Inference about the mean of a population of measurements (  ) is based on the standardized value of the sample mean (Xbar). The standardization involves.

Slides:



Advertisements
Similar presentations
BPS - 5th Ed. Chapter 181 Two-Sample Problems. BPS - 5th Ed. Chapter 182 Two-Sample Problems u The goal of inference is to compare the responses to two.
Advertisements

Estimating a Population Mean
Warm up The mean salt content of a certain type of potato chips is supposed to be 2.0mg. The salt content of these chips varies normally with standard.
AP STATISTICS LESSON 11 – 1 (DAY 3) Matched Pairs t Procedures.
Conditions with σ Unknown Note: the same as what we saw before.
AP Statistics Section 10.2 A CI for Population Mean When is Unknown.
BPS - 5th Ed. Chapter 171 Inference about a Population Mean.
Chapter 11: Inference for Distributions
+ DO NOW What conditions do you need to check before constructing a confidence interval for the population proportion? (hint: there are three)
Estimating a Population Mean
Chapter 8: Estimating with Confidence
+ The Practice of Statistics, 4 th edition – For AP* STARNES, YATES, MOORE Chapter 8: Estimating with Confidence Section 8.3 Estimating a Population Mean.
Ch 11 – Inference for Distributions YMS Inference for the Mean of a Population.
10.2 Estimating a Population Mean when σ is Unknown When we don’t know σ, we substitute the standard error of for its standard deviation. This is the standard.
The Practice of Statistics Third Edition Chapter 11: Inference for Distributions Copyright © 2008 by W. H. Freeman & Company Daniel S. Yates.
Chapter 11 Inference for Distributions AP Statistics 11.1 – Inference for the Mean of a Population.
BPS - 5th Ed. Chapter 171 Inference about a Population Mean.
Section 8.3 Estimating a Population Mean. Section 8.3 Estimating a Population Mean After this section, you should be able to… CONSTRUCT and INTERPRET.
CHAPTER 18: Inference about a Population Mean
When σ is Unknown The One – Sample Interval For a Population Mean Target Goal: I can construct and interpret a CI for a population mean when σ is unknown.
1 Happiness comes not from material wealth but less desire.
For 95 out of 100 (large) samples, the interval will contain the true population mean. But we don’t know  ?!
AP Exam Prep: Essential Notes. Chapter 11: Inference for Distributions 11.1Inference for Means of a Population 11.2Comparing Two Means.
CHAPTER 11 DAY 1. Assumptions for Inference About a Mean  Our data are a simple random sample (SRS) of size n from the population.  Observations from.
+ The Practice of Statistics, 4 th edition – For AP* STARNES, YATES, MOORE Chapter 8: Estimating with Confidence Section 8.3 Estimating a Population Mean.
BPS - 3rd Ed. Chapter 161 Inference about a Population Mean.
Essential Statistics Chapter 161 Inference about a Population Mean.
Week111 The t distribution Suppose that a SRS of size n is drawn from a N(μ, σ) population. Then the one sample t statistic has a t distribution with n.
Lesson 9 - R Chapter 9 Review.
+ The Practice of Statistics, 4 th edition – For AP* STARNES, YATES, MOORE Unit 5: Estimating with Confidence Section 11.1 Estimating a Population Mean.
Lecture PowerPoint Slides Basic Practice of Statistics 7 th Edition.
+ The Practice of Statistics, 4 th edition – For AP* STARNES, YATES, MOORE Chapter 8: Estimating with Confidence Section 8.3 Estimating a Population Mean.
Essential Statistics Chapter 171 Two-Sample Problems.
2003 Free Response Question #2 A simple random sample of adults living in a suburb of a large city was selected. The age and annual income of each adult.
+ Unit 5: Estimating with Confidence Section 8.3 Estimating a Population Mean.
+ Z-Interval for µ So, the formula for a Confidence Interval for a population mean is To be honest, σ is never known. So, this formula isn’t used very.
Chapter 7 Inference Concerning Populations (Numeric Responses)
Inference about the mean of a population of measurements (  ) is based on the standardized value of the sample mean (Xbar). The standardization involves.
+ Chapter 8 Estimating with Confidence 8.1Confidence Intervals: The Basics 8.2Estimating a Population Proportion 8.3Estimating a Population Mean.
In your groups, go through all four steps of a confidence interval: The “Country Taste” bread making company wants to estimate the actual weight of their.
AP Statistics Chapter 11 Section 1. TestConfidence Interval FormulasAssumptions 1-sample z-test mean SRS Normal pop. Or large n (n>40) Know 1-sample t-test.
16/23/2016Inference about µ1 Chapter 17 Inference about a Population Mean.
10.2 ESTIMATING A POPULATION MEAN. QUESTION: How do we construct a confidence interval for an unknown population mean when we don’t know the population.
CHAPTER 8 Estimating with Confidence
Chapter 8: Estimating with Confidence
Chapter 8: Estimating with Confidence
Inference for Distributions
Chapter 8: Estimating with Confidence
WARM - UP 1. How do you interpret a Confidence Interval?
Thus we have (Xbar - m )/(s/sqrt(n)) which has a Z distribution if:
Warmup To check the accuracy of a scale, a weight is weighed repeatedly. The scale readings are normally distributed with a standard deviation of
Chapter 8: Estimating with Confidence
Inference for Distributions
Estimating with Confidence
Chapter 8: Estimating with Confidence
Basic Practice of Statistics - 3rd Edition Two-Sample Problems
Essential Statistics Two-Sample Problems - Two-sample t procedures -
Chapter 8: Estimating with Confidence
Chapter 8: Estimating with Confidence
Chapter 8: Estimating with Confidence
Basic Practice of Statistics - 3rd Edition
Chapter 8: Estimating with Confidence
Chapter 8: Estimating with Confidence
2/5/ Estimating a Population Mean.
Chapter 8: Estimating with Confidence
Chapter 8: Estimating with Confidence
Chapter 8: Estimating with Confidence
Essential Statistics Inference about a Population Mean
Inference for Distributions
Presentation transcript:

