Biostatistics Case Studies 2016 Youngju Pak, PhD. Biostatistician Session 1 Understanding hypothesis testing, P values, and sample size.

Slides:



Advertisements
Similar presentations
Statistics Hypothesis Testing.
Advertisements

Our goal is to assess the evidence provided by the data in favor of some claim about the population. Section 6.2Tests of Significance.
Chapter 9 Hypothesis Testing Understandable Statistics Ninth Edition
Hypothesis Testing An introduction. Big picture Use a random sample to learn something about a larger population.
Chapter 12 Tests of Hypotheses Means 12.1 Tests of Hypotheses 12.2 Significance of Tests 12.3 Tests concerning Means 12.4 Tests concerning Means(unknown.
Inference Sampling distributions Hypothesis testing.
Last Time (Sampling &) Estimation Confidence Intervals Started Hypothesis Testing.
Our goal is to assess the evidence provided by the data in favor of some claim about the population. Section 6.2Tests of Significance.
INFERENCE: SIGNIFICANCE TESTS ABOUT HYPOTHESES Chapter 9.
Testing Hypotheses About Proportions Chapter 20. Hypotheses Hypotheses are working models that we adopt temporarily. Our starting hypothesis is called.
Probability & Statistical Inference Lecture 6
EPIDEMIOLOGY AND BIOSTATISTICS DEPT Esimating Population Value with Hypothesis Testing.
Fundamentals of Hypothesis Testing. Identify the Population Assume the population mean TV sets is 3. (Null Hypothesis) REJECT Compute the Sample Mean.
Introduction to Hypothesis Testing CJ 526 Statistical Analysis in Criminal Justice.
Introduction to Hypothesis Testing CJ 526 Statistical Analysis in Criminal Justice.
BCOR 1020 Business Statistics Lecture 20 – April 3, 2008.
BCOR 1020 Business Statistics
Inference about Population Parameters: Hypothesis Testing
Thomas Songer, PhD with acknowledgment to several slides provided by M Rahbar and Moataza Mahmoud Abdel Wahab Introduction to Research Methods In the Internet.
AM Recitation 2/10/11.
Overview Definition Hypothesis
1 © Lecture note 3 Hypothesis Testing MAKE HYPOTHESIS ©
Hypothesis Testing.
Introduction to Biostatistics and Bioinformatics
Testing Hypotheses About Proportions
Copyright © 2010, 2007, 2004 Pearson Education, Inc. Chapter 20 Testing Hypotheses About Proportions.
Testing Hypotheses Tuesday, October 28. Objectives: Understand the logic of hypothesis testing and following related concepts Sidedness of a test (left-,
Lesson 11 - R Review of Testing a Claim. Objectives Explain the logic of significance testing. List and explain the differences between a null hypothesis.
Biostatistics Case Studies 2015 Youngju Pak, PhD. Biostatistician Session 2: Sample Size & Power for Inequality and Equivalence Studies.
Statistical Inference Decision Making (Hypothesis Testing) Decision Making (Hypothesis Testing) A formal method for decision making in the presence of.
© 2003 Prentice-Hall, Inc.Chap 7-1 Business Statistics: A First Course (3 rd Edition) Chapter 7 Fundamentals of Hypothesis Testing: One-Sample Tests.
Introduction To Biological Research. Step-by-step analysis of biological data The statistical analysis of a biological experiment may be broken down into.
Chapter 8 Introduction to Hypothesis Testing
A Broad Overview of Key Statistical Concepts. An Overview of Our Review Populations and samples Parameters and statistics Confidence intervals Hypothesis.
Biostatistics Case Studies 2015 Youngju Pak, PhD. Biostatistician Session 1: Sample Size & Power for Inequality and Equivalence Studies.
Chapter 20 Testing hypotheses about proportions
Testing of Hypothesis Fundamentals of Hypothesis.
