Non-Probability Sampling and Evaluating Samples

Used because:

Commonly used in psychology research because most research is basic and concerns basic relationships between valuables and psychological processes
e.g., Memory, vision, etc.
Haphazard Sampling (Convenience) – “Take them where you can find them”
- Sampling bias: may not be representative of population
  e.g., Library – more studious
Purposive Sampling – sample with purpose in mind, looking for predetermined criterion
e.g., Females in 20s
- To attain sample of people with predetermined/predefined criterion
Quota Sampling – Goal to get sample of individuals to represent sub-groups of population (like stratified sampling)
- Haphazard selection to reflect composition
  e.g., 30% of participants ESL but try to get percentages
Snowball Sampling – Subjects lead to more subjects
We need to be aware of sampling bias

Evaluating Samples

To get completely unbiased sample, need
1) complete and random sample frame with
2) high response rate
Can be very, very difficult even with methods since people need to agree to participate
Sampling Bias occurs when sample is not completely representative of population
- People in population missing
- No longer members of population
- No sampling frame
  e.g. Homeless people in Vancouver; Pregnant women in Vancouver
- When large percentages of participants choose not to participate
Response Rate – Percentage of people who complete survey; indication of sampling bias
- Non-respondents may differ in gender, age, income, education, personal importance of issue under investigation
- Higher is better: 60% marginal; 70% reasonable; 80% good; 90% great!
- Would limit generalisability of findings

How to Increase Response Rate?