26

median(x) 24

sd(x) 30

std,error(x)

var(x) 30
binomci() 167
Robust

Trimmed Mean
Bootstrapped T Trimmed Mean Confidence Interval

Percentile Bootstrap Trimmed Mean Confidence Interval

Population Trimmed Mean Confidence Interval

Population Median Confidence Interval Hettmansperger & Sheather

Windsorized Mean

Standard Error Trimmed Mean

Standard Error of Median McKean-Schrader

Bootstrap Standard Error

Windsorized Variance

Median Absolute Deviation

Interquartile Range

Biweight Midvariance

Percentage Bend Midvariance

Tau Measure of Variation

Translated Biweight S
Agresti-Coull Binomial CI Agresti-Coull
tmean(x, tr=0.2) 26

trimcibt(x, tr=0.2, alpha=0.05, nboot=599, side=T) 281-282

trimpb(x, tr=0.2, alpha=0.05, nboot=2000) 278-279

trimci(x, tr=0.2, alpha=0.05) 153-159, 206

sint(x, alpha=0.05) 159-162

winmean(x, tr=0.2) 26-28

trimse(x, tr=0.2) 152-153

162
msmedse(x)
287
bootse(x, nboot=1000, est=onestep)
winvar(x, tr=0.2) 32

mad(x) 32-33

idealf(x) 30-32

bivar(x) 34-35

pbvar(x) 34-35

tauvar(x) 34-36
tbsvar(x) 34-36

acbinomci() 167
 Traditional

Univariate Outlier Detection

Mean Deviation

Multivaruate Outlier Detection

Mahalanobis Distance
Measures of Location After Outlier Detection
|𝑋−¯𝑋|/�>2 37

626-630
 Robust

Mad Median

Box Plot

Carling Box Plot Modification

Minimum Volume Elipsoid (MVE)

Minimum Covariance Determinant (MCD)

Median Ball Algorithm (RMBA)

Rocke's Translated Biweight S-Estimator (TBS)

Orthogonal Gnanadesikan-Kettenring (OGK) Estimator

Multiple Purpose Outlier Detection: MVE, MCD,

Wilcox Outlier Function RMBA, TBS, OGK

Projection Methods Out Pro (# Variables <9)

Out Pro (# Variables >9)

OP Estimators Skipped Mean

3D Outlier Detection using Out Pro and Theil-Sen

3D Graphs Regression Estimator
 Uses a modified mad-median rule to declare
One-Step M-Estimator outiers and creates a mean (M estimator)

One-Step M CI (Bootstrap) M estimator

Modified One-Step M-Estimator M estimator

Modified One-Step M-Estimator CI (Bootstrap) M estimator

out(x) 38-39, 633

boxplot(x) 39-41

outbox(x, mbox=F) 41-43

cov.mve(m) 630-632

covmve(m) 630-632

cov.mcd(m) 630-632

covmcd(m) 630-632

rmba(m) 633

tbs(m) 633

ogk(m) 633

out(m, x, cov.fun=cov.mve, plotit=T, SEED=T, xlab= "X",

ylab="Y", crit=NULL) 633

outpro(m, gval=NA, center=NA, plotit=T, op=T, MM=F, cop=3, 634-637

xlab="VAR1", ylab="VAR2")
outproad(m, gval=NA, center=NA, plotit=T, op=T, MM=F, 634-637
cop=3, xlab="VAR1", ylab="VAR2", rate=0.05)

smean(x, MM=F, MC=F) 638

out3d(x, outfun=outpro, xlab="VAR1", ylab="VAR2", 634-637

zlab="VAR3", reg.plane=F, regfun=tsreg)
onestep(x) 44-45

onesampb(x, est=onestep, alpha=0.05, nboot=2000) 278-279

mom(x) 44-45

momci(x, alpha=0.05, nboot=2000) 278-279

 Traditional

Graphical Summaries of Data

Histogram

Relative Frequency Histogram

Stem and Leaf

Error Bar Graphs

Distributions of Both Groups

Boxplots

4-Plot
 hist(x)

splot(x)

stem(x)

ebarplot(x, y=NULL, nse=1, xlab="Groups", yl

g2plot(x, y, op=4, xlab="X", ylab="")

boxplot(x, y)

Error bars, kernal density, boxplot, shift function sumplot2g(x, y)

 Robust

46-54
Kernal Densirty Estimator
46-54
Adaptive Kernal Densirty Estimator
57-58

363-367

371

371

371
 54-56
kdplot(x)
54-56
akerd(x)
 Traditional

Probability

Normal Probability Function

T Distribution

T Quantiles
 Robust

prnom(q, mean=0, sd=1)

pt(q, df)

qt(p, df)
 Traditional

T-test and Analogs

Independent Samples

One-Sample T

Two-Sample T var.equal=False (Welch Method)

