Psychology 320: Psychological Statistics

Professor: Howard B. Lee

Lecture Notes

Week 5: Chapters 6 and 7

Lecture 12

Data Table:

People	x	y	xy
1	4	10	40
2	8	9	72
3	6	14	84
4	3	8	24
5	5	10	50
6	7	17	119
7	2	5	10

Mx = 5
My = 10.429
SumXY = 399
SDx = 2.00
SDy = 3.659

How do you find a value and b value?

Regressing X onto Y

Predicting Y onto X

Raw score form
Look at page 152 in your book.

a1 is not equal to a2

a = r ( SDy/ SDx )

1. Use special function keys for SD.
2. Find the XY cross products.
3. Find the correlation ( r ).

b = My - r ( SDy / SDx ) Mx

1. Find the mean and SD of X and of Y.
2. Solve the above equation.
3. Find the correlation ( r ).

Make sure that you use .6637 in the continuing calculations.

*For Z scores, correlations, and regressions -----> carry to 3 significant digits, round to 2

a = .664 ( 3.659 / 2 )
a = 1.215 or 1.22

b = (10.429 - 1.215 ( 5 ))
b = 10.429 - 6.075
b = 4,354 or 4.35

Final equation Y'= 1.22X + 4.35

The least squares line predicting Y on the basis of Xx.
The formula means that as 1.22X + 4.35 changes, Y changes.
So if you plug in 4 for x variable, 1.22 (4) + 4.35 = 9.23
Therefore, 9.23 is the predicted value for y.

Standard score form

Z'y = r Zx

Predicted standard score of Y = correlation (Standard score of X)

* On the test, if it says " predict y..." that is in raw score form. If it says " find the standard score of y..." that is in standard score form.

success in graduate school = a GPA + b

final = a Midterm + b

Multiple regression : 1 dependent variable and more than 1 independent variables.

For ex. GPA, GRE, letters of recommendation, quality of school, autobiography are used in a multiple regression to predict success in graduate school.

Lecture 13

Review for exam...