Lecture 16 Dustin Lueker.  Charlie claims that the average commute of his coworkers is 15 miles. Stu believes it is greater than that so he decides to.
No criminal on the run The concept of test of significance FETP India.
Copyright © 2008 Pearson Education, Inc. Publishing as Pearson Addison-Wesley Chapter 20 Testing Hypotheses About Proportions.
Lecture 16 Section 8.1 Objectives: Testing Statistical Hypotheses − Stating hypotheses statements − Type I and II errors − Conducting a hypothesis test.
McGraw-Hill/Irwin Copyright © 2007 by The McGraw-Hill Companies, Inc. All rights reserved. Chapter 8 Hypothesis Testing.
Economics 173 Business Statistics Lecture 4 Fall, 2001 Professor J. Petry
Chapter 20 Testing Hypothesis about proportions
Lecture 18 Dustin Lueker.  A way of statistically testing a hypothesis by comparing the data to values predicted by the hypothesis ◦ Data that fall far.
Statistical Inference An introduction. Big picture Use a random sample to learn something about a larger population.
Lecture 17 Dustin Lueker.  A way of statistically testing a hypothesis by comparing the data to values predicted by the hypothesis ◦ Data that fall far.
3-1 MGMG 522 : Session #3 Hypothesis Testing (Ch. 5)
Fall 2002Biostat Statistical Inference - Confidence Intervals General (1 -  ) Confidence Intervals: a random interval that will include a fixed.
Copyright © 2006 Pearson Education, Inc. Publishing as Pearson Addison-Wesley Slide
Hypothesis Testing. “Not Guilty” In criminal proceedings in U.S. courts the defendant is presumed innocent until proven guilty and the prosecutor must.
A review of key statistical concepts. An overview of the review Populations and parameters Samples and statistics Confidence intervals Hypothesis testing.
© 2004 Prentice-Hall, Inc.Chap 9-1 Basic Business Statistics (9 th Edition) Chapter 9 Fundamentals of Hypothesis Testing: One-Sample Tests.
STA Lecture 221 !! DRAFT !! STA 291 Lecture 22 Chapter 11 Testing Hypothesis – Concepts of Hypothesis Testing.
What is a Hypothesis? A hypothesis is a claim (assumption) about the population parameter Examples of parameters are population mean or proportion The.
Course Overview Collecting Data Exploring Data Probability Intro. Inference Comparing Variables Relationships between Variables Means/Variances Proportions.
Copyright © 2007 Pearson Education, Inc. Publishing as Pearson Addison-Wesley Slide
Chapter 12 Tests of Hypotheses Means 12.1 Tests of Hypotheses 12.2 Significance of Tests 12.3 Tests concerning Means 12.4 Tests concerning Means(unknown.
6.2 Large Sample Significance Tests for a Mean “The reason students have trouble understanding hypothesis testing may be that they are trying to think.”
Today: Hypothesis testing p-value Example: Paul the Octopus In 2008, Paul the Octopus predicted 8 World Cup games, and predicted them all correctly Is.
Slide 20-1 Copyright © 2004 Pearson Education, Inc.
Hypothesis Tests Hypothesis Tests Large Sample 1- Proportion z-test.
Statistical Significance or Hypothesis Testing. Significance testing Learning objectives of this lecture are to Understand Hypothesis: definition & types.
Statistics 20 Testing Hypothesis and Proportions.
Lecture Nine - Twelve Tests of Significance.
Testing Hypotheses About Proportions
Georgi Iskrov, MBA, MPH, PhD Department of Social Medicine
Chapter 9 Hypothesis Testing.
Testing Hypotheses About Proportions
STA 291 Spring 2008 Lecture 17 Dustin Lueker.
Presentation transcript:

Biostatistics Case Studies 2016 Youngju Pak, PhD. Biostatistician Session 1 Understanding hypothesis testing, P values, and sample size determination. 1

