Data Collection Flashcards
Random
each member has equal chance of selection
therefore sample should be representative of population
not bias
simple random sample
each sample has equal chance of selection
need sampling frame
each member allocated number then number chosen at random
e. g. random number generator
e. g. lottery sampling - names drawn out of hat
population
complete set of items of interest
Census
measures every member of population
Census pro’s
everyone represented
accurate
Census con’s
costly
time consuming
lots of data
can’t use if destroying objects
sample
subset of population measured used to make generalisations about population
sample pro’s
quicker
cheaper
less data
used if destroyed
sample con’s
you need a representation thats fair
bigger better than smaller sample
not greatly accurate
sampling units
individual members of a population
sampling frame
list of sampling units
Systematic sampling
elements chosen at regular intervals from list
e.g. size 20 from 100 (every 5th interval)
first person chosen at random
stratified sampling
population divided into mutually exclusive strata (e.g. males and females)
random sampling taken from each
No. in sampled in stratum = (no. in stratum/no. in population) x 100
simple random sampling pro’s
no bias
easy and cheap
each sampling unit has easy and known chance of selection
simple random sampling con’s
not suitable when sample/population is large
sampling frame needed
systematic sampling pro’s
simple and quick
suitable for large samples and populations
systematic sampling con’s
sampling frame needed
can be bias if sampling frame isn’t random
stratified sampling pro’s
accurately reflects population structure
proportional representation of groups within population
stratified sampling con’s
population must be clearly divided into distinct strata
selection within stratum suffered similar disadvantages as simple random sampling
quota sampling (non-random)
interviewer/researcher selects sample that reflects characteristics of whole population
quota sampling steps
population divided into groups according to characteristic
size of group determines proportion of sample that should have that characteristic
interviewer assesses then allocates people into appropriate quota
if person refuses or group is full they are ignored.
Opportunity/convenience sampling
take sample from those available and fit the criteria pop what you’re looking for
e.g. !st 20 people outside a radio shop with records
quota sampling pro’s
allows small sample to reflect population
no sampling frame
quick, easy and cheap
allows comparison between groups in a population
quota sampling con’s
non-random bit might introduce bias
costly and maybe inaccurate dividing people into groups
increasing size, increases groups which adds to time taken and expense
non-responses aren’t recorded
opportunity sampling pro’s
easy to do
cheap
opportunity sampling con’s
unlikely to provide a representative sample
dependent on individual researched
quantitive variables/data
deal with numbers
e.g. shoe size
qualitative variables/data
deal non-numerically
e.g. hair colour
continuous variable
can take any value in given range
e.g. time : 2, 2.02, 2.001 secs etc.
discrete variable
only takes specific values in given range
e.g. girls in my bedroom ( not 2.5 but 10)
