180304P - SAMPLING: ERRORS, BIASES, AND ETHICS

Presentation by Professor Omar Hasan Kasule

SAMPLING ERRORS

• Total survey error = sampling error + non-sampling errors 
• Sampling errors are random errors. Easy to estimate. They decrease with the increase in the sample size.

NON SAMPLING ERRORS

• Non-sampling errors are 3 types: measurement error, non-response error, and coverage error. Non-sampling errors are difficult to estimate or eliminate 
• Non-sampling errors may be systematic like non-coverage of the whole sample
• Non-sampling errors may be non-systematic. Non-systematic errors cause severe bias. 

SAMPLING BIAS

• Sampling bias, positive or negative, arises when results from the sample are consistently wrong (biased) away from the true population parameter. 
• The sources of bias are incomplete or inappropriate sampling frame, use of a wrong sampling unit, non-response bias, measurement bias, coverage bias, and sampling bias.

SAMPLING ETHICS

• Misleading/non-representative sample
• Misrepresentation of unreliable non-scientific samples eg convenience, quota, etc
• Wrong conclusions from sample data
• Sample size: too big or too small OR ‘fixed’
• Deliberate sampling; exclusion/inclusion of some groups
• External validity vs internal validity