L15 - Statistical Testing - The Sign Test (When To Use & Calculations)

Question 1

Two types of statistics

Answer 1

• descriptive statistics
• inferential statistics

Question 2

Descriptive statistics

Answer 2

• measures central tendency & dispersion

Question 3

Inferential statistics

Answer 3

• Use statistics to infer something about target population
• Use inferential statistics to see if results are significant
  E.g. drinking speed up increases memory
• can’t test world population, test sample & infer about population

Question 4

Level of significance definition

Answer 4

• Level at which the decision is made to reject null hypothesis in favour of experimental hypothesis
• states how sure we can be that IV is affecting DV & it’s not due to chance

Question 5

Chance definition

Answer 5

Has no real cause, just happens

Question 6

Common significance level used

Answer 6

5%

Question 7

Strictest signifance level

Answer 7

1%

Question 8

What are significance levels

Answer 8

• Look at whether a real difference exists between control and experimental condition & how certain we are a real difference exists
• if 2 sets of data are very similar, test may indicate there is no real difference - accept null hypothesis
• if probability there is real difference between two conditions - accept experimental hypothesis

Question 9

Why is p<0.05 often used

Answer 9

• not too strict or lenient, middle fair value of significance
• minimises chance of making type 1/2 error

Question 10

What does p<0.01 mean

Answer 10

• 99% chance that the results are showing significant difference, 1% chance it’s due to chance
• often used for testing drugs on humans

Question 11

Reasons to uses the sign test

Answer 11

• we are looking for a difference in data
• looking at paired/related data - could come from repeated measures/matched pairs design as participants are paired for purpose of statistics , one person tested twice
• data is nominal (categories)

Question 12

How to do sign test

Answer 12

• state hypothesis
• calculate signs
• find S value - smaller value of signs
• find N value - number of participants (ignoring any 0’s)
• find critical value & state if significant

Question 13

When are results significant

Answer 13

• S value is LESS THEN OR EQUAL TO critical value for significance

Question 14

How do we write out results of experiment?

Answer 14

• results are/aren’t significant as…..
• calculated value for S is….. which is higher/lower/equal to critical value of…..
• when p<0.05 (one/two tailed test) when N=
• accept/reject hypothesis

Question 15

Which hypothesis do you accept/reject

Answer 15

• accept null hypothesis if results are NOT significant
• accept experimental hypothesis if results ARE significant