Chapter 13 Comparing Two Populations: Independent Samples.

Slides:



Advertisements
Similar presentations
The t Test for Two Independent Samples
Advertisements

1/2/2014 (c) 2001, Ron S. Kenett, Ph.D.1 Parametric Statistical Inference Instructor: Ron S. Kenett Course Website:
Introduction to Hypothesis Testing
C82MST Statistical Methods 2 - Lecture 2 1 Overview of Lecture Variability and Averages The Normal Distribution Comparing Population Variances Experimental.
Lecture 2 ANALYSIS OF VARIANCE: AN INTRODUCTION
Chapter 7 Sampling and Sampling Distributions
Chapter 10: The t Test For Two Independent Samples
INTRODUCTORY STATISTICS FOR CRIMINAL JUSTICE
Chi-Square and Analysis of Variance (ANOVA)
Hypothesis Tests: Two Independent Samples
Chapter 9 Introduction to the t-statistic
Comparing Two Population Parameters
Statistics for the Social Sciences Psychology 340 Spring 2005 Using t-tests.
Using t-tests Basic introduction and 1-sample t-tests Statistics for the Social Sciences Psychology 340 Spring 2010.
Statistics for the Social Sciences
Chi-square and F Distributions
Statistical Inferences Based on Two Samples
© The McGraw-Hill Companies, Inc., Chapter 10 Testing the Difference between Means and Variances.
Analysis of Variance Chapter 12 . McGraw-Hill/Irwin
Chapter Thirteen The One-Way Analysis of Variance.
Chapter 8 Estimation Understandable Statistics Ninth Edition
CHAPTER 15: Tests of Significance: The Basics Lecture PowerPoint Slides The Basic Practice of Statistics 6 th Edition Moore / Notz / Fligner.
Simple Linear Regression Analysis
T-tests continued.
Objective: To test claims about inferences for two proportions, under specific conditions Chapter 22.
Adapted by Peter Au, George Brown College McGraw-Hill Ryerson Copyright © 2011 McGraw-Hill Ryerson Limited.
Chapter 15 Comparing Two Populations: Dependent samples.
Testing means, part III The two-sample t-test. Sample Null hypothesis The population mean is equal to  o One-sample t-test Test statistic Null distribution.
PSY 307 – Statistics for the Behavioral Sciences
Lecture 8 PY 427 Statistics 1 Fall 2006 Kin Ching Kong, Ph.D
Statistics Are Fun! Analysis of Variance
Don’t spam class lists!!!. Farshad has prepared a suggested format for you final project. It will be on the web
BCOR 1020 Business Statistics Lecture 21 – April 8, 2008.
Hypothesis Testing Using The One-Sample t-Test
Chapter 9: Introduction to the t statistic
Chapter 10 The t Test for Two Independent Samples PSY295 Spring 2003 Summerfelt.
Hypothesis Testing and T-Tests. Hypothesis Tests Related to Differences Copyright © 2009 Pearson Education, Inc. Chapter Tests of Differences One.
Analysis of Variance. ANOVA Probably the most popular analysis in psychology Why? Ease of implementation Allows for analysis of several groups at once.
Psy B07 Chapter 1Slide 1 ANALYSIS OF VARIANCE. Psy B07 Chapter 1Slide 2 t-test refresher  In chapter 7 we talked about analyses that could be conducted.
Chapter 9 Hypothesis Testing and Estimation for Two Population Parameters.
COURSE: JUST 3900 TIPS FOR APLIA Developed By: Ethan Cooper (Lead Tutor) John Lohman Michael Mattocks Aubrey Urwick Chapter : 10 Independent Samples t.
One-sample In the previous cases we had one sample and were comparing its mean to a hypothesized population mean However in many situations we will use.
Hypothesis Testing Using the Two-Sample t-Test
Chapter 22: Comparing Two Proportions. Yet Another Standard Deviation (YASD) Standard deviation of the sampling distribution The variance of the sum or.
Chapter 17 Comparing Multiple Population Means: One-factor ANOVA.
Copyright (C) 2002 Houghton Mifflin Company. All rights reserved. 1 Understandable Statistics S eventh Edition By Brase and Brase Prepared by: Lynn Smith.
© Copyright McGraw-Hill 2004
- We have samples for each of two conditions. We provide an answer for “Are the two sample means significantly different from each other, or could both.
Chapter 10 Comparing Two Treatments Statistics, 5/E by Johnson and Bhattacharyya Copyright © 2006 by John Wiley & Sons, Inc. All rights reserved.
Chapter 10 Section 5 Chi-squared Test for a Variance or Standard Deviation.
 List the characteristics of the F distribution.  Conduct a test of hypothesis to determine whether the variances of two populations are equal.  Discuss.
Chapter 10: The t Test For Two Independent Samples.
Section Copyright © 2014, 2012, 2010 Pearson Education, Inc. Lecture Slides Elementary Statistics Twelfth Edition and the Triola Statistics Series.
Lecture Slides Elementary Statistics Twelfth Edition
Hypothesis Testing – Two Means(Small, Independent Samples)
Hypothesis Testing – Two Population Variances
STA 291 Spring 2010 Lecture 18 Dustin Lueker.
Math 4030 – 10b Inferences Concerning Variances: Hypothesis Testing
Lecture Slides Elementary Statistics Twelfth Edition
Math 4030 – 10a Tests for Population Mean(s)
Chapter 8 Hypothesis Testing with Two Samples.
Testing a Claim About a Mean:  Known
Hypothesis Tests for a Population Mean in Practice
Elementary Statistics
Chapter 10: The t Test For Two Independent Samples
Elementary Statistics: Picturing The World
Hypothesis Tests for a Standard Deviation
Hypothesis Testing: The Difference Between Two Population Means
Lecture Slides Elementary Statistics Twelfth Edition
Statistical Inference for the Mean: t-test
Presentation transcript:

Chapter 13 Comparing Two Populations: Independent Samples

Comparing more than 1 group Often psychologists are interested in comparing treatments, procedures, or conditions –Which drug is better in treating depression, Prozac or Zoloft? –Is the whole-language approach to teaching reading more effective than traditional methods?

A Research Study We are interested in the treatment of major depression Compare two drug therapies, Prozac and Zoloft Randomly select 16 people with major depression, 8 receive Prozac, 8 receive Zoloft

Measuring Depression Beck Depression Inventory (BDI) developed by Aaron Beck and his colleagues An “inventory” is a series of questions that are answered by the patient and the patient’s doctor Each answer contributes to an overall score That score is a “measure” of depression

Scores on the BDI Prozac Group Zoloft Group

Hypothesis test of Prozac vs. Zoloft 1. State and Check Assumptions –Normally distributed? - don’t know –σ? – don’t know –Interval data ? - probably –Independent Random sample? - yes

Hypothesis test of Prozac vs. Zoloft 2.Hypotheses H O : μ 1 = μ 2 (the effectiveness Prozac and Zoloft are the same) μ 1 - μ 2 = 0 (the difference between the effectiveness of Prozac and Zoloft is 0) H A : μ 1 ≠ μ 2 (the effectiveness of Prozac and Zoloft are not equal) μ 1 - μ 2 ≠ 0 (there is a difference between the effectiveness Prozac and Zoloft)

Hypothesis test of Prozac vs. Zoloft 3.Choose test statistic –parameter of interest - μ –2 groups independent samples –Not sure about Normal Distribution –Don’t know Population Standard Deviation

Hmm… What do we know about μ 1 – μ 2 ? What do we know about M 1 – M 2 ? Since we don’t know μ 1 or μ 2, we’ll concentrate on M 1 – M 2

Sampling Distribution The sampling distribution of M 1 – M 2 would help us predict values from random samples Three facts: –1. The mean of the M 1 – M 2 sampling distribution is equal to the mean of the sampling distribution of μ 1 – μ 2 –2. When the 2 populations have the same variance, then the standard deviation of the sampling distribution is –3. CLT

So… If we knew σ, we could transform the statistic M 1 – M 2 to a z score and use table A, but We don’t know σ But we know s 1 and s 2, that is, the standard deviations of the two samples Can we use them?

NO Not with a z, But we can use a t distribution That is to say: the differences in sample means, divided by the estimated SEM, is distributed as a t

t-test for 2 independent samples

Estimate of the Standard Error

Sampling Distribution The sampling distribution of M 1 – M 2 would help us predict values from random samples Three facts: –1. The mean of the M 1 – M 2 sampling distribution is equal to the mean of the sampling distribution of μ 1 – μ 2 –2. When the 2 populations have the same variance, then the standard deviation of the sampling distribution is –3. CLT

Hypothesis test of Prozac vs. Zoloft 1. State and Check Assumptions –Normally distributed? - don’t know –σ? – don’t know –Interval data ? - probably –Independent Random sample? – yes –Homogeneity of Variance (HoV): are the variances of the two population equal? – don’t know, but we’ll assume they are (can we check this out?)