Permutation Test Tests equality of Distribution

Dependent Samples

Paired T-Test
Graphical Methods of Two-Group Comparisons

Independent Groups

Error Bar Graphs

Distributions of Both Groups

Boxplots
Error bars, kernal density, boxplot, shift
4-Plot function
Comparing Variance

Power and Effect Size-2 Group Comparisons

Ttest

Stein Method
 127-128, 182
t.test(x, mu=0, conf.level=.95)
325
t.test(x, y, var.equal=FALSE, conf.level=0.95)

345-348
permg(x, y, alpha=0.05, est=mean, nboot=1000)

403-404
t.test(x, y , paired=TRUE)
 363-367
ebarplot(x, y=NULL, nse=1, xlab="Groups", ylab=NULL, tr=0)
371
g2plot(x, y, op=4, xlab="X", ylab="")
371
boxplot(x, y)
371
sumplot2g(x, y)
power.t.test(n=NULL, delta=NULL, sd=1, sig.level=0.05, power=NULL,
type=c("two.sided", "paired", "one.sample"), alternative=c("two.sided", 195-196, 317
"one.sided")
stein1(x, del, alpha=0.05, pow=0.8, oneside=F) 196-199

stein2(x1, x2, mu0=0, alpha=0.05) 196-199

 Robust

Independent Samples

One Sample Medians

Two-Sample T with Trimmed Means

Two-Sample Median Comparison

Two-Sample M-Estimator Comparison

Rank Based Methods

Dependent Samples

Trimmed Means
Bootstrap T

Percentile Bootstrap

Rank Based Methods

M-Estimator

Independent Groups

Quantile Plots (Shift Function)

Dependent Groups

Difference Score Plots

Adaptive Kernal Density Estimation

Independent Groups

Bootstrap Method

Bootstrap test of Population Variance

Difference in Percentage Bend Midvariance

Dependent Groups

HC4-Test

Percentile Bootstrap Method

All Quantiles

Stein Method Trimmed Means

Bootstrap Trimmed Means

Yuen's Method

Timmed Means Effect Size

Yuen's Method

Percentile Bootstrap of Trimmed Means

Yuen's Bootstrap T of Trimmed Means

Guo-Luh Bootstrap T of Trimmed Means (Good for trim <0.2)

Compares Medians using McKean-Schrader estimate of SE

Percentile Bootstrap of Medians

Percentile Bootstrap of Trimmed Means using a One-Step M-Estimator

Percentile Bootstrap of Trimmed Medians using a One-Step M-Estimator

Wilcoxon-Mann-Whitney: Rank Based Non-parametric test

Hodges, Ramsey, & Wechsler P value Computation

Cliff's Method: Rank Based: Heteroscedasticity with Tied Values

P value for Cliff's Method

Brunner-Munzel Method: Rank Based (Midranks): Heteroscedasticity with Tied Values

Kolmogorov-Smirnov Test: Tests Identical Distributions (low power with tied values)
Doksum & Sievers K-S Extension (Shift Function): Compares ALL Quantiles
Simultaneously

Trimmed Means
Trimmed Means Bootstrap T

Yuen's Bootstrap T of Trimmed Means

Bootstrap Test of Medians

Lin & Stivers Bootstrap T for Missing Data

Trimmed Means Percentile Bootstrap

Trimmed Means Percentile Bootstrap for Missing Data

Medians

Comparing Medians

Sign Test

Wilcoxon Sign Rank Test

M-Estimator

Plots of Quantiles between Independent Groups

Bootstrap two group comparison of variance

With Delta Specified

With Range of Deltas

Reports Effect Size

sintv2(x, alpha=0.05, nullval=0)

yuen(x, y, alpha=0.05, tr=0.2) 330-331

trimpb2(x, y, re=0.2, alpha=0.05, nboot=2000) 336-338

yuenbt(x, y, tr=0.2, alpha=0.05, nboot=599, side=F) 341-344

yht(x, y, tr=0.15, alpha=0.05, nboot=600, PV=F) 343-344

msmed(x, y, alpha=0.05) 331-332

medpb2(x, y, alpha=0.05, nboot=2000, SEED=T) 336-337

pb2gen(x, y, alpha=0.05, nboot=2000, est=mean, tr=0.2) 336-338

pb2gen(x, y, alpha=0.05, nboot=2000, est=median, tr=0.2) 336-338

349-351
wilcox.test(x, y)
349-351
wmw(x, y)
352-255
cid(x, y, alpha=0.05, plotit=F)
352-255
cidv2(x, y, plotit=F)
355-359
bmp(x, y, alpha=0.05)
359-360
ks(x, y, sig=T, alpha=0.05)
361-363
sband(x, y, flag=F, plotit=T, xlab="x (First Group)", ylab="Delta")

trimci() 405
trimcibt() 405

ydbt(x, y, tr=0.2) 406-407

loc2dif(x, y=NULL, est=median, na.rm=T, plotit=F, xlab="", ylab="", …) 416

rm2miss(x, y, tr=0) 415

trimpb() 405

rmmismcp(x, y, est=tmean) 415

sintv2() 405

dmedpb(x, y) 406

signt(x, y=NA, alpha=0.05) 417-418

wilcox.test(x, y, paired=F, exact=T) 418-420

mestci() 406

367-370
sband(x, y, flag=F, plotit=T, xlab="x (First Group)", ylab="Delta")

