unit 3: the normal approximation for data
TRANSCRIPT
![Page 1: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/1.jpg)
Unit 3: The Normal Approximation for Data
Marius Ionescu
09/13/2011
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 2: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/2.jpg)
Chapter 5: The Normal Distribution
Fact
The normal distribution is a very common approximate distribution
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 3: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/3.jpg)
Properties of the Normal Distribution
Fact
1 It has a single bump
2 It is symmetric about the average
3 Its shape depends only on average and SD
4 68% of the area lies within 1 SD of the average
5 95% lies within 2 SD
6 The height is given by
1√2πSD
e−1
2( x−Avg
SD )2
.
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 4: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/4.jpg)
Properties of the Normal Distribution
Fact
1 It has a single bump
2 It is symmetric about the average
3 Its shape depends only on average and SD
4 68% of the area lies within 1 SD of the average
5 95% lies within 2 SD
6 The height is given by
1√2πSD
e−1
2( x−Avg
SD )2
.
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 5: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/5.jpg)
Properties of the Normal Distribution
Fact
1 It has a single bump
2 It is symmetric about the average
3 Its shape depends only on average and SD
4 68% of the area lies within 1 SD of the average
5 95% lies within 2 SD
6 The height is given by
1√2πSD
e−1
2( x−Avg
SD )2
.
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 6: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/6.jpg)
Properties of the Normal Distribution
Fact
1 It has a single bump
2 It is symmetric about the average
3 Its shape depends only on average and SD
4 68% of the area lies within 1 SD of the average
5 95% lies within 2 SD
6 The height is given by
1√2πSD
e−1
2( x−Avg
SD )2
.
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 7: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/7.jpg)
Properties of the Normal Distribution
Fact
1 It has a single bump
2 It is symmetric about the average
3 Its shape depends only on average and SD
4 68% of the area lies within 1 SD of the average
5 95% lies within 2 SD
6 The height is given by
1√2πSD
e−1
2( x−Avg
SD )2
.
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 8: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/8.jpg)
Properties of the Normal Distribution
Fact
1 It has a single bump
2 It is symmetric about the average
3 Its shape depends only on average and SD
4 68% of the area lies within 1 SD of the average
5 95% lies within 2 SD
6 The height is given by
1√2πSD
e−1
2( x−Avg
SD )2
.
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 9: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/9.jpg)
Properties of the Normal Distribution
Fact
1 It has a single bump
2 It is symmetric about the average
3 Its shape depends only on average and SD
4 68% of the area lies within 1 SD of the average
5 95% lies within 2 SD
6 The height is given by
1√2πSD
e−1
2( x−Avg
SD )2
.
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 10: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/10.jpg)
Standard Units
De�nition
If X is approximated by a normal distribution with average Avg andSD, the the standard units are obtain by setting average to zeroand scaling SD to one:
Z =X − Avg
SD.
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 11: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/11.jpg)
Standard Units: Example
Example
Suppose that the average of one midterm is 65 and SD is 10. Howmany SDs away from the average is your score if it is
68
78
50
60
What was your score if you were 1.75 SD above the average?
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 12: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/12.jpg)
Standard Units: Example
Example
Suppose that the average of one midterm is 65 and SD is 10. Howmany SDs away from the average is your score if it is
68
78
50
60
What was your score if you were 1.75 SD above the average?
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 13: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/13.jpg)
Standard Units: Example
Example
Suppose that the average of one midterm is 65 and SD is 10. Howmany SDs away from the average is your score if it is
68
78
50
60
What was your score if you were 1.75 SD above the average?
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 14: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/14.jpg)
Standard Units: Example
Example
Suppose that the average of one midterm is 65 and SD is 10. Howmany SDs away from the average is your score if it is
68
78
50
60
What was your score if you were 1.75 SD above the average?
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 15: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/15.jpg)
Standard Units: Example
Example
Suppose that the average of one midterm is 65 and SD is 10. Howmany SDs away from the average is your score if it is
68
78
50
60
What was your score if you were 1.75 SD above the average?
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 16: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/16.jpg)
Standard Units: Example
Example
Suppose that the average of one midterm is 65 and SD is 10. Howmany SDs away from the average is your score if it is
68
78
50
60
What was your score if you were 1.75 SD above the average?
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 17: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/17.jpg)
Standard Units: Example
Example
Suppose that the average of one midterm is 65 and SD is 10. Howmany SDs away from the average is your score if it is
68
78
50
60
What was your score if you were 1.75 SD above the average?
