Week 3 Video Notes

Question 1

r = coefficient of correlation = pearson’s r

Answer 1

-1<r<+1 (or equal to -1, 1)

Question 1

types of relationships illustrated by scatter point graphs

Answer 1

perfect negative -1
strong negative -0.76 to -0.99
moderate negative -0.25 to -0.75
weak negative -0.1 to -0.24
no correlation 0
weak positive 0.1 to 0.24
moderate positive 0.25 to 0.75
strong positive 0.76 to 0.99
perfect positive 1

Question 2

r equation

Answer 2

n(SIG XY) - (SIG X)(SIG Y) / SQRROOT [n(SIG X^2) - (SIG X)^2][n(SIG Y^2) - (SIG Y)^2]

Question 3

coefficient of determination

Answer 3

r^2
- shows the percentage in changes in x value can be seen in y
- percentage

Question 4

curvilinear

Answer 4

strong correlation but NOT linear (upside-down U-shaped)

Question 5

regression analysis (least-squares method)

Answer 5

‘predicting for the future’
- we use regression equation (equations for the line, close to most of the points, to predict for the other values, or for future - LINEAR MODEL): (y-bar/ y-hat) y=bx + a
- y-hat is the predicted value

Question 6

residual

Answer 6

the difference between an observed value (actual value) and the predicted value (thru regression equation)
- residual = observed value - predicted value

Question 7

least squares method

Answer 7

residual^2 for all points (to find area of each square
- minimize SIG(RESIDUALS)^2 (sum of area of squares)

Question 7

regression line

Answer 7

y = bx + a
- b = r * STDEV(Y) / STDEV(X)
- a = y(bar) - bx(x-bar)
SHORCUT
a = ((SIG Y) / n) - b ((SIG X)/n)
b = n (SIG XY) - (SIG X)(SIGY ) / n (SIG X^2) - (SIG X)^2