Inferential Statistics

Question 1

What is the role of inferential statistics?

Answer 1

• Allow researchers to determine whether results are significant
• In this case, significant means that there is an actual difference/correlation between findings and is not due to chance
• Not significant = no actual difference/correlation - results were due to chance
• Neither = outcome still tells us something useful

Question 2

How are inferential statistics different to descriptive statistics?

Answer 2

• Explain characteristics of the data e.g. the means, standard deviations but not whether the difference is significant
• Significance can be subjectively interpreted from descriptive statistics whereas inferential statistics explain significance objectively

Question 3

What is the difference between a one-tailed and two-tailed test?

Answer 3

• One-tailed tests have a directional hypothesis and a specific outome being tested
• Whereas two-tailed tests have a non-directional hypothesis as a non-specific outcome is being tested

Question 4

What is a p-value?

Answer 4

• The p-value refers to the probability findings have occurred due to chance
• Shows us if the difference in conditions is a fluke or due to IV manipulation
• Written as a % or decimal
• Low p value e.g. p<0.01 (1% likelihood findings are due to chance) = no pevious research on the topic or the researcher needs to be confident of the result e.g. testing drug efficiency
• Higher p value e.g. p<0.1 (10% likelihood findings are due to chance) = repeating existing research so the researcher doesn’t need to be confident of the result
• p<0.05 (5%) is the standard significance level in psychology

Question 4

When do we use a sign test?

Answer 4

To judge if difference in conditions are significant. Only used if the experiment was:
- a test of difference between conditions
- repeated measures or matched pairs design
- using nominal data - in categories or can be converted into categories

ALL 3 OF THIS CRITERIA MUST BE MET

Question 5

How do we do a sign test?

Answer 5

1. State the hypothesis + identify if one-tailed or two-tailed
2. Gather and record data
3. Work out if the difference between scores is positive, negative, or 0
4. Work out the calculated value for S: add up how many +s there are and how many -s there are and ignore any 0s. Whatever value is smaller is the calculated S value
5. Find the critical value of S from the critical values table
6. Report the conclusion and state which hypothesis to accept

Question 6

How do you write a conclusion for a sign test?

Answer 6

Either of these:
1. Since the calculated value of S ( _ ) is LESS than/equal the critical value of S ( _ ), the difference in conditions is significant for N = _ and p<0.05
2. Since the calculated value of S ( _ ) is MORE than the critical value of S ( _ ), the difference in conditions is NOT significant for N = _ and p<0.05