Correlation Flashcards

1
Q

Correlation….

A

measures and describes how two variables relate to one another

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Three characteristics of correlation

A

Direction (+/-)
Form (linear)
Strength (0-1)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Pearson correlation (r)

A

Measures the degree and direction of the linear relationship of the 2 variables

Perfect linear relationship (change in x has a corresponding change in y)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Sums of Products (SP)

A

Similar to sums of squares

Measures the amount of coverability between two variables

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

SP definitional formula

A
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

SP Computational formula

A
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Pearson Correlation Calculation (Formula)

A

Ratio comparing coverability of x and y (numerator) with the variability of X and Y separately (denominator)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Partial correlation…

A

Measures the relationship between two variables while controlling the influence of a third variable by holding it constant

Formula:

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Pearson correlation can be expressed as a relationship of z scores (formulas)

A
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Correlation is used for…

A

Prediction
Validity
Reliability
Theory verification

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Correlation does not…

A

Mean causation

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Correlation coefficient is affected by range…

A

therefore, never generalize correlation beyond the sample range data

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

An (extreme deviant) outlier has…

A

A disproportionately large impact on the correlation coefficient (r)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Correlation is not a…

A

Proportion of variability

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Squared correlation (coefficient of determination)…

A

(r^2) is interpreted as the proportion of shared variability

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Pearson correlation is usually computed for sample data, but…

A

used to test hypothesis’ about the relationship in the population

17
Q

Population correlation is shown by…

A

Greek letter rho (ρ)

Non-directional: H0:
ρ = 0 and H1: ρ ≠ 0

Directional: H0: ρ ≤ 0 and H1: ρ > 0

or

Directional: H0: ρ ≥ 0 and H1: ρ < 0

18
Q

Sample correlation is used to test population ρ

A

Degree of freedom
n-2

Hypothesis test can be computed using either t or F

Use to t table to find critical value with df=n-2

19
Q

Reporting Correlation in Literature

A

Report
- Whether it is statistically significant

Concise test results
- Value of correlation
- Sample Size
- p-value or level
- Type of test (one- or two- tailed)

Example
r=-.76, n=48, p<.01, two tails

20
Q

What type of data has Pearsons’s correlation been developed for?

A

Linear relationships
Interval or ratio measurement scales

21
Q

Other correlations have been developed for…

A
  1. Non-linear relationships
  2. Data with nominal or ordinal measurement scales
22
Q

Spearman (rs) correlation formula is used with…

A

Data from an ordinal scale (ranks)

  1. Both variables are measured on an ordinal scale
  2. May be used when measurement scale is interval or ration when relationship is consistently directional but not linear
23
Q

Ranking Tied Scores

A

Tie scores need for Spearman correlation

Method:
- List ranks smallest to largest
- Assign a rank to each position in the list
- When two (or more) scores are tied, compute the mean of their ranked position, and assign this mean value as the final rank for each score

24
Q

Special formula for Spearman Correlation

A

The ranks for scores are simply integers

Use D as the difference between the X rank and the Y rank for each individual to compute the rs statistic

Formula:

25
Q

Point-Biseral Corrlation…

A

measures the relationship between two variables
Also r^2

  • One continuous variable (e.g., Height)
  • One dichotomous variable (e.g., Male=1, female=0)
26
Q

Phi coefficent…

A

Measures the relationship between two dichtomous variables
- Both coded 0 and 1
- Regular Persons correlation is used

27
Q

r=.50

A

Large correlation

28
Q

r=.30

A

Moderate correlation

29
Q

r=.10

A

Small correlation

30
Q

Proportion of variance accounted for (coefficient of determination)

31
Q

Confidence Interval on r

A

It is easy to work out a confidence interval for r too. It works just like the CI on the mean. It is centred on the observed value of r and extends a set number of standard errors in each direction …

32
Q

Hypothesis Testing (r) steps

A
  1. State Hypothesis
  2. Determine alpha and critical values
  3. Compute Pearson r
  4. Compute effect size
  5. Report the results