# Statistical test

HideShow resource information

## Mann Whitney U test

Mann Whitney U test - When to use

• When hypothesis predicts a difference between 2 sets of data
• When the 2 sets of data are from seperate groups (Independent measures)
• Condition 1 and 2 DO NOT have to have the same amount of participants
• Data is ordinal or interval NOT nominal

How it works

• Assigning a point score to each result - If they're the same value do the rank in the middle
• Point for each set of data are added up and the lower of the two scores is our U
• U is the lowest number of points

Oberserved value must be LESS THAN critical value to REJECT null hypothesis

1 of 4

## Chi-Square

Chi-Square - When to use

• Used when the hypothesis predicts a difference and Nominal data (random)
• Used when counting occurance in a category
• Must have at least 20 participants and set of data must be independent

Method

1) Do all of the totals in the contingency table including the total of the whole table

2) E = total of row X total of column DIVIDED by total

3) X squared = (total of cell - E) squared DIVIDED by E

4) Calculate the degrees of freedom (row-1) X (columns-1)

5) X squared value must be equal to or exceed that found in the table

2 of 4

## Spearman's Rho

Spearman's Rho When is it used?

• Test of correlation- to see if relationship is significant between two variables
• +1.0 = Positive correlation  0 = No correlation   -0.1 = Negative correlation
• Data Ordinal or Interval NOT nominal

Method

1) Scores must be ranked and then the difference between the ranks calculated

2) Square the 'd' and then add them together

3) complete the equation rho = 1 - 6 X (total of d squared) DIVIDED by n(n squared -1)

If observed value is higher than critical value then significant correlation

3 of 4

## Wilcoxon T test

Wilcoxon T test - When to use

• Use when hypothesis predicts a difference
• 2 sets of data are from 1 person - repeated measures design
• Data needs to be ordinal or interval NOT nominal

Method

1) Rank and work out the difference between the ranks

2) Find T - the sum of the ranks of the less frequent sign

Observe value same or lower than critical value there is a significant difference

4 of 4