research planning and design 5

Question 1

issues with small samples

Answer 1

• easily distorted by extreme values
• less certainty about the location of the true mean
• larger confidence intervals can mask real differences between groups/ conditions

Question 2

issues with large samples

Answer 2

• sensitive to extreme outliers
• detecting insignificant differences which are too small to be meaningful
• sampling bias may still persist
• can be unethical
• waste of resources
• burden on participants

Question 3

probability sampling methods

Answer 3

= used to make a precise statement about a specific population
- - simple random sampling
- stratified random sampling
- cluster sampling
- systematic sampling

Question 4

non- probability sampling methods

Answer 4

= not as sophisticated, more common
- convienience sampling
- purposive sampling
- snowball sampling
- quota sampling

Question 5

simple random sampling

Answer 5

= every member of the population has an equal chance of being selected

• representative sample, cheap and easy to implement, unbiased.
• needs information about entire population -> might not be possible

Question 6

stratified random sampling

Answer 6

= divide population into smaller sub-populations based on specific characteristics( age, gender) -> then random sampling between thes sub groups.

• diversity of sample is reflected, includes sub groups that can usually be underrepresented with small sample sizes
• requires detailed knowledge of population, unknown characteristics?,

Question 7

proportional vs disproportional

Answer 7

proportional = The size of the sample from each stratum is proportional to the size of the stratum in the overall population

disproportional = The size of the sample from each stratum does not match the proportions of the strata in the population

Question 8

cluster sampling

Answer 8

= divides a population into smaller groups or clusters -> cluster is selected randomly.

• time and cost efficient, large populations, high external validity
• too few clusters may be selected, difficulty defining clusters, might be biased if clusters are nor representative

Question 9

systematic sampling

Answer 9

= select members of the population at regular intervals eg every third person

• easy, time-efficient, using fixed interval reduces bias, good when you dont know details about full population
• not recommended for periodically ordered groups (employee shift patterns)

Question 10

convienience sampling

Answer 10

= take them where you find them, people who are easy to contact/ in geographical proximity/ volunteers

• participants readily available, quick and easy
• high risk of bias, self selection has issues with external validity

Question 11

purposive sampling

Answer 11

= obtaining a sample with pre-determined criteria

• theoretical generalisability
• prone to research bias/ no statisitical inference

Question 12

snowball sampling

Answer 12

= start with one or more initial participants, then continue to recruit on the basis of referrals from those initial participants -> until desired sample is reached

• addresses research Q, efficient for hard to find populations
• research bias, time intensive recruitment

Question 13

quota sampling

Answer 13

• sample represents the estimated numerical compositiob population but the sample is not randomly selected
• similar to stratified sampling

benefits
quick and easy
allows for comparisons between subgroups
can accurately represent entire population

costs
selection bias
onl focuses on specific charecteristics
can not always divide population into mutally exclusive groups