424
akerd(x[,1]-x[,2])
425
g2plotdifxy(x, y)
 374-375
comvar(x,y)
376-377
comvar2(x, y)
376-377
pb2gen(x, y, est=pbvar)

421
pcorhc4(x1~x2, x1+x2)
422
rmrvar(x, y=NA, alpha=0.05, est=pbvar, plotit=F)
423
lband(x, y=NA, alpha=0.05, plotit=T)

stein1.tr(x, del, alpha=0.05, pow=0.8, tr=0.2) 208

stein2.tr(x, y, alpha=0.05, tr=0.2) 208

283-286
powt1est(x, delta=0, ci=F, nboot=800)
283-286, 344
powt1ant(x, ci=F, plotit=T, nboot=800)
384-385
yuenv2(x, y)
384-385
akp.effect(x, y, EQVAR=T, tr=0.2)
Traditional

ANOVA

Independent Groups

ANOVA
Dependent Groups (Repeated Measures)

Random Effects ANOVA

Graphical Methods of Comparisons

Interaction Plots

Power and Effect Size-Multiple Group Comparisons

ANOVA

Multiple Comparisons(Controlling FWE)

Independent Groups

Fisher's Least Significant Difference

Tukey-Kramer Honestly Significance

Difference

Bonferronni

Rom's Method

Hochberg's Method

Benjamini and Hochberg Method

Method SR

Scheffe's Homoscedastic Method

Kaiser-Bowden Method

Welch-Sidak Method
Interaction

Interaction Plots
When predictor is factor

When data is a matrix or list

When data is a data frame or list

easures)

mparisons
ltiple Group Comparisons

ntrolling FWE)

Fisher LSD: Normality, Homoscedasticity, and J<=3

Tukey HSD: Normality, Homoscedasticity, and J>3

α/C

Rank Based. Similar to Hochberg, but looks up values for k<10 in

table.

dk=α/k

Pk=((C-k+1)α) /C

Rank Based. Similar to Hochberg, but looks up values for k<10 in

table.

Two-Way (Btwn-by-Within)
 434-436, 501,
aov(x~g) 521

anova1(x) 436

anova(x) 437, 501, 521

rananova(x, tr=0.0, grp=NA) 470-472

aov(dv ~ myfactor + Error(subject/myfactor), data=mydata) 530-531

interaction.plot(x.factor, trace.factor, response, fun=mean,
trace.label=deparse(substitute(trace.factor)), 494-495
xlab=deparse(substitute(x.factor)), ylab=ylabel)

anova.power(groups=NULL, n=NULL, delta=NULL, sig.level=0.05, 439-441

power=NULL)
power.anova.test(groups=NULL, n=NULL, between.var=NULL, 439-441
within.var=NULL, sig.level=0.05, power=NULL)

556-568

TukeyHSD(x, conf.level=0.95) 568-570

578

578-579

579

579-581

582

588

kbcon(x, con=0, tr=0.0, alpha=0.05) 589-597

589-597
interaction.plot(x.factor, trace.factor, response, fun=mean,
trace.label=deparse(substitute(trace.factor)), 494-495
xlab=deparse(substitute(x.factor)), ylab=ylabel)
Robust

Independent Groups

Trimmed-Means

Medians

Bootstrap T with Trimmed Means

Percentile Bootstrap Methods

Rank Based Methods

M-Estimators

Dependent Groups (Repeated Measures)

Trimmed-Means

Bootstrap T with Trimmed Means

Percentile Bootstrap Methods

Rank Based Methods

Bishop-Dudewicz

Tamhane's Procedure

Hochberg's Procedure

Independent Groups

Dunnett's T3

Games-Howell

Create Contrast Coefficients

Trimmed Means

Bootstrap T

Percentile Bootstrap Methods

Rank Based Methods

Medians

M-Estimators

Dependent Groups (Repeated Measures)

Trimmed Means

Bootstrap T

Percentile Bootstrap Methods

M-Estimators

Rank Based Methods

One-Way

One-Way: Reports the measure of effect size

One-Way: Same as t1way. Where x is a column of data

Two-Way

Two-Way: Percentile Bootstrap

Three-Way

One-Way

Two-Way: Method SHPB

One-Way

Two-Way

One-Way: Method SHPB (M estimator)

One-Way: Method LSPB (M estimator)

Two-Way: Method LSPB (M estimator)

Two-Way: Trimmed Means

One-Way: Kruskall-Wallis Test

One-Way: Method BDM-Rank Based ANOVA

Two-Way: Method BDM-Rank Based ANOVA

Two-Way: Patel-Hoel Interaction Test

 Full Two-Way with Contrasts (M estimator)

ed Measures)