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 18: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/18.jpg)
Finding areas under the curve
Fact
We use a table to �nd the area under the normal curve: A(z)=Areabetween −z and +z
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 19: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/19.jpg)
Examples
Example
A(1) = 68.27%
A(2) =
95.45%
Example
Find the area:
between −z and 0
between −z and w
below 0
above w
below −w
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 20: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/20.jpg)
Examples
Example
A(1) = 68.27%
A(2) =
95.45%
Example
Find the area:
between −z and 0
between −z and w
below 0
above w
below −w
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 21: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/21.jpg)
Examples
Example
A(1) = 68.27%
A(2) =
95.45%
Example
Find the area:
between −z and 0
between −z and w
below 0
above w
below −w
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 22: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/22.jpg)
Examples
Example
A(1) = 68.27%
A(2) =
95.45%
Example
Find the area:
between −z and 0
between −z and w
below 0
above w
below −w
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 23: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/23.jpg)
Examples
Example
A(1) = 68.27%
A(2) =95.45%
Example
Find the area:
between −z and 0
between −z and w
below 0
above w
below −w
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 24: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/24.jpg)
Examples
Example
A(1) = 68.27%
A(2) =95.45%
Example
Find the area:
between −z and 0
between −z and w
below 0
above w
below −w
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 25: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/25.jpg)
Examples
Example
A(1) = 68.27%
A(2) =95.45%
Example
Find the area:
between −z and 0
between −z and w
below 0
above w
below −w
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 26: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/26.jpg)
Examples
Example
A(1) = 68.27%
A(2) =95.45%
Example
Find the area:
between −z and 0
between −z and w
below 0
above w
below −w
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 27: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/27.jpg)
Examples
Example
A(1) = 68.27%
A(2) =95.45%
Example
Find the area:
between −z and 0
between −z and w
below 0
above w
below −w
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 28: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/28.jpg)
Examples
Example
A(1) = 68.27%
A(2) =95.45%
Example
Find the area:
between −z and 0
between −z and w
below 0
above w
below −w
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 29: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/29.jpg)
Examples
Example
A(1) = 68.27%
A(2) =95.45%
Example
Find the area:
between −z and 0
between −z and w
below 0
above w
below −w
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 30: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/30.jpg)
Example
Example
Suppose that the Math SAT has a normal distribution with anaverage score of 500 and SD=100.
What percent of scores are above 650?
What percent of scores are between 600 and 650 or between350 and 400?
What percent of scores are between 600 and 650?
What percentile would 650 be?
What percentile would 400 be?
If you score in the 70th%ile, what is your score?
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 31: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/31.jpg)
Example
Example
Suppose that the Math SAT has a normal distribution with anaverage score of 500 and SD=100.
What percent of scores are above 650?
What percent of scores are between 600 and 650 or between350 and 400?
What percent of scores are between 600 and 650?
What percentile would 650 be?
What percentile would 400 be?
If you score in the 70th%ile, what is your score?
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 32: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/32.jpg)
Example
Example
Suppose that the Math SAT has a normal distribution with anaverage score of 500 and SD=100.
What percent of scores are above 650?
What percent of scores are between 600 and 650 or between350 and 400?
What percent of scores are between 600 and 650?
What percentile would 650 be?
What percentile would 400 be?
If you score in the 70th%ile, what is your score?
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 33: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/33.jpg)
Example
Example
Suppose that the Math SAT has a normal distribution with anaverage score of 500 and SD=100.
What percent of scores are above 650?
What percent of scores are between 600 and 650 or between350 and 400?
What percent of scores are between 600 and 650?
What percentile would 650 be?
What percentile would 400 be?
If you score in the 70th%ile, what is your score?
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 34: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/34.jpg)
Example
Example
Suppose that the Math SAT has a normal distribution with anaverage score of 500 and SD=100.
What percent of scores are above 650?
What percent of scores are between 600 and 650 or between350 and 400?
What percent of scores are between 600 and 650?
What percentile would 650 be?
What percentile would 400 be?
If you score in the 70th%ile, what is your score?
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 35: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/35.jpg)
Example
Example
Suppose that the Math SAT has a normal distribution with anaverage score of 500 and SD=100.
What percent of scores are above 650?
What percent of scores are between 600 and 650 or between350 and 400?
What percent of scores are between 600 and 650?
What percentile would 650 be?
What percentile would 400 be?
If you score in the 70th%ile, what is your score?
