Data Collection Flashcards
What’s a population
Entire collection of things you’re investigating.
Size of population
N
Mieuw symbol
Mean of population
Looks like 6 squared symbol.
Variance of population
Census
Survey of all members of population
Advantages of census
Accurate because all members considered
Unbiased
No worries about sampling frames or representative samples
Disadvantages of census
Long costly
Cant test to destruction
Difficult to make sure all pop surveyed and done accurately?
Difficult to process large data sets may not be relevant.
What’s a sample
Selection subset of population
Size of sample
n
Sample mean
X with bar on top
Variance of sample denoted
S squared
Sample unit
Single item of population
Sampling frame
List database of all members of population
Sample survey
Investigation collects data From every member of sample.
Well conducted survey can provide very good estimates.
Advantages of sampling
Large well mixed sample can represent pop
Cheaper and quicker than census
Can test products to destruction
Data readily available can be more detailed and better quality
Disadvantages of sampling
Can be biased because of poorly designed sample methods.
Can be inaccurate because of random variations.
Sample size may not give information about sub groups of pop.
There may be no sampling frame.
Sampling error
Error due to a sample survey there are two of those
Sampling error
Random error
Error caused by random variations. These errors can be positive or neg, on average size of random errors is zero.
Sampling error
Bias or systematic error
Error caused by non random method of choosing sample. These errors can be positive or neg, on average this error is non zero (biased)
Random sampling
Each item chosen by pure chance.
Free from bias, sampling frame required, difficult more expensive than non random samples.
Three types of random samples
Simple random sampling
Stratified sampling
Systematic sampling
Simple random sampling
Sample of size n that has equal chance of being selected as any other sample of size n.
Stratified sampling
Population of size N divided into mutually exclusive subgroups called strata.
Strata can be decided by one or more criteria.
Random sample taken from each stratum. Total sample size is n.
Sample size for each stratum =
Population of each stratum x n/N
Stratified sampling advantage and disadvantage
Good with diverse populations
Can be difficult to identify meaningful strata
Systematic sampling
Elements chosen at regular intervals from an ordered list.
First term is chosen from 1 to k. To introduce some randomness.
k=N/n. And every kth element thereafter is chosen from ordered list.
Only random is ordered list is truly random.
Different methods of picking random samples
Random number sampling
Lottery sampling
With replacement
Without replacement
Random number sampling
Every member of sampling frame is allocated a number. Random numbers are created (table generator) elements chosen if their numbers come up.
Lottery sampling
Every member of sampling frame is allocated a ticket or ball which is identified by number or name. Tickets drawn one at a time. Difficult to do with large numbers.
With replacement
A chosen item is replaced, so can appear more than once in a sample.
Without replacement
A chosen item is not replaced, so can appear only once in a sample.
Non random samples
Quota sampling
Opportunity sampling
Advantages and disadvantages of non random samples
Cheaper, quicker, convenient as dont require sampling frame.
Introduce bias.
Quota sampling
Alternative to stratified sampling.
Strata sampled in non random way.
An interview has a quota (a set number) of people to interview from different groups (strata).
The interviewer selects individuals and once the quota reached no more interviews in that group.
Convenient and cheap but biased. Commonly used in street/market research.
Opportunity sampling
Uses people from target population available at the time and willing to take part.
Selecting a sample of students from those leaving a library.
Convenient but biased.