One-Way

One-Way

Two-Way (Btwn-by-Within)

Two-Way (Within-by-Within)

Three-Way (Btwn-by-Btwn-by-Within)

Three-Way (Btwn-by-Within-by-Within)

Three-Way (Within-by-Within-by-Within)

One-Way

Two-Way (Btwn-by-Within)

One-Way: Method RMPB3 (M estimator)

One-Way: Difference Scores (M estimator)

Two-Way (Btwn-by-Within): Test Factor A (M estimator)

Two-Way (Btwn-by-Within): Test Factor B (M estimator)

Two-Way (Btwn-by-Within): Test Interaction A*B (M estimator)

One-Way: Freidman's Test

One-Way: Method BPRM

Two-Way (Btwn-by-Within):
Two Stage Method

Two Stage Method (Controls FWE)

Two Stage Method (Controls FWE) with Trimming

Means, Heteroscedasticity. Applies Welch's Method. Uses Studentized

Maximum Modulus Distribution for Critical Value

Same as T3. Uses Studentized Range for Critical Value

Two Way: Create Contrast Coefficients

Three-Way: Create Contrast Coefficients

Two-Way: Dunnett's T3 with Trimmed Means

Two-Way: Kaiser-Bowden Method with Trimmed Means

Two-Way: Bootstrap T with Trimmed Means

Two-Way: Trimmed Mean Comparison (20% Trim)

Two-Way: Trimmed Means

Two-Way: Median Comparison

Two-Way: M-Estimator Comparison (M estimator)

 Three-Way: Trimmed Means

Two-Way: Cliff's Method: Heteroscedastic with Tied Values

Two-Way: Cliff's Method AND Hochberg's Method

Two-Way: Brunner-Munzel Method: Heteroscedastic with Tied Values

Two-Way: Patel-Hoel Interaction Method

Two-Way: Percentile Bootstrap Method

Two-Way: McKean-Schrader

Two-Way: All pairwise contrasts and interactions

Two-Way: M-Estimator Comparison (M estimator)

Two-Way: Full Two-Way with Contrasts (M estimator)

ed Measures)

Two-Way: Rom's Method

Three-Way: Rom's Method (Within-Within-Within)

Two-Way: Trimmed Means (Within-Within)

Two-Way: Method BWMCP (Btwn-by-Within): Overall

Two-Way: Method BWAMCP (Btwn-by-Within): Factor A

Two-Way: Method BWBMCP (Btwn-by-Within): Factor B

Two-Way: Method BWIMCP (Btwn-by-Within): Interaction

Three-Way: Method BBWMCP (Btwn-Btwn-Within)

Three-Way: Method BWWMCP (Btwn-Within-Within)

Two-Way: Method SPMCP (Btwn-by-Within): Overall

Two-Way: Method SPMCPA (Btwn-by-Within): Factor A

Two-Way: Method SPMCPB (Btwn-by-Within): Factor B

Two-Way: Method SPMCPI (Btwn-by-Within): Interaction

Three-Way: Method BBWMCP (Btwn-Btwn-Within)

Three-Way: Method BWWMCP (Btwn-Within-Within)

Three-Way: Method WWWMCP (Within-Within-Within)

Two-Way: Rom's Method with Percentile Bootstrap (M estimator) (Within-

Within)
Two-Way: Hochberg's Method with Percentile Bootstrap (M estimator)
(Within-Within)
Two-Way: Method SR with Percentile Bootstrap (M estimator) (Within-
Within)

Patel-Hoel Interaction Method (Controls FWE)

t1way(x, tr=0.2, grp=NA) 450-455

t1wayv2(x, tr=0.2, grp=NA) 450-455

t1wayF(x, fac, tr=0.2, nboot=100, SEED=T) 450-455

t2way(J, K, x, tr=0.2, grp=c(1:p), p=J*K) 505-509

bbmcppb(J, K, x, tr=0.2, JK=J*K, alpha=0.05, grp=c(1:JK), nboot=500, bhop=F, SEED=T) 599-600

t3way(J, K, L, x, tr=0.2, grp=c(1:p)) 521-522

med1way(x, grp=NA) 454-455

m2way(J, K, x, alpha=0.05) 457-462, 514-516

t1waybt(x, tr=0.2, alpha=0.05, grp=NA, nboot=599) 456-457

t2waybt(J, K, x, tr=0.2, grp=c(1:p), p=J*K, nboot=599, SEED=T) 509-514

b1way(x, est=onestep, alpha=0.05, nboot=599) 457-462

pbadepth(x, est=onestep, con=0, alpha=0.05, nboot=2000, grp=NA, op=1, MM=F, MC=F, 457-462
SEED=T, na.rm=F, …)

pbad2way(J, K, x, est=onestep, conall=T, alpha=0.05, nboot=2000, grp=NA, pro.dis=F,…) 509-514

bbmcppb(J, K, x, tr=0.2, JK=J*K, alpha=0.05, grp=c(1:JK), nboot=500, bhop=F, SEED=T) 599-600