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 36: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/36.jpg)
Example
Example
Suppose that the Math SAT has a normal distribution with anaverage score of 500 and SD=100.
What percent of scores are above 650?
What percent of scores are between 600 and 650 or between350 and 400?
What percent of scores are between 600 and 650?
What percentile would 650 be?
What percentile would 400 be?
If you score in the 70th%ile, what is your score?
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 37: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/37.jpg)
Example
Example
For a certain group of people the 25th%ile is 62.2 inches, the75th%ile is 65.8 inches. The histogram follows a normaldistribution. Find the 90th%ile.
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 38: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/38.jpg)
Measurement Error
Fact
Any measurement has error associated with it.
Measurement=Exact Value + Variation
Error can be of two types:
1 Bias
2 Chance Error
Measurement = Exact Value + Bias + Chance Error
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 39: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/39.jpg)
Measurement Error
Fact
Any measurement has error associated with it.
Measurement=Exact Value + Variation
Error can be of two types:
1 Bias
2 Chance Error
Measurement = Exact Value + Bias + Chance Error
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 40: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/40.jpg)
Measurement Error
Fact
Any measurement has error associated with it.
Measurement=Exact Value + Variation
Error can be of two types:
1 Bias
2 Chance Error
Measurement = Exact Value + Bias + Chance Error
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 41: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/41.jpg)
Measurement Error
Fact
Any measurement has error associated with it.
Measurement=Exact Value + Variation
Error can be of two types:
1 Bias
2 Chance Error
Measurement = Exact Value + Bias + Chance Error
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 42: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/42.jpg)
Measurement Error
Fact
Any measurement has error associated with it.
Measurement=Exact Value + Variation
Error can be of two types:
1 Bias
2 Chance Error
Measurement = Exact Value + Bias + Chance Error
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 43: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/43.jpg)
Measurement Error
Fact
Any measurement has error associated with it.
Measurement=Exact Value + Variation
Error can be of two types:
1 Bias
2 Chance Error
Measurement = Exact Value + Bias + Chance Error
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 44: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/44.jpg)
Chance Error
Fact
Chance Error is due to randomness.
It has zero average
If bias is zero then we can �nd the exact value by measuring
lots of times and taking the average.
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 45: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/45.jpg)
Chance Error
Fact
Chance Error is due to randomness.
It has zero average
If bias is zero then we can �nd the exact value by measuring
lots of times and taking the average.
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 46: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/46.jpg)
Chance Error
Fact
Chance Error is due to randomness.
It has zero average
If bias is zero then we can �nd the exact value by measuring
lots of times and taking the average.
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 47: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/47.jpg)
Chance Error
Fact
Chance Error is due to randomness.
It has zero average
If bias is zero then we can �nd the exact value by measuring
lots of times and taking the average.
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 48: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/48.jpg)
How can we spot bias?
Fact
There is no way to tell if bias exists without comparison with
another kind of experiment.
Example
To determine the age of a tree one can do the followingexperiments:
carbon dating
tree ring counting
gene variation analysis
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 49: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/49.jpg)
How can we spot bias?
Fact
There is no way to tell if bias exists without comparison with
another kind of experiment.
Example
To determine the age of a tree one can do the followingexperiments:
carbon dating
tree ring counting
gene variation analysis
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 50: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/50.jpg)
Outliers
De�nition
If one or two data points are wild results, statisticians oftenthrow them out.
They assume that something fundametally di�erent occuredfor that point.
For example, a bias for that measurement that is not in othermeasurements.
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 51: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/51.jpg)
Outliers
De�nition
If one or two data points are wild results, statisticians oftenthrow them out.
They assume that something fundametally di�erent occuredfor that point.
For example, a bias for that measurement that is not in othermeasurements.
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 52: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/52.jpg)
Outliers
De�nition
If one or two data points are wild results, statisticians oftenthrow them out.
They assume that something fundametally di�erent occuredfor that point.
For example, a bias for that measurement that is not in othermeasurements.
Marius Ionescu Unit 3: The Normal Approximation for Data
![Page 53: Unit 3: The Normal Approximation for Data](https://reader030.vdocument.in/reader030/viewer/2022012506/618112f3d2a0a835215c6df3/html5/thumbnails/53.jpg)
Outliers
De�nition
If one or two data points are wild results, statisticians oftenthrow them out.
They assume that something fundametally di�erent occuredfor that point.
For example, a bias for that measurement that is not in othermeasurements.
Marius Ionescu Unit 3: The Normal Approximation for Data