Errors

Question 1

What are errors?

Answer 1

When a conclusion from an inferential test is inaccurate

Question 2

What are the two types of error?

Answer 2

Type I
Type II

Question 3

Type I errors:

Answer 3

• WHEN RESULTS ARE ACTUALLY DOWN TO CHANCE
• when conducting a test, it may show a significant difference/ correlation and based on that we accept the alternate hypothesis and reject the null HOWEVER:
• in reality, this difference/correlation was down to chance and we should’ve accepted the null and rejected the alternate

Question 4

Type II

Answer 4

• WHEN RESULTS AR ACTUALLY DOWN TO A GENUINE DIFFERENCE/ CORRELATION (false negative).
• conclusion that results are not significant and due to chance so accept the null and reject the alternate
• HOWEVER: in reality, there is a genuine difference/ correlation so need to accept alternate and reject null

Question 5

When are Type I errors more likely to happen?

Answer 5

• The chance of a Type I error is the same as the significance level we set. This is typically P≤0.05 so 5% is the chance of a Type I error.
• When a p-value or significance level is of a more lenient value, (e.g. P≤0.1) there is a higher chance we make a Type I error (e.g.10%)

Question 6

To reduce the chance of a Type I error:

Answer 6

Make the p-value more strict (e.g. P≤0.01 =1%)
- Limitation: we are more likely to get Type II errors with a more stringent P-value

Question 7

When are Type II errors more likely to happen

Answer 7

• When we use a more stringent (strict) p-value
• we only accept results are significant if there is a very low probability that they happened by chance
• genuine differences/correlations will often be dismissed as due to chance

Question 8

Explain why the researcher decided to use the 5% level of significance rather than the 1% level in this study.
[2 marks]

Answer 8

• the 5% level of significance is the conventional level of probability employed by
  psychologists/balances the risks of making a Type I and Type II error
• the researcher is investigating the relationship between _______________, this is not a sensitive topic nor one which may affect individual’s health (as in clinical trials).