Statistical Testing

When to use the sign test?

1. You must have a DIFFERENCE not association
2. Repeated measures design
3. Nominal data (categories)

Concept of Probability

Probability: likelihood of certain events occurring
Sample of participants
Null hypothesis: no difference/association in population
Alternative hypothesis: Difference found in sample → true for population

Accepted probability/level of sig. = 0.05 (5%)
(sometimes an even higher level of sig/certainty is needed - 0.01 → e.g. new drug trial (human cost))

“This suggests” > “This proves”

The critical value

statistical test calculated → researcher left with calculated value

compare calc v. to critical value

critical values given in table
to use table you need
1. significance level - usually 0.05
2. number of participants
3. hypothesis is directional (one-tailed) or non-directional (2-tailed → 0.10)

Sign test

1. convert to nominal data - subtract data of second column from first - positive number = plus sign, negative number = minus sign

2. total plus number & total minus number. void equals

3. less frequent sign = calc value

4. compare calc to crit - 0.05
   number of participants = N → determines row on table
   calc value must be equal to or less than the crit value at sig level of 0.05 = significant
   calc value = more → not significant = accept null