Estimate of the Standard Error

More on the estimated SEM s 2 p is called “pooled variance” it is the variance of the two samples, put together, or pooled s 2 1 (n 1 -1) looks familiar, doesn’t it? (it’s variance times n-1)

SS(X 1 ), right? s 2 1 (n 1 -1) = SS(X 1 ) Thus:

df in a 2-sample t-test Since the calculation of each mean has n -1 degrees of freedom, then The 2-sample t-test has (n 1 -1) + (n 2 - 1) df, or df = n 1 + n 2 - 2

estimated SEM, again So, when we left the est SEM, we had: But, n 1 + n 2 – 2 = df, right? Thus:

Back to the hypothesis test 4.Set Significance Level α =.05 Critical Value Non-directional Hypothesis with df = n 1 + n = = 14 From Table C t crit = 2.145, so we reject H O if t ≤ or t ≥ 2.145

Hypothesis test of Prozac vs. Zoloft 5.Compute Statistic –We need:

Scores on the BDI Prozac Group Zoloft Group

Hypothesis test of Prozac vs. Zoloft 6. Draw Conclusions –because our t does not fall within the rejection region, we cannot reject the H O, and –conclude that we did not find any evidence that Prozac and Zoloft are different in their effectiveness to treat depression

What if? What if we have unequal sample sizes?

Unequal Sample Sizes In the previous example, n 1 = n 2 = 8, but What if n 1 ≠ n 2 ? In this case we make an adjustment to the calculation of the SEM But, since we calculate the pooled variance (a weighted mean), we’re OK

Just so we’re on the same page If n 1 is larger than n 2, then n will be larger than n This is larger than that

So… If n 1 is larger than n 2, then s 1 2 (n 1 - 1) will be weighted more than s 2 2 (n 2 - 1) This is weighted more than that

This makes sense If we make the homogeneity of variance assumption (the sampled populations have the same variance), then The best estimate of the population standard deviation will use information from both samples, But when we have more observations in one sample than the other, than we have more information from that sample than the other We should use that additional information, which is precisely what weighting accomplishes

Effect size estimates After conducting a t-test, you should report: –t –df –p But, it is becoming a standard practice to report effect size as well (Cohen’s d is a good measure)

Effect Size review Effect size – the strength of the relationship (between IV and DV) in the population, or, the degree of departure from the null hypothesis Important points: –rejecting the null hypothesis doesn’t imply a large effect, and –failing to reject the null does not mean a small effect

Example (from Rosenthal and Rosnow, 1991 – a great book on research methodology) Smith conducts an experiment with 40 learning disabled children –half undergo special training (“experimental group”) and – half receive no special training (“control group”) She reports that the experimental group improved more than the control group (p <.05)

But Jones is skeptical about Smith’s results and attempts to repeat (replicate) the experiment with 20 children, –half in the experimental and –half in the control group He reports a p >.10, and claims that Smith’s results are not-replicable

The Data Smith ’ s ResultsJones ’ Results t(38) = 1.85t(18) = 1.27 p <.05, Reject Hop >.10, Don’t Reject d =.15 power =.33power =.18

As you can see Even though Jones did not reject the null hypothesis, he had the same effect size as Smith Jones lacked power (but Smith had pretty low power as well)

Statistic = Effect Size X Size of Study

Statistic = effect size X size of study

And, if

What if one or more of the assumptions are violated? Gross, meaning large, violations may cause the real α to be different from the stated significance level Gross violations of the normality and H of V assumptions will cause these problems with a t-test

Alternative Test When gross violations of the assumptions of normality or variance with a 2-independent samples t-test becomes apparent, Use a Rank Sum T test

Rank Sum T test Rank all the scores (across both groups) Sum the ranks of each group (T = the sum of the ranks of group 1) Turns out that the T sampling distribution is approximately normal

Rank Sum T test

When to use Rank SumT Turns out, the t-test is fairly ROBUST to violations of HoV. But not large violations… What is a large violation of HoV? Recommendation: greater than 10x, use Rank Sum…