(4) STATISTICAL TESTING

Question 1

What does statistical testing aim to achieve

Answer 1

Statistical Testing:
• Know whether to accept/reject hypothesis
• Are variable relationships statistically significant or by chance?
• Ensures diffs between mean of two conditions is large enough

Question 2

What is a significance level used for and what is the standard

Answer 2

• Significance level
○ Used to find if results are due to chance or due to the IV
○ Standard significance level = 5%
§ Use this unless otherwise specified in test

Question 3

How is a sign test carried out?

Answer 3

1. Sign Test
   § Determines whether the diff between conditions is significant
   § Used when:
   □ We look for a DIFFERENCE rather than ASSOCIATION
   □ Repeated measures design (or matched pairs in rare cases)
   1. Convert data into + or - sign (consistently for all participants)
   2. Add up no. of each sign
   3. Find least frequent sign and call it S (eg S = 6)
   □ Calculated value
   4. Critical value is found using table and compared to calculated value. If S value =< critical value, result is significant

Question 4

What are the 3 things we look at when choosing a sign test

Answer 4

To choose which stats test to use, think of 3 D’S:
1. Levels of data
○ Look at DV
○ Nominal
§ Data that can be put into specific categories
□ Eg green cars, blue cars in a car park
○ Ordinal
§ Data with an order of ranking
□ Eg rate stress from 1 to 10
§ Differences between ranks may not be equal
□ Eg babette may think 6 is a higher level of stress than dalia
○ Interval
§ Type of data where numbers = meaningful order
□ Eg weight. A kg is always a kg
2. Design
○ Related designs
§ Repeated measures
□ Same ppts used in all conditions
§ Matched pairs
□ Ppts related to each other
○ Unrelated designs
§ Independent groups
3. Test of difference/correlation

Question 5

Draw out the stats test table

Answer 5

(Add photo)

Question 6

What are the two types of errors

Answer 6

Type I, Type II

Question 7

What is a Type I error

Answer 7

• null rejected, alternative accepted
• should be other way around
• ‘false positive’
• researchers say say there is sig diff when there isn’t
• usually when sig level is 0.1/10%

Question 8

What is a Type II error

Answer 8

• Type II
○ Opposite of type I
○ Alternative rejected, null accepted
○ SHOULD be other way around
○ “Pessimistic error/false negative”
○ Researchers say no significant difference when there is
○ More likely when significance level = 0.01/1%
only used if psychologist doing clinical trials as it could affect human life

Question 9

Why do psychologists use 0.5% significance level

Answer 9

Balances out chances of errors