Correlations

Question 1

Correlation is often the first step in what?

Answer 1

identifying potential causal relationships

Question 2

What does correlation express

Answer 2

quantitatively the size and direction of relationship between variables

Question 3

Correlation coefficient

Answer 3

statistic which expresses numerically the magnitude and direction of association between two variables

Question 4

3 types of different relationships

Answer 4

positive linear correlation
negative linear correlation
non-linear correlation

Question 5

Correlation coefficient characteristics: calculated from

Answer 5

pairs of measurements on variables x and y for the same group of individuals

Question 6

Correlation coefficient characteristics: values range from

Answer 6

+1 to -1

Question 7

Correlation coefficient characteristics: +1 means what, 0 means what, and -1 means what?

Answer 7

+1 = perfect positive correlation
-1 = perfect negative correlation
0 = no correlation

Question 8

Correlation coefficient characteristics: square of the correlation coefficient is

Answer 8

coefficient of determination

Question 9

The type of correlation coefficient expressed depends on

Answer 9

scale of measurements

Question 10

Condition where phi is appropriate

Answer 10

both x and y measure on a nominal scale

Question 11

Conditions where rho is appropriate

Answer 11

both x and y measure on, or transformed to, ordinal scales

Question 12

Conditions where r is appropriate

Answer 12

both x and y measure on an interval ratio scale

Question 13

Pearson’s r is for when

Answer 13

both x and y measured on interval or ratio scale

Question 14

r assumes what?

Answer 14

both variables x and y are normally distributed, and have linear relationship

Question 15

Pearson’s r is a measure of

Answer 15

the extent to which paired scores are correlated

Question 16

To calculate r, we need to represent the

Answer 16

position of each paired score within its own distribution

Question 17

To calculate r, we need to represent the position of each paired score within its own distribution: convert each

Answer 17

raw score to z or standard score

Question 18

To calculate r, we need to represent the position of each paired score within its own distribution: corrects for

Answer 18

x and y having different means and standard deviations

Question 19

The closer the relationship between two variables approximates a _________ the more r approaches _______

Answer 19

straight line
1.0

Question 20

In biological/social sciences, r is not usually _______

Answer 20

1.0

Question 21

r > 0.7 =

Answer 21

quite high

Question 22

r = 0.3 - 0.7 =

Answer 22

moderate

Question 23

r < 0.3 =

Answer 23

weak

Question 24

Spearman’s p

Answer 24

when at least one of x or y measured on ordinal scale

Question 25

Spearman's p can readily convert

Answer 25

variable on higher scale to ordinal scale

Question 26

GO OVER EXPLANATION OF TERMS FOR SPEARMAN'S P

Answer 26

GO OVER EXPLANATION OF TERMS FOR SPEARMAN'S P

Question 27

2 uses of correlation in the health sciences

Answer 27

prediction reliability and predictive validity assessment

Question 28

Estimating shared variance

Answer 28

coefficient of determination (r^2)

Question 29

Coefficient of determination

Answer 29

proportion of variance in one variable accounted for by the other

Question 30

Criteria of a causal relationship: the ________ must ________ the _______ in time

Answer 30

cause precede effect

Question 31

Criteria of a causal relationship: _________ of cause and effect

Answer 31

covariation

Question 32

Criteria of a causal relationship: _________ of _______ explanations

Answer 32

elimination rivals

Question 33

Even high correlation alone does not

Answer 33

necessarily indicate causal relationship

Question 34

Correlations can provide

Answer 34

hypotheses for further testing

Question 35

If using sample to estimate correlation in population, could also have

Answer 35

biased sampling or sampling error

Question 36

If there is biased sampling or sampling error you need to show

Answer 36

correlation coefficient has statistical significance