WEEK 4 - L8.2 sampling

Question 1

elements/ sampling units types

Answer 1

1. finite vs. infinite
2. known vs. unknown

Question 2

sampling frame

Answer 2

a list of all eligible members of the target population from which the sample will be selected

Question 3

sample unit

Answer 3

those who are selected for the sample

Question 4

sampling biases

Answer 4

1. research bias
2. response bias (participant artifact)
   - self selection bias
   - non response

Question 5

self selection bias

Answer 5

those who are interested join

Question 6

random selection

Answer 6

every unit in the pop has an equal probabilty of being chosen

Question 7

probability sampling ex.

Answer 7

simple random sampling
- sampling w or w out replacement
- w conforms to srs standards ( if you dont replace they have a higher chance of being chosen)

Question 8

systemic random sample

Answer 8

picking the 10th person every time for example
still same probabilty of being selected
for example if you say always 5th house, but the houses in the right are always more expensive this leads to bias

Question 9

stratified random sampling

Answer 9

SRS within known subgorups
(specific ethnicities)
unless, there’s correction with reweignthing, it may be disproportionate

Question 10

multistage cluster sampling

Answer 10

first divides the population into equivalent and internally heterogenous groups, and than starts sampling at the group level before sampling the final units of interest
- sampling in stages (first larger than smaller units)
- you only need info for the communities you select first

Question 11

3 types of probability sampling

Answer 11

1. simple random sampling + systematic random sample
2. stratified random sampling
3. multistage cluster sampling

Question 12

non probability sampling 4

Answer 12

1. convenience smapling: easily available ppl (students)
2. purposive: researcher decides
   e.g quota sampling (half of sample male/female)
3. snowball (chain/ referral/ network) sampling: known participants are asked to identify others
4. theoretical sampling: during data collection and analysis (used in grounded theory)

low external validity

Question 13

response rate types (4)

Answer 13

1. contact rate (% of selected individuals contacted)
2. cooperation rate (% of individuals participating)
3. surveyed raate (% of respondents surveyed too often)
4. response rate (completed interviews/ eligible sample)

Question 14

types of weighting 2

Answer 14

1. based on information available priori (correction for over and under sampling of known subgroups)
2. post ad hoc corrections: reweighting of realized sample (e.g due to nonresponse) to correspond to known proportions- but no fix for sampling bias

Question 15

when should weighting be used?

Answer 15

for inferences on the whole population - it’s highly reccomended
for testing patterns that support causal relationships - it’s optional

Question 16

sample error (…) with bigger sample size

Answer 16

decreases
ters oranti

Question 17

larger samples (…) sampling error and (…) statistical power

Answer 17

decreases sampling error
increases statistical power