kruskal.test(x~g) 472-475

bdm(x) 475-477

bdm2way(J, K, x, alpha=0.05) 516-517

rimul(J, K, x, alpha=0.05, p=J*K, grp=c(1:p)) 517-518, 600

mcp2a(J, K, x, est=onestep, con=0, alpha=0.05, nboot=NA, grp=NA, …) 599

rananova(x, tr=0.2, grp=NA) 470-472

rmanova(x, tr=0.2, grp=c(1:length(x))) 531-533

bwtrim(J, K, x, tr=0.2, grp=c(1:p)) 546-550

wwtrim(J, K, x, grp=c(!:p), p=J*K, tr=.2, bop=F) 560-561

bbwtrim(J, K, L, x, grp=c(1:p), tr=0.2) 561-562

bwwtrim(J, K, L, x, grp=c(1:p), tr=0.2) 561-562

wwwtrim(J, K, L, x, grp=c(1:p), tr=0.2) 561-562

rmanovab(x, tr=0.2, alpha=0.05, grp=0, nboot=599) 535-536

tsplitbt(J, K, x, tr=0.2, alpha=0.05, grp=c(1:JK), nboot=599) 550-551

bd1way(x. est=onestep, nboot=599, alpha=0.05) 536-538

rmdzero(x, est=onestep, grp=NA, nboot=NA, ...) 538-541

sppba(J, K, x, est=onestep, grp=c(1:JK), avg=F, nboot=500, …) 552-554

sppbb(J, K, x, est=onestep, grp=c(1:JK), nboot=500, …) 552-554

sppbi(J, K, x, est=onestep, grp=c(1:JK), nboot=500, …) 552-554

friedman.test(x) 541-542

bprm(x) 543-544

bwrank(J, K, x) 554-560
bdanova1(x, alpha=0.05, power=0.9, delta=NA) 446-450

bdanova2(x1, x2, alpha=0.05, power=0.9, delta=NA) 446-450

tamhane(x, x2=NA, cil=NA, crit=NA) 601-603

hochberg(x, x2=NA, cil=NA, crit=NA, con=0, tr=0.2, alpha=0.05, iter=1000, SEED=T) 603-605

lincon(x, con=0, tr=0, alpha=0.05) 575-576

576-577

con2way(J, K) 594-597

con3way(J, K, L) 616-618

lincon(x, con=0, tr=0.2, alpha=0.05) 577-578

kbcon(x, con=0, tr=0.2, alpha=0.05) 592-597

linconb(x, con=0, tr=0.2, alpha=0.05, nboot=599) 583-584, 599

tmcppb(x, alpha=0.05, nboot=NA, grp=NA, est=tmedan, con=0, bhop=F, ...) 582-583

bbmcppb(J, K, x, tr=0.2, JK=J*K, alpha=0.05, grp=c(1:JK), nboot=500, bhop=F, SEED=T) 599-600

medpb(x, alpha=0.05, nboot=NA, grp=NA, est=median, con=0, bhop=F) 582-583

pbmcp(x, alpha=0.05, nboot=NA, grp=NA, est=onestep, con=0, bhop=F) 582-583

bbbmcppb(J, K, L, x, tr=0.2, JKL=J*K*L, alpha=0.05, grp=c(1:JKL), nboot=599, SEED=T, ...) 619

cidmul(x, alpha=0.05) 584-585

cidmulv2(x, alpha=0.05, g=NULL, dp=NULL, CI.FWE=F) 584-585

bmpmul(x, alpha=0.05) 584-585

rimul(J, K, x, alpha=0.05, p=J*K, grp=c(1:p)) 600

medpb(x, alpha=0.05, nboot=NA, grp=NA, est=median, con=0, bhop=F) 582-583

msmed(x, y=NA, con=0, alpha=0.05) 597-598

mcp2med(J, K, x, con=0, alpha=0.05, grp=NA, op=F) 597-598

pbmcp(x, alpha=0.05, nboot=NA, grp=NA, est=onestep, con=0, bhop=F) 582-583

mcp2a(J, K, x, est=onestep, con=0, alpha=0.05, nboot=NA, grp=NA, …) 599

rmmcp(x, con=0, tr=0.2, alpha=0.05, dif=T) 606-607, 615

rm3mcp(J, K, L, x, tr=0.2, alpha=0.05, dif=T, op=F, grp=NA) 618

bptd(x, tr=0.2, alpha=0.05, con=0, nboot=599) 609

bwmcp(J, K, x, tr=0.2, alpha=0.05, con=0, nboot=599) 609-614

bwamcp(J, K, x, tr=0.2, alpha=0.05) 609-614

bwbmcp(J, K, x, tr=0.2, alpha=0.05, dif=T, pool=F) 609-614

bwimcp(J, K, x, tr=0.2, alpha=0.05) 609-614

bbwmcp(J, K, L, x, tr=0.2, JKL=J*K*L, con=0, alpha=0.05, grp=c(1:JKL), nboot=599, 618

