Sampling Methods Flashcards
What is Sampling Method
Karamoutsou 2024- “The (process of) selecting a number of respondents (sample) to represent a population of interest.
Why use Sample
Karamoutsou 2024:
Cannot access the entire population, cheaper, reduces time
Aim of sampling
Karamoutsou 2024:
Ensures reliability of findings
Find as close to truth as possible to the characteristics of population
what is the sampling process
1) Define the population of interest
2) determine whether to sample or census
3) select sampling frame
4) choose sampling method
5) determine sample size
6) Implement the sampling procedure
Step 1: Define the population of interest
Wilson 2019- population of interest is the total group of people that the researcher wishes to examine, study or obtain information from. Population of interest normally reflect the target market or potential target market for product or service such as gender, age, hobbies and more.
Utilise screening questions to ensure these are meet
Step 2: determine whether to sample or census
Karamoutsou 2024- defines them as
* Census: attempting to study every member of your defined target population
* Sample: selecting a subgroup of that population for study
Resources and accessibility will drive the decision
* Cost
* Time
* Access to entire market
Census is preferred- uncommon though typically only in specialised industries
Step 3: select sampling frame
Wilson 2019: Sampling frame is a list of population of interest from which researcher selects individuals for inclusion such as age, gender, location and more
Sampling frame error wilson 2019
* A bias that occurs as a result of the population implied by the sampling frame being different from the population of interest
○ Reduce by adding a number of lists together to create sampling frame
Use of Access Pannels wilson 2019
* A sample frame which uses a database of individuals who have agreed to be available for surveys of various types and topics
○ Saves time and money
Concerns for sample frame- Karamoutsou 2024
* Is it up-to-date
* Inclusive of entire population?
Step 4: choose sampling method
Karamoutsou 2024- there are 2 main sampling methods
* Probability Sampling:
○ each sampling unit has a known, non-zero chance of being selected.
* Non-probability Sampling:
○ each sampling unit’s chances of selection are unknown
what are the probability sample methods
Systematic random sampling
Stratified random sampling
Cluster sampling
What are the non-probability sample methods
Convenience Sampling
Quota Sampling
What is Simple Random sampling
everyone has the same chance of being selected
- Advantages - Every unit has the same probability of being chosen. Gives a good representation if all subjects participate.
- Disadvantage - Not possible without a complete list of population members. Potentially uneconomical to achieve.
What is convenience sampling
A sample is taken from those who are easy to sample. This could create some very obvious problems!
What is Systematic sampling
Number all units from the population consecutively and systematically select units. For example take a randomly generated number and select all the units which end in that number.
what is cluster sampling
- With cluster sampling, the researcher divides the population into separate groups, called clusters. Then, a simple random sample of clusters is selected from the population. This helps keep costs down.
Clusters are considered to be similar to each other because each cluster should reflect the characteristics of the population.
what is Stratified sampling
The population is split into non overlapping and identifiable groups called strata. Units are selected from each strata at random. Strata are groups that contain similar traits. Strata vary significantly between each other.
○ Advantages - Can ensure that all groups are represented proportionally in the sample by selecting individuals from each strata.
○ Disadvantage - More complex than simple random sampling and so requires greater effort. Strata must be clearly defined.
what is quota sampling
Quota is given for different characteristics of a population, for example age, and individuals are sampled (not randomly) until quotas have been met
Step 5: determine sample size
Wilson 2019
*Hard decision for researcher as increasing sample size reduces sample error but is more expensive and time consuming so depends on the finance and time available to the researcher:
* Budget availability key determinant
* If many different views on topic in population bigger sample size the better
* Number of subgroups of data analysed- increase sample size if more subgroups
3 ways to determine sample size
* Statistical
* Observation to question ratio
* analytical factors
Statistical Karamoutsou 2024
* Requires estimates for
○ standard deviation for
population
○ acceptable level of precision
expressed as sampling error
○ desired confidence level that
the result of the survey will fall
within a certain range of the
true population
Variation
* Variation (“ Var”): The amount
of disparity of a certain characteristic or phenomenon in a population.
- Depends on the VAR in the population for the key variable(s) I’m interested in! The higher the VAR the larger the size of the sample; and vice versa!
Various formulas depending on which sampling method is used and dependent on what kind of data it is after
Observation to question ratio
* The more complex the analysis the higher population size required
○ Min 5 respondents per question
○ 10:1 more acceptable now
Step 6: Implement the sampling procedure
Wilson 2019
Once sampling procedure has been selected researcher can start selecting the members of sample and begin the survey
What is Sample error
Sample error - wilson 2019
* Always exist often in 3 main forms
○ Sampling frame error
- A bias that occurs due to
population implied by sample
frame being different to
population interest
○ Non-response errors
- Errors in a study that arise
when some of the potential
respondents do not respond
○ Data errors
-Non-sampling errors that occur
during data collection or analysis
that impact on the accuracy of
inferences made about the
population of interest
Main types: interviewer and data
analysis errors
balls
balls
