Topic 4 (statistics)

Question 1

correlation

Answer 1

measure of linear relationship between 2 variables (straight line of best fit)

-only gives evidence of linear relationship

Question 2

types of correlation

Answer 2

positive (stronger correlation stronger relationship between variables) (majority of data in 1st and 3rd quadrant)

negative (majority of data in 2nd and 4th quadrant)

zero (could still mean non linear relationship) (points in all 4 quadrants

Question 3

Bivariate data

Answer 3

data for which each item requires the values of 2 variables

Question 4

independent variable

Answer 4

explanatory variable
one that is set independently of other variable
plotted along x-axis

Question 5

dependent variable

Answer 5

response variable
values are determined by values of independent variable
plotted along y-axis

Question 6

spurious

Answer 6

correlation without a causal connection

Question 7

causal

Answer 7

one variable has an impact on other

Question 8

Product moment correlation coefficient (PMCC)

Answer 8

-measure of strength of correlation

Question 9

PMCC

Answer 9

r= 1 perfect positive linear correlation (exact straight line with positive gradient)

r= -1 perfect negative linear correlation (exact straight line with negative gradient)

r=0 no linear correlation (may be relationship of some sort)

remains unchanged through coding (if coded using linear transformation to make data manageable PMCC is same)

Question 10

correlation & regression:

Answer 10

• methods for investigating relationship between variables in statistics
  correlation: measure to which 2 variables are related
  regression: describing the relationship between 2 variables (eq of line of best fit for data)

Question 11

interpolated value

Answer 11

sits within the plotted range of data

Question 12

extrapolated value

Answer 12

sits outside range of plotted values

Question 13

residual

Answer 13

compares the true value compared to the value on the regression line

• data value above regression line = residual >0
• data value below regression line = residual <0

perfect scenario:
sum of residuals = 0
sum of residuals squared = as close to 0 as possible

Question 14

regression line

Answer 14

of y on x
y= a + bx

a= y int 
b = gradient (b>0 PMCC >0).  (b<0 PMCC <0)

Question 15

interpolation

Answer 15

estimating dependent variable within range of data

Question 16

extrapolation

Answer 16

estimating dependent variable outside range of data