SEED=T, …)
bwwmcp(J, K, L, x, tr=0.2, JKL=J*K*L, con=0, alpha=0.05, grp=c(1:JKL), nboot=599, 619
SEED=T, …)

bwmcppb(J, K, x, tr=0.2, alpha=0.05, nboot=500, bhop=F) 609-615 (614)

spmcpa(J, K, x, tr=0.2, alpha=0.05) 609-614

spmcpb(J, K, x, tr=0.2, alpha=0.05, dif=T, pool=F) 609-614

spmcpi(J, K, x, tr=0.2, alpha=0.05) 609-614

bbwmcppb(J, K, L, x, tr=0.2, JKL=J*K*L, con=0, alpha=0.05, grp=c(1:JKL), nboot=599, 619

SEED=T, …)
bwwmcppb(J, K, L, x, tr=0.2, JKL=J*K*L, con=0, alpha=0.05, grp=c(1:JKL), nboot=599, 619
SEED=T, …)
wwwmcppb(J, K, L, x, tr=0.2, JKL=J*K*L, con=0, alpha=0.05, grp=c(1:JKL), nboot=599, 619
SEED=T, …)
rmmcppb(x, y=NA, alpha=0.05, con=0, est=onestep, plotit=T, dif=T, grp=NA, nboot=NA, 608-609
BA=F, hoch=F, xlab="Group 1", ylab="Group 2", pr=T, SEED=T, …)
rmmcppb(x, y=NA, alpha=0.05, con=0, est=onestep, plotit=T, dif=T, grp=NA, nboot=NA, 608-609
BA=F, hoch=T, xlab="Group 1", ylab="Group 2", pr=T, SEED=T, …)
rmmcppb(x, y=NA, alpha=0.05, con=0, est=onestep, plotit=T, dif=F, grp=NA, nboot=NA, 608-609
BA=F, hoch=F, xlab="Group 1", ylab="Group 2", pr=T, SEED=T, …)

rimul(J, K, x, alpha=0.05, p=J*K, grp=c(1:p)) 600

 Traditional

Comparing Binomials

Power and Effect Size-Binomial

Comparison of Two Proportions

power.prop.test(n=NULL, p1=NULL, p2=NULL, sig.level=0.05, power=NULL, alternative= 390
c("two.sided", "one.sided"), strict=FALSE)
 Robust

Storer-Kim Method

Beal's Method
twobinom(r1=sum(x), n1=length(x), r2=sum(y), n2=length(y), x=NA, y=NA) 389-390

twobici(r1=sum(x), n1=length(x), r2=sum(y), n2=length(y), x=NA, y=NA, alpha=0.05) 389-390

 Traditional

Regression
Linear Regresssion
Variable Selection

Stepwise Regression

Regression Hypothesis Testing

Graphical Smoothers

Detecting Curvature

Detecting Interactions
Plotting Interactions

Correlation
Pearson Correaltion

Comparing Depedent Correlations H0: XY=XZ

Graphical Methods of Regression Exploration

Test of Homoscedasticity

Koenker's Test
lm(y~x) 225
cor.test(x, y, conf.level=0.95) 251

TWOpov(x,y) 308
khomreg(x, y)
 Robust

Linear Regression with Outliers Removed (MAD)

Linear Regression with Heteroscedasticity

Wild Bootstrap Regression

Modified Bootstrap Regression

Bootstrap Test of Independence

Medians and Quantiles

Median Regression

Theil-Sen

Theil-Sen: Using Harrell-Davis

Least Median of Squares (LMS)

Least Trimmed Squares (LTS)

Least Trimmed Absolute Value (LTA)

Quantile Regression

M-Estimators

Coakley-Hettmansperger
Deepest Regression Line

Deepest Regression Line

Skipped Estimators

OP Regression Estimator

S-Estimators & E-Type Estimators

TSTS Estimator

Variable Selection

0.632 Bootstrap

Least Angle Regression

Method SA1

Method SA2

Method SA3

Comparing All Slopes

TS Reg Test

TS Reg Test-CI

Comparing 2 Independent Slopes

TS 2 Slope Test

Comparing >2 Independent Groups

Smoothed-ANCOVA with Trimmed Means

Quantile Smoothed-ANCOVA
Smoothed-ANCOVA with Trimmed Means and outlier
removal

% Bootstrap Smoothed-ANCOVA

Bootstrap T Smoothed-ANCOVA with Trimmed Means

Bootstrap Bagging Smoothed-ANCOVA

Cleveland's Smoother (lowess)

Advanced Lowess

Running Interval Smoother

Running Interval Smoother w/Bootstrap Bagging

Running Interval Smoother: Projection Based

Kernal Smoother

Constrained B-Spline Smoothing (COBS)

