Statistics T-test Black: pp 362-364 and 402-416.

Slides:



Advertisements
Similar presentations
Statistics Review – Part II Topics: – Hypothesis Testing – Paired Tests – Tests of variability 1.
Advertisements

Comparing One Sample to its Population
PSY 307 – Statistics for the Behavioral Sciences
Tests of significance Confidence intervals are used when the goal of our analysis is to estimate an unknown parameter in the population. A second goal.
Statistics Are Fun! Analysis of Variance
Chapter 9 Hypothesis Testing.
1 (Student’s) T Distribution. 2 Z vs. T Many applications involve making conclusions about an unknown mean . Because a second unknown, , is present,
Hypothesis Testing: Two Sample Test for Means and Proportions
The t-test Inferences about Population Means when population SD is unknown.
AM Recitation 2/10/11.
Chapter 13 – 1 Chapter 12: Testing Hypotheses Overview Research and null hypotheses One and two-tailed tests Errors Testing the difference between two.
Week 9 Chapter 9 - Hypothesis Testing II: The Two-Sample Case.
Education 793 Class Notes T-tests 29 October 2003.
McGraw-Hill/Irwin Copyright © 2007 by The McGraw-Hill Companies, Inc. All rights reserved. Statistical Inferences Based on Two Samples Chapter 9.
T tests comparing two means t tests comparing two means.
Chapter 9 Hypothesis Testing and Estimation for Two Population Parameters.
Chapter 9 Hypothesis Testing II: two samples Test of significance for sample means (large samples) The difference between “statistical significance” and.
Chapter 9: Testing Hypotheses
1 Objective Compare of two matched-paired means using two samples from each population. Hypothesis Tests and Confidence Intervals of two dependent means.
Statistics for the Behavioral Sciences Second Edition Chapter 11: The Independent-Samples t Test iClicker Questions Copyright © 2012 by Worth Publishers.
Chapter 12 Tests of a Single Mean When σ is Unknown.
© Copyright McGraw-Hill 2000
1 Objective Compare of two population variances using two samples from each population. Hypothesis Tests and Confidence Intervals of two variances use.
Chapter 8 Parameter Estimates and Hypothesis Testing.
Ex St 801 Statistical Methods Inference about a Single Population Mean.
Chapter 9: Testing Hypotheses Overview Research and null hypotheses One and two-tailed tests Type I and II Errors Testing the difference between two means.
Inferences Concerning Variances
Sampling Distribution of Differences Between Means.
Monday, October 21 Hypothesis testing using the normal Z-distribution. Student’s t distribution. Confidence intervals.
Chapter 8 Single Sample Tests Part II: Introduction to Hypothesis Testing Renee R. Ha, Ph.D. James C. Ha, Ph.D Integrative Statistics for the Social &
CHAPTER 7: TESTING HYPOTHESES Leon-Guerrero and Frankfort-Nachmias, Essentials of Statistics for a Diverse Society.
Hypothesis Testing – Two Means(Small, Independent Samples)
HYPOTHESIS TESTING.
Section 6 Comparing Two Samples William Christensen, Ph.D.
Testing the Difference between Means, Variances, and Proportions
Testing the Difference between Means and Variances
Hypothesis Testing: One Sample Cases
Statistical Inference
Statistical Inference
Inference and Tests of Hypotheses
Math 4030 – 10b Inferences Concerning Variances: Hypothesis Testing
Estimation & Hypothesis Testing for Two Population Parameters
Math 4030 – 10a Tests for Population Mean(s)
Hypothesis testing using contrasts
Statistical Inference
Hypothesis testing March 20, 2000.
Hypothesis Testing and Confidence Intervals (Part 1): Using the Standard Normal Lecture 8 Justin Kern October 10 and 12, 2017.
Lecture 2 2-Sample Tests Goodness of Fit Tests for Independence
John Loucks St. Edward’s University . SLIDES . BY.
Two Sample Tests When do use independent
Hypothesis Testing: Two Sample Test for Means and Proportions
Hypothesis Testing: Hypotheses
INTRODUCTORY STATISTICS FOR CRIMINAL JUSTICE Test Review: Ch. 7-9
Chapter 11 Inferences About Population Variances
The t distribution and the independent sample t-test
Review: What influences confidence intervals?
Chapter 11: Inference About a Mean
Monday, October 19 Hypothesis testing using the normal Z-distribution.
Comparing Populations
Quantitative Methods in HPELS HPELS 6210
Essential Statistics Introduction to Inference
Hypothesis Testing.
What are their purposes? What kinds?
Intro to Confidence Intervals Introduction to Inference
Hypothesis Testing and Confidence Intervals
Statistical Inference for the Mean: t-test
Presentation transcript:

Statistics T-test Black: pp 362-364 and 402-416

Example problem Is it possible that this sample is coming from a high school educated population?

Example problem METHOD B – Confidence interval from sample to population Is it possible that this sample is coming from a high school educated population? It is 95% likely that this sample is coming from a population with a mean between 9.2 and 11.6. Therefore this sample is probably not coming from a high school educated population.

Problem METHOD B What kinds of populations can this sample be coming from? Compute the sample mean. Compute the estimated standard deviation for the population Compute the SEM Find out 95% confidence interval for the population that generated this sample Conclude if the given population could have generated THIS sample?

INTODUCING METHOD C– t-test method Is it likely that this sample came from a population with a mean ____ ? Is the mean of this sample close enough to the population mean of ____ ? Testing the Null Hypothesis: The sample mean and the given population mean are equal. Alternative hypothesis: The sample mean is different from the population mean The sample mean is greater than (smaller than) the population mean ANSWER: One sample t-test

More on the one sample t-test Population parameter: It is FIXED. It does not change from sample to sample. Sample statistic: It VARIES. It changes from sample to sample. Test statistic: It has a KNOWN distribution and its probability of occurrence can be looked up. Estimate of the Variability in Sample statistic: How much does sample statistic change from sample to sample.

METHOD C: t-test method Is it likely that this sample came from a population with a mean ____ ? Is the mean of this sample close enough to the population mean of ____ ? Write the null and alternative hypotheses Calculate the sample mean Calculate the difference of sample mean from population mean Take the absolute value of the difference Estimate of the POPULATION SD Calculate the SEM Divide the difference of means by SEMTHIS IS THE T-STATISTIC Look up the probability of the t statistic using the excel function: TDIST(t-value, df, 2)

Example problem – t-test method Is it possible that this sample is coming from a high school educated population? The probability that this sample is coming from a population with a mean of 12 is only 2.9%. Therefore it is HIGHLY UNLIKELY that this sample is coming from a high school educated population.

METHOD C– t-test method for testing SPECIFIC hypotheses “Is it likely that this sample came from a population with a mean ____ ?” OR “Is the mean of this sample close enough to the population mean of ____ ?” Null Hypothesis: The sample mean and the given population mean are equal. Alternative Hypothesis: Option 1: The sample mean is different from the population mean  TWO TAILED TEST Option 2: The sample mean is greater than the population mean  ONE TAILED TEST Option 3: The sample mean is smaller than the population mean  ONE TAILED TEST Apply one sample t-test 

Example Problem 1 vs. 2-tailed tests Depression scores from a sample of men who just lost their jobs. Is the depression level in this sample normal? (mean level of depression in a normal population is 10)? Null Hyp: The sample is coming from a population with mean depression = 10 Alt Hyp: The sample is coming from a population with mean depression GREATER THAN 10

Example problem – t-test method Is the depression level in this sample normal?

Practice for 1 and 2 tails Is the sample of children from Hakkari coming from a population with normal level of anxiety? Do American Indian children have a normal level of language development? Do the 2006 entries to KU have a similar average GPA as the 2005 entries? Do rural women have the same mean age at marriage as the urban women? Is the post-diet weight of women who went on Weight-Watchers diet similar to the group of women who did not diet?