Overview of biostatistical supports Biostatistics consulting services available to LABioMed investigators: Assistance with study design and protocol development Developing Data Analysis Plans Power and sample size calculation Creating randomization schedules Guidance in data analysis and interpretation of results Advice on statistical methods and use of statistical software Discussion with journal club presenters on statistical aspects of the article 2

Announcements All lecture materials will be uploaded in the following website research.labiomed.org/Biostat  statistics Education  Courses  Biostatistics Case studies: Spring 2016 Try to read posted articles before the class you can and pay more attention to statistical components when you read them Send me an so I can communicate with you if necessary. 3

Five stages when carrying out a hypothesis test 1.Define the null (H 0 ) and alternative(Ha) hypothesis under the study. 2.Collect relevant data from a sample of individuals. 3.Calculate the value of the test statistics specific to the null hypothesis 4.Compute the P-value by compare the value of the test statistics to values from a known probability distribution 5.Interpret the P-value and results 4

A criminal prosecution in U.S. justice system 1.Define the null (H 0 ) and alternative(Ha) hypothesis under the study : a primary suspect is arrested and assumed to be “Not Guilty” (H 0 ) until proven, H a to be “Guilty” 2.Collect relevant data from a sample of individuals: works from a prosecutor and a lawyer to find the evidence to prove “Guilty (H a ) ” & evidence against “Guilty (H a ) ” 3.Calculate the value of the test statistics specific to the null hypothesis : a prosecutor aggregate all possible evidences/witness statements to make “Not Guilty (H 0 ) ” to be rejected BEYOUND a reasonable doubt by jury 5

A criminal prosecution in U.S. justice system 4.Compute the P-value by comparing the value of the test statistics to values from a known probability distribution: a jury decide how rare all evidences presented by a prosecutor if a defendant is “Not Guilty”. Is it a beyond reasonable doubt? 5.Interpret the P-value and results : How RARE what I see from all prosecutor’s evidences if a defendant is “Not Guilty”? 6

How to interpret P Value, in general ? A P Value is predicted probability on the assumption that H 0 is true A P Value measure the degree of “RARENESS” of what your data show if H0 is true. A P Value is NOT a probability of the alternative being correct. A P Value should be used as an evidence to DISPROVE H 0, not to prove the Ha. ( Not innocent enough ! Thus we are favor toward the defendant to be GUILTY, but we DO NOT prove the defendant to be GUILTY). 7

Justice system-Trial/Hypothesis test Two sides of the coin 8 Defendant Not guilty (H0) Defendant Guilty (Ha) Reject “Not guilty(H 0 )” beyond reasonable doubt Type I error (α)Correct decision Fail to Reject H 0 Correct decisionType II error (β)  Statistical Power = Prob.(Reject H0 when Ha is true) = 1-β Different factors play the role in sample size calculation depending on a statistical test to test a primary hypothesis. But common parameters to determine the sample size are statistical power, type I error rate, and the effect size ( how much mean difference between two groups relative to the standard deviation) for a two sample t-test.

Hypothesis test to test Inequality Two or more treatments are assumed equal (H 0 )and the study is designed to find overwhelming evidence of a difference (Superiority and/or Inferiority). Most common comparative study type. It is rare to assess only one of superiority or inferiority (“one-sided” statistical tests), unless there is biological impossibility of one of them. Hypotheses: H a : | mean(treatment ) - mean (control ) | ≠ 0 H 0 : | mean(treatment ) - mean (control ) | = 0 9

Insignificnat p-values for Inequality tests Insignificant p-values (> 0.05) usually mean that you don’t find a statistically sufficient evidence to support Ha and this doesn’t necessary mean H 0 is true. H 0 might or might not be true => Your study is still “INCONCLUSIVE”. Insignificant p-values do NOT prove your null ! 10

