L8 - Testing for Differences

Question 1

What is the need for Statistics Tests

Answer 1

Once a difference has been found has it been?

• Is the difference reliable/meaningful
• Could the results be due to chance?

Question 2

What are the steps for choosing a stats test?

Answer 2

1. What is your question?
2. Select your data
3. Know your data
4. Parametric vs Non-Parametric Data

Question 3

What is step 1 when choosing a stats test?

Answer 3

• Questions must be clear
  
  • This must be considered before data collection
• Different tests require different statistical tests
• Tests of relationship or Tests of Difference

Question 4

What are tests of realtionship?

Answer 4

• Correlation (Are variable X and Y realted)
• Multiple regression - which combination of independent variables the most accurate predition of a given dependent variable
  
  • Function + Age = ✅/ Function OR age = ❌
• Chi Squared - Is there an association between two categorial variables
  
  • Is smoking associated with smoking

Question 5

What are tests of difference?

Answer 5

• T-tests - Measure the difference between two groups over some continuous variables
  
  • Whats the difference between sexes in height, sales and job satisfaction
• ANOVA (Analysis of Varience) - Measures differences when there is:
  
  • More than two groups
  • More than two independent variables

Question 6

What is step 2 when choosing a stats test?

Answer 6

• Identify IV, DV
  
  • Independent - What you control, what effects the dependent variable
  • Dependent - What you measure, which is effected by IV
• Identify level of measurement
  
  • Categorical - Nominal, ordinal
  • Continuous - Interval, ratio
• Level of dependent measure = Parametric OR Non-Parametric

Question 7

What is step 3 of chooisng a stats test?

Answer 7

Know your data
- To find characteristics → Descriptive Statstics
- Nominal Data requires that theres a balance in each group
- Continous Data requires normal distrabution (Not skewd)

Question 8

What is step 4 of choosing a stats test?

Answer 8

Question 9

What is parametric tests?

Answer 9

• IN DOUBT USE NONP
• Robust/Tolerant of minor violations
• Makes assumptions on:
  
  • Normal distribution
  • Variances
  • Measurements are in equal intervals or ration level data

Question 10

What are non-parametric tests?

Answer 10

Non-Parametric

• No assumptions about distribution
• Analyses ordinal properties
• Less powerful (> Reject Null Hypothesis)

Question 11

When should you choose P or NONP?

Answer 11

Question 12

What are T tests?

Answer 12

• PARAMETRIC TEST → Check for Normality,Level Of Measurement,Homogeneity,Independence of observations
• Testing significance of an observed difference between two groups or treatments
  
  • Independent-samples t-test → Compare 2 different groups
    
    • Ordinal/Nominal IV and Continous Level DV
  • Paired-Samples t-test → Compare same subject on two different treatment conditions
    
    • Before and after some intervention (Memory Test)

Question 13

How do you perform independent T tests in SPSS?

Answer 13

Question 14

How do you interpret T Test data?

Answer 14

Question 14

How do you perform a paired T test in SPSS?

Answer 14

Question 15

How do you interpret paired T test data?

Answer 15