Inference about the mean of a population of measurements (  ) is based on the standardized value of the sample mean (Xbar). The standardization involves subtracting the mean of Xbar and dividing by the standard deviation of Xbar – recall that –Mean of Xbar is  ; and –Standard deviation of Xbar is  /sqrt(n) Thus we have (Xbar -  )/(  /sqrt(n)) which has a Z distribution if: –Population is normal and  is known ; or if –n is large so CLT takes over…

But what if  is unknown?? Then this standardized Xbar doesn’t have a Z distribution anymore, but a so-called t-distribution with n-1 degrees of freedom… Since  is unknown, the standard deviation of Xbar,  /sqrt(n), is unknown. We estimate it by the so-called standard error of Xbar, s/sqrt(n), where s=the sample standard deviation. There is a t-distribution for every value of the sample size; we’ll use t(k) to stand for the particular t-distribution with k degrees of freedom. There are some properties of these t- distributions that we should note…

Every t-distribution looks like a N(0,1) distribution; i.e., it is centered and symmetric around 0 and has the same characteristic “bell” shape… however, the standard deviation of t(k) {sqrt(k/(k-2))} is greater than 1, the s.d. of Z so the t-distribution density curve is more spread out than Z. Probabilities involving r.v.s that have the t(k) distributions are given by areas under the t(k) density curve … Table D in the back of our book gives us the probabilities we need…

The good news is that everything we’ve already learned about constructing confidence intervals and testing hypotheses about  carries through under the assumption of unknown  … So e.g., a 95% confidence interval for  based on a SRS from a population with unknown  is Xbar +/- t * (s.e.(Xbar)) Recall that s.e.(Xbar) = s/sqrt(n). Here t * is the appropriate tabulated value from Table D so that the area between –t * and +t * is.95 As we did before, if we change the level of confidence then the value of t * must change appropriately…

Similarly, we may test hypotheses using this t- distributed standardized Xbar… e.g., to test the H 0 :    against H a :    we use (Xbar -    /(s/sqrt(n)) which has a t- distribution with n-1 df, assuming the null hypothesis is true. See page 422 (7.1, 3/7) for a complete summary of hypothesis testing in the case of “the one-sample t-test” … HW: Read section 7.1 thru p. 433; go over all the examples carefully and answer the HW questions following them: # Work on the following problems (p.441 ff) (use JMP as needed): # , 7.25, 7.32, , 7.41.

To summarize the analysis: – when the data comes in matched pairs, the analysis is performed on the differences between the paired measurements – then use the t-statistic with n-1 d.f. (n = # of pairs) to construct confidence intervals and test hypotheses on the true mean difference. Is there a difference in aggressive behavior of patients on "moon days" compared with "non-moon days"?

In a matched pairs design, subjects are matched in pairs and the outcomes are compared within each matched pair. A coin toss could determine which of the two subjects gets the treatment and which gets the control… One special kind of matched pairs design is when a subject acts as his/her own control, as in a before/after study… See example 7.7 on page 428ff (7.1, 4/7). Note that the paired observations (# of agressive behaviors) are subtracted and the difference in scores becomes the single number analyzed with a one-sample t-statistic with n-1 df, where n=the number of pairs… see the top of page 431 and the next page for a summary of the process. HW Read through p.433. Go over Example 7.7 then do #7.32, 7.35, 7.41.

Read the section on Robustness of the t procedures (starting p.432 (7.1, 5/7))… note the definition of the statistical term robust – essentially, a statistic is robust if it is insensitive to violations of the assumptions made when the statistic is used. For example, the t-statistic requires normality of the population… how sensitive is the t-statistic to violations of normality?? Look at the practical guidelines for inference on a single mean at bottom of p.432… –If the sample size is < 15, use the t procedures if the data are close to normal. –If the sample size is >= 15 then unless there is strong non-normality or outliers, t procedures are OK –If the sample size is large (say n >= 40) then even if the distribution is skewed, t procedures are OK