Smoother for Discrete X

Partial Residual Plot

Generalized Additive Model

Bootstrap Test of Independence

ADTEST
Two Way Interaction

Two Way Interaction with Trimmed Mean

Two Way Interaction using Theil-Sen

Regression Surface Plot

TS Regression Surface Plot

Correlation with Heteroscedasticity

OP Correlation

Op Correlation with Bootstrap CI

Polychoric

Polyserial

Rank Based Methods

Kendall's Tau

Spearman's Rho

Winsorized Rho

Comparing Two Independent Correlations

Modified Bootstrap Method

Least Squares Estimator

Heteroscedastic Wild Bootstrap

Medians and Quantiles

Theil-Sen Regression Scatter Plot

Wicox(?) Test
HC4-Test of all slopes = 0

HC4-Test of each slope = 0

HC4-Test of all slopes = 0

HC4-Test of each slope = 0

PM1-Bootstrap Method (Modified Percentile)

Detects Curvature

Can also be used to test Quartiles. Uses Bootstrap

Median of the Slopes (when xout=T, leverage points removed)

Uses Harrell-Davis estimate of medians

Minimizes the Median of Squared Residuals

With MAD-median rule applied for xout=T

Test of all slopes = 0

Test of each slope = 0

Uses the LTS estimator and adjusts the estimate based on Residuals
Similar to median regression, but provides protection against outliers

Remove outliers and fit regression line after

Removes outliers via projection method, then uses Theil-Sen estimator

1) Minimize % Bend Midvariance 2) Eliminate outlier residuals using MAD-Median

3) Use Theil-Sen

Estimates prediction error for all possible combinations of predictors (n ≤ 5)

chooses predictors with lowest prediction error. Good for large n of predictors

Uses Median-Ball variance estimation

Uses TS Estimator and Bootstap to compare strength of association between predictors

Test of all slopes = 0 using Theil-Sen regression estimator and Bootstrap CI

Test of each slopes = 0 using Theil-Sen regression estimator and Bootstrap CI

Tests Slope1=Slope2 using Theil-Sen regression estimator and Bootstrap CI

Tests Slope Grp 1 = Slope Grp 2 = ... using running interval smoother with 20% Trimming

Tests Slope Grp 1 = Slope Grp 2 = ... using Quantile running interval smoother
Tests Slope Grp 1 = Slope Grp 2 = ... using running interval smoother with 20% Trimming

Tests Slope Grp 1 = Slope Grp 2 = ... using running interval smoother with Percentile Bootstrap

Tests Slope Grp 1 = Slope Grp 2 = ... using running interval smoother with 20% Trimming and
Bootstrap T

Tests Slope Grp 1 = Slope Grp 2 = ... using running interval smoother with Bootstrap Baggin

Advanced Lowess plot with outlier removal and 3D plots.

with 20% Trimmed Mean

with 20% Trimmed Mean

with 20% Trimmed Mean

using Theil-Sen regression estimator

Detects Curvature

Compares General Additive Model with and without interaction

Y on X1 at Quartiles of X2

Y on X1 at given Values of X2

Y on X1 at given Values of X2

using OLS

using Theil-Sen

HC4

PM1-Bootstrap Method (Modified Percentile)

PM2-Bootstrap Method

Skipped correlation using projection method for outlier detection

Skipped Correlation using Bootstrap for

Ordinal-Ordinal

For correlation of Ordinal Variables

Pearson Correlation of Ranks

Scatter plot with Theil-Sen estimator of Slope

more power than Koenker. Uses a quantile regression approach.

ols(x, y, xout=T)

hc4test(x, y, xout=F, outfun=out)

olshc4(x, y, alpha=0.05, xout=F, outfun=out)

hc4wtest(x, y, nboot=500)

olswbtest(x, y, nboot=500)

lsfitci(x, y, nboot=599)

indt(x, y, nboot=500, alpha=0.05, flag=1)

medind(x, y, qval=0.5, nboot=1000, com.pval=F, alpha=0.05, xout=F)

tsreg(x, y, xout=F, outfun=out, iter=10, varfun=pbvar, corfun=pbcor, …)

tshdreg(x,y, ….)

lmsreg(x, y)

ltsreg(x, y)

ltareg(x, y, tr=0.2, h=NA)

rqfit(x, y, qval=.5, xout=T, outfun=out, res=F)

rqtest(x, y, qval=0.5, nboot=200, alpha=0.05)

qregci(x, y, nboot=100, alpha=0.05, qval=0.5)

chreg(x, y, xout=F)
mdepreg(x, y)

opreg(x, y)

tstsreg(x, y)

regpre(x, y, regfun=lsfit, error=absfun, nboot=100, model=NA)

larsR(x, y)

regpord(x, y)

ts2str(x, y)

sm2strv7(x, y, regfun=tsreg, nboot=500, alpha=0.05)

regtest(x, y, regfun=tsreg, nboot=600, alpha=0.05, plotit=T, grp=c(1:ncol(x)), nullvec=c(rep(0, length(grp))), xout=F,

outfun=out)

regci(x, y, regfun=tsreg, nboot=599, alpha=0.05, SEED=T, pr=T, xout=F, outfun=out, …)

reg2ci(x1, y1. x2, y2, regfun=tsreg, nboot=599, alpha=0.05, plotit=T)

ancsm(x1, y1, x2, y2, nboot=200, plotit=T, sm=F, tr=0.2)

