SAMPLING
- Created by: Shumaira Ahmed
- Created on: 18-05-22 11:21
View mindmap
- Sampling
- A good sample should...
- Represent the target population (the set of people that researchers intend to investigate and apply findings to)
- Be unbiased - sample may not accurately represent the target population (eg. some
groups in the population may accidently be excluded.)
- Represent the target population (the set of people that researchers intend to investigate and apply findings to)
- Population Validity: Being able to generalise results from the sample to the target population and be true.
- Types of sample bias:
- Gender Bias: The sample only represents one gender despite the target
population containing multiple genders.
- Androcentric: disproportionate amount of males Gynocentric: disproportionate amount of females
- Cultural Bias: Too focussed on one culture/ ignores other cultures
- Ethnocentric: Research results that are generalised to different cultures despite only being conducted on one culture, thus disregarding how different cultures may have different results.
- Gender Bias: The sample only represents one gender despite the target
population containing multiple genders.
- Sampling Techniques
- Opportunity Sampling: People who are
available + consenting at the time of the research
- Quick and time efficient
- Not representative- chance of researcher bias (eg. only choosing certain people based on looks)
- Self-selected Sample: Volunteered to partcipiate
- Volunteers are more likely to commit to the study and be fully involved.
- Less chance of participant attrition- they chose to participate- so unlikely to change their minds.
- Chance not many people will see advert or be interested enough to respond, leading to small sample, creating invalid results.
- Can gain a large sample quite quickly, that reaches a broad and representative audience
- Inherent bias in participants as similar types of people may
sign up due to interest/need for money/ experience doing studies.
- Could lead to bias + unrepresentativeness.
- Volunteers are more likely to commit to the study and be fully involved.
- Snowball Sampling: Existing participants recruit other participants
- Access people that you may not otherwise be able to recruit
- Bias- people’s friends/family likely to share characteristics- so sample is not representative
- Little effort by researcher
- Random Sampling: Each member of target population has same chance of being selected as any other member
- Time consuming to gather details of every member of target population if it's quite large.
- Not everyone picked may consent to take part- so can take a long time for experiment to actually begin
- Opportunity Sampling: People who are
available + consenting at the time of the research
- A good sample should...
- Stratified Sampling: All types of members of the population are represented proportionally by
selecting different numbers of participants from all strata
- If proportion of target population included 60% people over 50, the sample would too.
- Sample will be representative + have no biases.
- Method is difficult, costly + time inefficient. Target population must also be fully accessible
Comments
No comments have yet been made