Statistical Testing

Question 1

Mann Whitney

Answer 1

• Unrelated data
• Data is at least ordinal (meaning it can be interval too).
• You are predicting a difference between two conditions.

Question 2

Unrelated t Test

Answer 2

• Unrelated data
• Data is at least interval (Cannot be used on ordinal or nominal data)
• Predicting a difference between two conditions.

Question 3

Related t Test

Answer 3

• Related data
• The data is at least interval (meaning it cannot be used on ordinal or nominal data).
• You are predicting a difference between two conditions.

Question 4

Wilcoxon

Answer 4

• Data is related
• The data is at least ordinal (meaning it can be interval too)
• You are predicting a difference between two conditions.

Question 5

The Sign Test

Answer 5

• Data is related
• The data is at least nominal (the test can be used if data is ordinal or interval too).
• You are predicting a difference between two conditions.

Question 6

Chi Square

Answer 6

• Data is unrelated
• The raw data is nominal
• Can be used for associations and differences between variables.

Question 7

Spearman’s rho

Answer 7

• Data is related (by definition as it is used for correlations only).
• The raw data is at least ordinal (meaning it can be interval too).
• A correlational method was used. (i.e you are looking at the relationship between two variables).

Question 8

Pearson’s r

Answer 8

• Data is related (by definition as it is used for correlations only).
• The raw data must be interval.
• A correlational method was used. (i.e you are looking at the relationship between two variables).

Question 9

Carrots Should Come…

Answer 9

• Carrots Should Come
• Mashed With Swede
• Under Roast Potatoes

—- Unrelated / Related / Correlation
Nom. (Chi) (Sign) (Chi)
Ord. (Mann.W) (Wilcox) (Spear.)
Int. (Un T.test) (Rel T.test) (Pears)

Question 10

How to: Sign Test

Answer 10

Step 1: Record the scores in a table
Step 2: Work out the differences between each set of scores
Step 3: Record whether each difference is a positive or negative value
Step 4: Work out ‘S’. This is known as the calculated or observed value. S is the number of the least frequent sign. In the example above we have 5 minus signs and 2 plus signs so S is 2 Step 5: Compare the calculated value to a critical value in a table:

To compare the calculated (observed) value to the critical value you need the following information:
> N = number of participants. If any participants had the same score in both conditions omit them
> One tailed is a directional hypothesis, two tailed is a non directional hypothesis.
> The level of significance is the probability of the test giving you the wrong information (as the tests aren’t 100% accurate). 0.05 is the most commonly used
> For the sign test a calculated value is considered significant if it is equal to or less than the critical value (the value in the table).
> If it is significant it means the hypothesis can be supported. If not, it means the null hypothesis must be accepted.