Qancsm(x1, y1, x2, y2, nboot=200, qval=0.5, xlab="X", ylab="Y", plotit=T, xout=F, outfun=out, …)
ancova(x1, y1, x2, y2, fr1=1, fr2=1, tr=0.2, alpha=0.05, plotit=T, pts=NA, xout=T, outfun=out, …)

ancpb(x1, y1, x2, y2, est=onestep, pts=NA, fr1=1, fr2=1, nbppt=599, plotit=T, …)

ancboot(x1, y1, x2, y2, fr1=1, fr2=1, tr=0.2, nboot=599, pts=NA, plotit=T)

ancbbbpb(x1, y1, x2, y2, fr1=1, fr2=1, nboot=200, pts=NA, plotit=T, SEED=T, alpha=0.05)

lowess(x, y, p=2/3)
plot(x, y); lines(lowess(x,y)
lplot(x, y, span=0.75, pyhat=F, esout=F, xout=F, outfun=out, plotit=T, expand=0.5, low.span=2/3, varfun=pbvar, scale=F,
xlab="X", ylab="Y", zlab="", theta=50, phi=25)
rplot(x, y, est=tmean, scat=T, fr=NA, plotit=T, pyhat=F, efr=0.5, theta=50, phi=25, scale=F, expand=0.5, SEED=T,
varfun=pbvar, nmin=0, xout=F, outfun=out, eout=F)
rplotsm(x, y, est=tmean, nboot=40, atr=0, scat=T, fr=1, plotit=T, pyhat=F, efr=0.5, theta=50, phi=25, scale=F, expand=0.5,
SEED=T, varfun=pbvar, nmin=0, xout=F, outfun=out, pr=T)
runpd(x, y, est=tmean, fr=0.8, plotit=T, pyhat=F, nmin=0, theta=0, phi=25, expand=0.5, scale=F, xlab="X", ylab="Y", MC=F,
…)
kerreg(x, y, pyhat=F, pts=NA, plotit=T, theta=50, phi=25, expand=0.5, scale=F, zscale=F, eout=F, xout=F, outfun=out,
np=100, xlab="X", ylab="Y", zlab="", varfun=pbvar, e.pow=T)

qsmcobs(qval=0.5, xlab="X", ylab="Y", FIT=T, pc=".", plotit=T)

rundis(x, y, est=mom, plotit=T, pyhat=F, …)

prplot(x, y, pvec=ncol(x), regfun=tsreg, …)

indt(x, y, nboot=500, alpha=0.05, flag=1)

adtest(x, y, nboot=100, alpha=0.05, xout=F, outfun=out, SEED=T, …)

kercon(x, y, cval=NA, eout=F, xout=F, outfun=out=xlab="X", ylab="Y")

runsm2g(y, x1, x2, val=median(x2), est=tmean, sm=F, xlab="X", ylab="Y", …)

regi(y, x, z, val=median(z), est=onestep, regfun=tsreg, testit=F, …)

ols.plot.inter(x, y, pyhat=F, eout=F, xout=F, outfun=out, plotit=T, expand=0.5, scale=F, xlab="X", ylab="Y", zlab="",
theta=50, phi=25, family="gaussian", duplicate="error")

reg.plot.inter(x, y, regfun=tsreg, pyhat=F, eout=F, xout=F, outfun=out, plotit=T, expand=0.5, scale=F, xlab="X", ylab="Y",
zlab="", theta=50, phi=25, family="gaussian", duplicate="error")

pcorhc4(x, y)

lsfitci(x, y, nboot=599)

pcorb(x, y)

scor(x, y, plotit=T, xlab="VAR1", ylab="VAR2", MC=F)

corb(x, y, corfun=scor, plotit=F)

library(polycor); polychor(x)

cor.test(x, y, "kendall") OR cor(x,y, method="kendall")

cor(x,y, method="spearman")

wincor(x, y, tr=0.2)

twopcor(x1, y1, x2, y2)

two1sreg(x1, y1, x2, y2)

tworegwb(x1, y1, x2, y2)

regplot(x, y, regfun=tsreg, xlab="X", ylab="Y")

qhomt(x, y)
225

242

242

292

292

293

294-297

296

678-680

678-680

680-681

680-681

680-681

257

298

298

681-682
682

683

683-684

300-305

305-306

713-717

713-717

713-717

686-689

686-689

708-709

725-731

725-731
725-731

725-731

725-731

725-731

691-695

691-695

696-697

696-697

697-698

697-698

697-698

697-698

698-700

701

294-297

717-720
720-724

720-724

720-724

720-724

720-724

254

288-289

289-290

704-705

706

702-703

703

704

386-387

386-387
386-387

680

298-300