choosing and interpreting a statistical test

Question 1

What is the purpose of a statistical test?

Answer 1

• tells us which hypothesis is most likely to be true
• to determine the likelihood that the results/difference/correlation are due to chance or iv/co-variables
• if our results are significant or not significant

Question 2

What is the accepted significance level?

Answer 2

• conventional level
• p < 0.05
• less than 5% probability that the results occurred by chance/EV
• more than 95% probability that the results were due to changing IV/relationship between co-variables

Question 3

Process of inferential/statistical testing…

Answer 3

1. carry out statistical tests
2. gives us the calculated/observer
3. compare to a critical value (from the table)

• critical value decides whether or not calculated value is significant
• calculated value is significant = accept alternative hypothesis
• calculated value not significant = accept null hypothesis

Question 4

3 things to identifying or justifying a stats test…

Answer 4

1. test of difference OR correlation
2. experimental design: repeated measures, matched pairs, independent groups
3. level of measurement: nominal, ordinal, interval

Question 5

NO RIC table…

Answer 5

N = nominal
O = ordinal
! = interval
R = repeated measures (or matched pairs)
I = independent groups
C = correlation
STUDENTS = sign test
COME = chi squared
COLLEGE = chi squared
WHEN = wilcoxon
MISS = mann whitney u
SAYS = spearman’s rho
RIC = related t test
U = unrelated t test
PART-TIMER = pearson’s r

Question 6

Writing frame: choosing a stats test

Answer 6

The appropriate statistical test is …
This is because the study is a test of difference/correlation between … and …
They used an …. experimental design because CONTEXT (only use if it is a test of difference)
The level of measurement is … because CONTEXT

Question 7

How to decide whether results are significant…

Answer 7

1. One (directional) or two (non-directional) tailed hypothesis?
2. Number of ppts (N=) or degrees of freedom (df)
3. Level of significance
4. Identify the calculated value (in question)
5. Identify the critical value (in table)
6. interpret the findings using the statement under the critical value table

Question 8

Calculated value must be equal to or LESS than critical value…

Answer 8

• sign test
• mann whitney u
• wilcoxon

Question 9

Calculated value must be equal to or MORE than critical value…

Answer 9

• moRe
• chi squaRed
• spearman’s Rho
• unRelated t-test
• Related t-test
• pearson’s R

Question 10

Writing frame: Are the results significant?

Answer 10

The critical value is … due to the P value being …
a …-tailed hypothesis and where N/df=…
The calculated/observed value (…) must be … than or equal to the critical value (…) to be significant.
In this case the results are … and we reject the … hypothesis and accept the …
So there IS/IS NOT a significant DIFFERENCE/CORRELATION between …