Research Methods A2 L9 - 12 (Type I and II errors, testing)

Question 1

Type 1 error:

Answer 1

• Rejected null hypothesis and accepted experimental hypothesis
• However, the results are actually due to chance rather than statistical significance

Question 2

Type 2 error:

Answer 2

• Rejected experimental hypothesis and accepted null hypothesis
• However, the results are due to statistical significance rather than chance

Question 3

What are levels of measurement used for?

Answer 3

Categorise data into one of 3 types

Question 4

What are the 3 types of data?

Answer 4

• Nominal
• Ordinal
• Interval

Question 5

Nominal data:

Answer 5

• Organised into categories
• Fit in one category only not the other

Question 6

Ordinal data:

Answer 6

• Can be placed in rank order
• Can consist of measurements eg height

Question 7

Parametric tests

Answer 7

• More robust and powerful
• Based on actual data collected rather than rank order
• More likely to detect significance

Question 8

What are the 3 parametric tests?

Question 9

What 3 factors mean a parametric test can be conducted?

Answer 9

• Level of measurement must be interval
• Data must form a normal distribution
• Data should have similar variance of scores

Question 10

What are the 4 non-parametric tests?

Answer 10

• Chi Squared
• Spearman’s Rho
• Mann Whitney
• Wilcoxon

Question 11

What 3 questions should be asked when deciding the test that should be used?

Answer 11

1) Does the research involve a correlation, a test of difference or an association?
2) Which research design is being used?
3) What level of measurement is being used?

Question 12

Correlation, ordinal:

Answer 12

Spearman’s rho

Question 13

Correlation, interval:

Answer 13

Pearson’s r

Question 14

Association:

Answer 14

Chi squared

Question 15

Test of difference, independent measures, nominal:

Answer 15

Chi squared

Question 16

Test of difference, independent measures, ordinal:

Answer 16

Mann Whitney U

Question 17

Test of difference, independent measures, interval:

Answer 17

Unrelated t-test

Question 18

Test of difference, repeated measures/matched pp, nominal:

Answer 18

Sign test

Question 19

Test of difference, repeated measures/matched pp, ordinal:

Answer 19

Wilcoxon

Question 20

Test of difference, repeated measures/matched pp, interval:

Answer 20

Related t-test

Question 21

Structure to answer whether results are significant or not:

Answer 21

• Not significant/significant
• Because calculated value is greater/less than critical value
• when p = ? And n = ? Under a one/two tailed test
• Therefore, experimental hypothesis should be accepted/rejected and null hypothesis should be rejected/accepted

Question 22

N:

Answer 22

Number of pps

Question 23

d:

Answer 23

Difference

Question 24

Na:

Answer 24

Number of Pps in smaller sample

Question 25

Nb:

Answer 25

Number of Pps in larger sample

Question 26

T:

Answer 26

Sum of ranks in smaller sample

Question 27

Observed value:

Answer 27

Calculated value

Question 28

Degrees of freedom:

Answer 28

Number of values that are free to vary

Question 29

Contingency table:

Answer 29

2 way table

Question 30

How is Wilcoxon calculated (2):

Answer 30

1) Calculate sum of positive ranks
2) Calculate sum of negative ranks

Question 31

What does (N-1) mean?

Answer 31

Number of Pps minus 1