Equivalence Study: Two treatments are assumed to differ (H 0 ) and the study is designed to find overwhelming evidence that they are equal. Usually, the quantity of interest is a measure of biological activity or potency (the amount of drug required to produce an effect) and “treatments” are drugs or lots or batches of drugs. AKA, bioequivalence. Sometimes used to compare clinical outcomes for two active treatments if neither treatment can be considered standard or accepted. This usually requires LARGE numbers of subjects. 11

Hypotheses for equivalence tests H a : mean (trt 1) – mean (trt 2) = 0 H 0 : mean(trt 1) - mean (trt 2 ) ≠ 0 With a finite sample size, it is very hard to find two group means are exactly the same. So we put a tolerability level for the equivalence, AKA, the equivalence margin, usually denoted as Δ Practical hypotheses would be H a : Δ 1 < mean(trt 1) – mean (trt2) < Δ 2 H 0 : mean(trt 1) – mean (trt2) ≤ Δ 1 or mean(trt 1) – mean (trt2) ≥ Δ 2 Non-inferiority 12

Today, we are going to learn how to determine sample size for Inequality tests using software using two published studies. 13

Study #1 14

How was 498 determined? Back to: 15

From earlier design paper (Russell 2007): Δ = 0.85(0.05) mm = mm 16

Need to Increase N for Power Need to increase N to: 2SD 2 Δ 2 ( ) 2 Power is the probability that p<0.05 if Δ is the real effect, incorporating the possibility that the Δ in our sample could be smaller. 2SD 2 Δ 2 (1.96) 2 N =for 50% power. for 80% power.N = 2SD 2 Δ 2 ( ) 2 for 90% power. from Normal Tables 17

Info Needed for Study Size: Comparing Means 1.Effect 2.Subject variability 3.Type I error (1.96 for α=0.05; 2.58 for α=0.01) 4.Power (0.842 for 80% power; for 95% power) ( ) 2 2SD 2 Δ 2 N = Same four quantities, but different formula, if comparing %s, hazard ratios, odds ratios, etc. Δ/SD = Effect size 18

Comparing two independent means using G*Power (Free software for power calculations) 19

Comparing two independent means using G*Power (Free software for power calculations) 20

Comparing two independent means using G*Power (Free software for power calculations) 21

SD Estimate Could be Wrong Should examine SD as study progresses. May need to increase N if SD was underestimated. 22

Study #2 23

24

Sample size justification 25

Comparing two independent proportions using G*Power

Comparing two independent proportions using G*Power

Comparing two independent proportions using G*Power

A statistical power primarily depends on what statistical test to be used. The choice of statistical tests depends the data type of two variables (dependent v.s independent variables). Dependent variables are outcomes of interest while independent variables are the hypothesized predictors of outcomes. Independent variables are also called explanatory variables 29

Variable CategoricalNumerical Ordinal Categories are mutually exclusive and ordered Examples: Disease stage, Education level, 5 point likert scale Counts Integer values Examples: Days sick per year, Number of pregnancies, Number of hospital visits Measured (continuous) Takes any value in a range of values Examples: weight in kg, height in feet, age (in years) QualitativeQuantitative Nominal Categories are mutually exclusive and unordered Examples: Gender, Blood group, Eye colour, Marital status Types of Data 30

31 Choosing a statistical test ► DV: Dependent variable, IV: Independent variable, where IV affects DV. For example, treatment is IV and clinical outcome is DV when treatments affect clinical outcomes.

A statistically significant result --- is not necessarily an important or even interesting result may not be scientifically interesting or clinically significant. With large sample sizes, very small differences may turn out to be statistically significant. In such a case, practical implications of any findings must be judged on other than statistical grounds. Statistical significance does not imply practical significance 32

Assumptions Random samples from the population –Beware of convenience samples Population is Gaussian (Normal distribution) if sample size is “small” (n<30) Independent observations –Beware of double counting or repeated measures 33

Other Sample Size Software 34

Free Sample Size Software 35

Study Size Software in GCRC Lab ncss.com ~$500 36

nQuery - Used by Most Drug Companies 37