Selecting Among 2-Sample Tests

Slides:



Advertisements
Similar presentations
One-way nonparametric ANOVA with trigonometric scores by Kravchuk, O.Y. School of Land and Food Sciences, University of Queensland.
Advertisements

Sampling: Final and Initial Sample Size Determination
Computing the ranks of data is only one of several possible so- called scoring methods that are in use... Section 2.7 reviews three of them – we’ll look.
Sampling Distributions (§ )
Reject H o Accept H o Accept H o Reject H o Accept H o Reject H o Left Tailed Right Tailed Two tailed
Nonparametric tests and ANOVAs: What you need to know.
Tests Jean-Yves Le Boudec. Contents 1.The Neyman Pearson framework 2.Likelihood Ratio Tests 3.ANOVA 4.Asymptotic Results 5.Other Tests 1.
Descriptive statistics Experiment  Data  Sample Statistics Sample mean Sample variance Normalize sample variance by N-1 Standard deviation goes as square-root.
Descriptive statistics Experiment  Data  Sample Statistics Experiment  Data  Sample Statistics Sample mean Sample mean Sample variance Sample variance.
Intro to Statistics for the Behavioral Sciences PSYC 1900 Lecture 9: Hypothesis Tests for Means: One Sample.
CSE 221: Probabilistic Analysis of Computer Systems Topics covered: Continuous random variables Uniform and Normal distribution (Sec. 3.1, )
Chapter 11: Inference for Distributions
15-1 Introduction Most of the hypothesis-testing and confidence interval procedures discussed in previous chapters are based on the assumption that.
1 (Student’s) T Distribution. 2 Z vs. T Many applications involve making conclusions about an unknown mean . Because a second unknown, , is present,
Normal and Sampling Distributions A normal distribution is uniquely determined by its mean, , and variance,  2 The random variable Z = (X-  /  is.
Choosing Statistical Procedures
The paired sample experiment The paired t test. Frequently one is interested in comparing the effects of two treatments (drugs, etc…) on a response variable.
Statistics for Engineer Week II and Week III: Random Variables and Probability Distribution.
(c) 2007 IUPUI SPEA K300 (4392) Outline Normal Probability Distribution Standard Normal Probability Distribution Standardization (Z-score) Illustrations.
Where are we?. What we have covered: - How to write a primary research paper.
Lesson Inferences about the Differences between Two Medians: Dependent Samples.
What are Nonparametric Statistics? In all of the preceding chapters we have focused on testing and estimating parameters associated with distributions.
Ordinally Scale Variables
Determination of Sample Size: A Review of Statistical Theory
Simulation Example: Generate a distribution for the random variate: What is the approximate probability that you will draw X ≤ 1.5?
Research Methods: 2 M.Sc. Physiotherapy/Podiatry/Pain Inferential Statistics.
Single-Factor Studies KNNL – Chapter 16. Single-Factor Models Independent Variable can be qualitative or quantitative If Quantitative, we typically assume.
IE 300, Fall 2012 Richard Sowers IESE. 8/30/2012 Goals: Rules of Probability Counting Equally likely Some examples.
NON-PARAMETRIC STATISTICS
t Test Assumptions: Paired/dependent/matched Samples:
Chapter 7: The Distribution of Sample Means. Frequency of Scores Scores Frequency.
Computing the ranks of data is only one of several possible so-called scoring methods that are in use... Section 2.7 reviews three of them – we’ll look.
Week 21 Order Statistics The order statistics of a set of random variables X 1, X 2,…, X n are the same random variables arranged in increasing order.
Week 21 Statistical Model A statistical model for some data is a set of distributions, one of which corresponds to the true unknown distribution that produced.
Nonparametric Statistics
Sampling and Sampling Distributions
University of Palestine
Correlation I have two variables, practically „equal“ (traditionally marked as X and Y) – I ask, if they are independent and if they are „correlated“,
STATISTICAL INFERENCE
IEE 380 Review.
Randomized Block Design
The greatest blessing in life is in giving and not taking.
Basic Practice of Statistics - 5th Edition
Behavioral Statistics
Two Sample Tests When do use independent
Sample Mean Distributions
Chapter 7: Sampling Distributions
Statistics in Applied Science and Technology
Lesson Inferences about the Differences between Two Medians: Dependent Samples.
Environmental Modeling Basic Testing Methods - Statistics
Georgi Iskrov, MBA, MPH, PhD Department of Social Medicine
Comparing Groups.
Single-Factor Studies
Sampling Distributions and the Central Limit Theorem
Single-Factor Studies
Some Nonparametric Methods
Nonparametrics.Zip (a compressed version of nonparamtrics) Tom Hettmansperger Department of Statistics, Penn State University References: Higgins (2004)
Assessing Normality.
Chapter 6 Confidence Intervals.
Estimation Goal: Use sample data to make predictions regarding unknown population parameters Point Estimate - Single value that is best guess of true parameter.
Tutorial 7 Consider the example discussed in the lecture, concerning the comparison of two teaching methods A and B, and let W denote the sum of the.
Chapter 7: The Distribution of Sample Means
Chapter 24 Comparing Two Means.
Sampling Distributions (§ )
Nonparametric Statistics
Distribution-Free Procedures
pairing data values (before-after, method1 vs
Statistical Inference
1/2555 สมศักดิ์ ศิวดำรงพงศ์
Introductory Statistics
Presentation transcript:

Selecting Among 2-Sample Tests Selection among tests is often based on the underlying distributions of the two populations. We often consider the ones below. Here x is a vector of values of the variable… uniform (density is dunif(x, min, max) where min and max are the endpoints of the uniform density. normal (density is dnorm(mean,sd) exponential (density is dexp(rate), where rate is the parameter of the exponential (so 1/rate = mean of the exponential distribution) Cauchy (density is dcauchy(location,scale) where location and scale are its parameters Laplace (density is (1/2b)(exp(-abs(x-a)/b) where -inf<x<+inf, and b>0 is the scale parameter and a is the location parameter) Write R code that will plot all these densities on the same axis for comparison, especially comparison of the tails of the distributions. “Fat tailed” underlying densities often call for nonparametric methods …

Assume Xi from treatment 1 and Yj from treatment 2 and that they have cdf's given by: Also assume the cdf's are continuous (no ties). The hypothesis tested is If the cdfs are normal with unknown but equal variances, then the t-test is the best. Departure from normality has little effect on Type I error, because of the Central Limit Theorem, but there are problems with power if they are non-normal. See Table 2.9.1 for a comparison of the Wilcoxon rank-sum test versus the t-test. Note that for small samples and "light-talied" distributions (those without much chance of outliers), the t-test is probably better. Otherwise, use Wilcoxon Another way of comparing these two tests is via relative efficiency (and ARE, asymptotic relative efficiency). See the definition on p. 62: basically, test 1 is more efficient than test 2 if it requires a smaller sample size to achieve the same power.

The ARE of the Wilcoxon test versus the t-test is >= The ARE of the Wilcoxon test versus the t-test is >= .864 but can be arbitrarily large (see Table 2.9.2). Even for normal cdfs, the ARE of Wilcoxon vs. t-test is .955 The permutation test vs. the Wilcoxon test and the t-test is discussed in 2.9.4. There are a couple things to note here: little is gained by doing more than 1600 randomly sampled permutations see Table 2.9.3 for comparison with t-test see Table 2.9.4 for comparison with Wilcoxon test