How linear?

Report 0 Downloads 39 Views
How linear?

Ani Adhikari and Philip Stark

Statistics 2.1X

Lecture 6.2

1/7

y

y

0

0

0

-1

-1

-1

-2

-2

-2

-3

-3

y

1

1

1

2

2

2

3

3

3

How linear?

-2

-1

0 x

Ani Adhikari and Philip Stark

1

2

-2

-1

0 x

Statistics 2.1X

1

2

-2

-1

0

1

2

x

Lecture 6.2

1/7

y

y -3

-3

-2

-2

-2

-1

-1

-1

0

0

0

y

1

1

1

2

2

2

3

3

3

How linear?

-2

-1

0 x

1

2

-2

-1

0 x

1

2

-2

-1

0

1

2

x

correlation coefficient (r ): a number between −1 and 1; it measures linear association, that is, how tightly the points are clustered about a straight line.

Ani Adhikari and Philip Stark

Statistics 2.1X

Lecture 6.2

1/7

Example

Data: (1, 2) (2, 3) (3, 1) (4, 6) (5, 6)

Ani Adhikari and Philip Stark

Statistics 2.1X

Lecture 6.2

2/7

Ani Adhikari and Philip Stark

4 y 3 2 1

Data: (1, 2) (2, 3) (3, 1) (4, 6) (5, 6)

5

6

Example

1

2

3

4

5

x

Statistics 2.1X

Lecture 6.2

2/7

4 y 3 2 1

Data: (1, 2) (2, 3) (3, 1) (4, 6) (5, 6)

5

6

Example

1

2

3

4

5

x

Expect r to be positive but not 1.

Ani Adhikari and Philip Stark

Statistics 2.1X

Lecture 6.2

2/7

Calculating r

Ani Adhikari and Philip Stark

Statistics 2.1X

Lecture 6.2

3/7

Calculating r

x 1 2 3 4 5 mean = 3 SD = 1.41

y 2 3 1 6 6 mean = 3.6 SD = 2.06

Ani Adhikari and Philip Stark

x in std. units −1.41 −0.71 0 0.71 1.41

Statistics 2.1X

y in std. units −0.78 −0.29 −1.26 1.16 1.16

product of std. units 1.10 0.21 0 0.82 1.64 mean = 0.75 =r

Lecture 6.2

3/7

The formula in two languages Formula for r 1. Convert both lists to standard units. 2. Multiply corresponding pairs of standard units. 3. r is the average of the products.

Ani Adhikari and Philip Stark

Statistics 2.1X

Lecture 6.2

4/7

The formula in two languages Formula for r 1. Convert both lists to standard units. 2. Multiply corresponding pairs of standard units. 3. r is the average of the products.

For those who like math notation and have read the algebra supplement: If the data are (xi , yi ), 1 ≤ i ≤ n, then n

1 X  xi − µx  yi − µy  r = n σx σy i=1

Ani Adhikari and Philip Stark

Statistics 2.1X

Lecture 6.2

4/7

Properties of r

1. The calculation uses only standard units. So r is a pure number with no units.

Ani Adhikari and Philip Stark

Statistics 2.1X

Lecture 6.2

5/7

Properties of r

1. The calculation uses only standard units. So r is a pure number with no units. 2. −1 ≤ r ≤ 1. Trust me.

Ani Adhikari and Philip Stark

Statistics 2.1X

Lecture 6.2

5/7

Properties of r

1. The calculation uses only standard units. So r is a pure number with no units. 2. −1 ≤ r ≤ 1. Trust me. The extreme cases: r = −1 is when the scatter is a perfect straight line sloping down; r = 1 is when the scatter diagram is a perfect straight line sloping up.

Ani Adhikari and Philip Stark

Statistics 2.1X

Lecture 6.2

5/7

Properties of r

1. The calculation uses only standard units. So r is a pure number with no units. 2. −1 ≤ r ≤ 1. Trust me. The extreme cases: r = −1 is when the scatter is a perfect straight line sloping down; r = 1 is when the scatter diagram is a perfect straight line sloping up. 3. It doesn’t matter if you switch the variables x and y ; r stays the same.

Ani Adhikari and Philip Stark

Statistics 2.1X

Lecture 6.2

5/7

20 0

10

final

30

40

Switching axes doesn’t affect linearity

0

5

10

15

20

midterm

Ani Adhikari and Philip Stark

Statistics 2.1X

Lecture 6.2

6/7

20 15 0

0

5

10

midterm

20 10

final

30

40

Switching axes doesn’t affect linearity

0

5

10

15

20

midterm

Ani Adhikari and Philip Stark

0

10

20

30

40

final

Statistics 2.1X

Lecture 6.2

6/7

Linear transformations

4. Adding a constant to one of the lists just slides the scatter diagram, so r stays the same.

Ani Adhikari and Philip Stark

Statistics 2.1X

Lecture 6.2

7/7

Linear transformations

4. Adding a constant to one of the lists just slides the scatter diagram, so r stays the same. 5. Multiplying one the lists by a positive constant does not change standard units, so r stays the same.

Ani Adhikari and Philip Stark

Statistics 2.1X

Lecture 6.2

7/7

Linear transformations

4. Adding a constant to one of the lists just slides the scatter diagram, so r stays the same. 5. Multiplying one the lists by a positive constant does not change standard units, so r stays the same. 6. Multiplying just one (not both) of the lists by a negative constant switches the signs of the standard units of that variable, so r has the same absolute value but its sign gets switched.

Ani Adhikari and Philip Stark

Statistics 2.1X

Lecture 6.2

7/7