Doing a t-test We must find the probability of this t value: Depends on sample size (df=n-1) Depends on the alternative hypothesis TDIST(t,df, tails)

Example Is it likely that this sample is coming from a population with normal reading ability (mean of 100)?

Example Is it likely that this sample is coming from a population with normal reading ability (mean of 100)? Alternatively: Is the mean of this sample close enough to 100?

Example Is it likely that this sample is coming from a population with normal reading ability (mean of 100)? Alternatively: Is the mean of this sample close enough to 100? Null Hyp: The mean of this sample is not different from 100. Alternative Hyp: The mean of this sample is lower than 100.

Example Calculate the mean: 70.6

Example Calculate the mean: 70.6 Calculate the SD: 10.2

Example Calculate the mean: 70.6 Calculate the SD: 10.2 Calculate the SEM: 2.05

Example Calculate the mean: 70.6 Calculate the SD: 10.2 Calculate the SEM: 2.05 Calculate the t-value: 14.4 Calculate the probability of that t: TDIST(14.4, 24,1) = 0.000000

Example Calculate the mean: 70.6 Calculate the SD: 10.2 Calculate the SEM: 2.05 Calculate the t-value: 14.4 Calculate the probability of that t: TDIST(14.4, 24,1) = 0.000000 CONCLUSION?

Example – t-test Mean weight of golden retriever dogs are 42kg. A sample of 20 golden retrievers were fed a new flavor of Purina dog food, and their average weight was found to be 46kg. The SD was estimated to be 5 kg. Can we say that the new Purina flavor is not healthy (i.e. they eat too much, thus they get too fat) for golden retrievers?

What if I have two samples? Children from divorced families have average aggression scores of 67. Children from intact families have average aggression scores of 52. DO CHILDREN FROM DIVORCED FAMILIES HAVE HIGHER AGGRESSION SCORES THAN CHILDREN IN INTACT FAMILIES?

PROBLEM TYPE 2: Comparing means Is the achievement of children from divorced families different from that of children from intact families? Do boys have a higher level of aggression than girls? Are working women happier than women who stay at home? Are students at Koc more satisfied than students at Sabanci? Do children who attended kindergarten perform better in first grade than children who did not attend kindergarten? Do religious women have less authority in the family than non-religious women? Do students from affluent families have less independent decision making skills than students from modest family backgrounds?

? ? TWO SAMPLES: Population Sample Mean=x Mean=x Population Sample Mean=y Sample Mean=y ?

TWO SAMPLES: Population Mean=z Sample Mean=x ? Sample Mean=y ?

TWO SAMPLE t-test Are the two means “close enough”? What is “close enough”? Are the two samples coming from the same population?

“Close enough” has to be related to the variability of each sample mean Difference of means variability of the difference in means

“Close enough” has to be related to the variability of each sample mean Difference of means variability of the difference in means Variance of the difference = variance of first mean + variance of second mean SQRT(Variance of the difference) = SQRT (variance of first mean + variance of second mean)

Application Mean of sample 1 = 150 Mean of sample 2 = 170 Est SD of pop 1 = 34 Est SD of pop 2 = 32 N of sample 1 = 50 N of sample 2 = 50 Are the sample means significantly different from each other? Step 1: Null and Alternative Hypotheses Null Hyp: The means of the two samples are equal. OR: The two samples are COMING FROM THE SAME POPULATION.

Application - continued Step 2: SEM of each mean Population 1: 34/sqrt(50) = 4.81 Population 2: 32/sqrt(50) = 4.53 Step 3: t-statistic Numerator = 170-150=20 Denominator = Sqrt(4.812 + 4.532) = sqrt(23.14 + 20.52) = 6.61 t-statistic = 20/6.61 = 3.03 df = 50+50-2 = 98 P-value = 0.003 Step 4: conclusion Conclusion: The two groups have significantly different means. OR The two groups are NOT coming from the same population.

Example Sample A: Intact families Sample B: Divorced families Mean 108 102 SD 16 13 N 112 84

QUIZ 3 on Wed Oct 24