Summary of issues involved

 

Sampling

  • population and sample
  • sample vs. census
  • representative sample? Watch:
    1. voluntary response
    2. convenience sampling
    3. nonresponse
    4. wrong answers
    5. ‘silly’ answers
    6. subjective aspects
    7. etc, etc,etc,!!!

 

  • sampling techniques
    1. SRS (simple random sample)
    2. stratified sample
    3. cluster sample

 

DOE (Design of Experiments)

  • sampling vs. DOE
  • causality
  • confounding? Avoid by

1.      control group

2.      random assignment

  • still problems! E.g.

1.      representative?

2.      realistic?

3.      hidden bias

  • techniques:

1.      CRD (Completely Randomized Design)

2.      RCBD (Randomized Complete Block Design)

 

Measuring

  • validity and reliability
  • nominal, ordinal, interval/ratio
  • check for
    1. internal consistency
    2. exaggerated precision
    3. right context
  • location: mean, median, mode
  • scale: sample variance, IQR (inter quartile range)
  • correlation: sample (rank) correlation coefficient ; interpretation: use square!

 

 

Literature:

Moore, D.S. (1985).Statistics: concepts and controversies. Freeman, New York (CBa 517.3 m078)