Sampling and population surveys

Question 1

Probability vs non-probability sampling

Answer 1

Probability: involves random selection
+ random sampling to make inferences from the statistics,
- quite difficult in reality

Non: more realistic but not random.
+Usually cost-effective and time efficient
- High risk of sampling bias

Question 2

Random Sample
- S_mple ra_dom sa_ple
- S_stemat_c
- Mul_i-st_ge

Answer 2

1. Simple random sample: Obtain a list of names (marked w/ a number) then randomly choose/sample
2. Systematic sample: Fixed interval, e.g. every 5th person
3. Multi-stage - Different stages
   => first sample unis
   => second sample students at random within those unis

Question 3

Stratified random sample
- St_uctured
- d_spr_portionate
- Cl_ster

Answer 3

Stratified sample structured, proportion of participants match the population
1. divide the pop. into strata (smaller subgroups) based on characteristics => uni level
2. take a random sample that is proportional => 50 first years, 30 second years, etc…
3. pool the subsets to form a random sample

Disproportionate: oversample some groups, usually the minority, to ensure the sample size is sufficient and balance the ss of different groups

Cluster sample: use geographical areas (eg district neighborhoods), randomly select entire clusters
1. define the population
2. cluster them
3. randomly select
4. collect data
Note: more risk of error, could have substantial differences between clusters bc ppl within one are likely to be more similar to one another, therefore less representative

Question 4

Non probability sampling
- Qu_ta sampling
- C_nven_ence
- Sn_wb_l_
- Pur_osi__
- The_re_c__

Answer 4

1. Quota: participants fir one of the specified categories
   - usually marketing researching
2. Convenience: due to easily recruited
   - usually quantitative research in psychology
3. ## Snowball: asked to suggest other SIMILAR PEOPLE to participate
4. Purposive: select participants as they are of theoretical interest
5. Theoretical: select further participants to “test” as ideas develop
   - occurs after some data has been collected, analysis formulated so that further recruits to the study may inform/challenge the developing theory

Question 5

Socio-demographic info

Answer 5

• abt the nature of the sample studied
• characteristics described depend on the kind of participants and purpose of the study

Question 6

Considerations when carrying out research (sample size)

Answer 6

• population size
• number of people you can contact, proportion that will participate
• variability of their responses
• how confidence (sig level) for the results
• how accurate for the estimate to be compared with actual pop.

Question 7

How to determine whether our sample is representative of the population?

Answer 7

1. check sample against other available data abt the pop.
2. compare it with census/national data on characteristics eg marital status

Question 8

Sampling error

Answer 8

• occurs due to a sample being a subset of the population
• may not precisely represent it: difference between mean of sample and mean of population

Question 9

How to reduce sampling error?

Answer 9

• use proper sampling methods
• make the sample size larger so there is more variability
• know your population: ensure only target the sample that matters

Question 10

Important characteristics for sample and population
- inclu_ion cri_eria
- repre_entativ_ness
- state of constant fl_x + ti_e and c_st

Answer 10

inclusion criteria
- specific what population to focus on
- all possible cases that meet the criteria (not possible cuz pop. always changing)

representativeness
- sample mirrors the pop. in important respects for generalizability

State of constant flux + time and cost
- turn a sample, not pop. (portion of population)

Question 11

What is sampling frame?

Answer 11

• a tool
• set of info abt the acessible units in a sample
• a list of ALL possible cases within a pop. that can be sampled
  (e.g telephone directory, school lists, employment records, etc…)

Question 12

What is the purpose of sampling?

Answer 12

• to approximate the characteristics that are relevant to the RQ abt a larger population

Note: sample must be representative so that researchers can make “accurate” inferences abt the larger population

Question 13

Descriptive vs Inferential statistics

Answer 13

• summarize or describe (data distributions from samples and populations)
• statistic = characteristic of sample
• parament = of the pop.
  => statistic is used to infer the parameter(s)