| Sec.6.1 , all from Moore
.
p.302, 6.1 poll of women 6.2 95% confidence? 6.3 density of x-bar, and confidence intervals This problem combines the pictures 6.2 and 6.4 For part d, to draw the confidence interval: just choose any point on the horizontal axis of your graph to be x-bar. Measure off the distance m (half the width of the shaded interval) and extend a bar m wide to the left and the right of your point,below the curve. (Like fig. 6.4, the bars with arrows at the ends. The red dots show what the x-bar is for that confidence interval) Choose another point, and repeat.. If your first x-bar was in the shaded interval, pick your second outside the shaded interval, and vice versa. You should note that if x-bar is in the shaded interval, then the confidence interval bar covers mu (280) and if x-bar isn't, then the bar doesn't. -- - - - - - - - - - - - - - - - - - - - - - - - - - Using formula p. 306 for C.I.: 6.6 potassium again. 6.7 comparing CI's for different confidence levels. Also write down the m (margin of error) for each interval. 6.9 comparing CI's for different sample sizes. 6.5 IQ test scores Read pp. 312-13 before doing this one. - - - - - - - - - - - - - - - - - - - - - - - - - - - Postpone:Cautions; general review and extension p. 314 6.14 internet, response rate p.317, 6.19 newts p. 318, 6.22 men/women CI's p.316, 6.18 consumers/pharmacies |
Read, to discuss
- - - - - - - -
- - - - - - 6.13, 6.15 (cautions)
|
Optional |
Question from last time: Do people really use sampling
to decide if they should accept an order of something?
Activstats p. 10-1, activity 2
Sample Size and Population Size (Activstats 10-1 activity
4, or Activstats, 16-1, green star)
S.D of Xbar is population sigma
divided
by square root of sample size.
"It may seem that the sample size alone should not determine
the variability of statistics computed on random samples, but rather that
the fraction of the population that has been sampled should
somehow be important..... Consider a common form of sampling; tasting
soup from a pot. If the soup is well-stirred, we are not surprised
that a spoonful will tell us how the entire pot tastes. Nor would
anyone argue that to test a small pot of soup requires a spoonful, but
to test a large vat of soup requires eating a bowlful. Now consider
tasting the soup only by dipping a toothpick into it. This small
sample may be too little to provide an accurate impression of how the soup
tastes; we need at least a spoonful or so.
In the same way, the size of the sample is what determines how precisely
statistics from that sample describe the population. But the sample
size has the same effect regardless of the population size." [As long
as the population is at least about 10 times the size of the sample.]
Fuzzy Central Limit Theorem, again:
Data whose variation is due to many small independent random
influences will have an approximately normal distribution.
"Quincunx "probability board
* * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Chapter 6, Introduction to Inference Go
to Day 28
Will continue with Assumptions p. 312-13, sample
size Friday
| Sievers home | Math151-Fall02/Day-29.htm | 2pm | 11/6/02 |