Topic 1 (statistics) Flashcards
Population
Complete set of items you are interested in
-individual items of population are known as sampling units
Census
Measures a value from every member of the population
Adv
-get completely accurate view of population
Dis
- time consuming and expensive
- cant be used when testing process destroys items
- not possible with a changing population (continually)
Sample
Selection observations from subset of population
info on whole population
Size = affect validity of conclusions (larger=accurate)
Different samples = different conclusions = natural variation in population
Adv
- quicker and cheaper than census
- fewer People respond= ( larger population)
Dis
- not be representative of original population
- not be large enough to give info about small minority sub-groups of population
Sampling frame
When sampling units of a population are individually named or numbered to form a list called a sampling frame
Parameter
Number that describes the entire population
Statistic
Number taken from a single sample
Can use one or more of these to estimate the parameter
Random sampling
Every member of population = equal chance of selection
Sample representative of population & removes bias from sample
3 methods of random sampling
- Simple random sampling
- Systematic sampling
- Stratified sampling
Simple random sampling
Every sample of size n has equal chance of being selected
Need sampling frame, list — each thing allocated unique number and these numbers chosen at random
2 methods of choosing number
- Lottery (placed in a hat)
- Generating random nos (calculator/computer/random number table)
Adv
Fair way, probably representative of population, each sampling unit = same chance of being chosen
Dis
Not possible without sampling frame, time consuming/disruptive or expensive when large population, minority groups may be missed
Systematic sampling
Chose starting point at random systematically select objectives a certain number apart
Only random when sampling frame has no order
Adv
Quick and easy suitable for large samples and large populations
Dis
Not possible without sampling frame
If sampling technique conincides with a periodic trait in population, sampling technique no longer representative = introduce bias
May be missing values in population and minority groups might be missed
Stratified sampling
Population split into distinguishable groups different to each other and cover whole population groups are called strata within each group sample is selected
Adv
Minimises sample selection bias ensuring certain segments aren’t over or under represented
Frequencies for each sampled group proportional to frequencies for each group in population
Minor groups included and sample reflects whole population
Dis
Need sampling frame, strata must be carefully defined
Sometimes difficult to split population into naturally occurring groups
How to decide on a sampling method
list every member of population
sources of bias/difficulties = taking certain samples
methods available = best suiting
Sampling method biased = sample that doesn’t represent population
Types of non random sampling
- Opportunity sampling
- Quota sampling
(Volunteer sample: often biased only those passionate about that area being investigated may contribute their data/thoughts)
Opportunity sampling
Taking sample from, target population available at time study is being carried out and fit criteria
Adv
East to select sample, inexpensive
Dis
Unlikely to produce a sample representative of population
Highly dependant on individual researcher ‘pick nice people’ = biased
Quota sampling
Population split into groups like stratified
Size of each group determines proportion of sample having characteristic
Judgement is used to select members and the quotas are filled by interviewers
Adv
Even small sample representative of population, no sampling frame needed, quick easy inexpensive and different group responses can be compared
Dis
Non random could be biased, split into groups = inaccurate or time consuming
Non responses aren’t recorded may distort interpretations
Different to stratified = as could be biased when choosing the people once in groups not everyone gets a chance of selection stratified is random
Quantitative data
Associated with numerical observations
Can be continuous (can take any value in a given range you measure it ) or discrete (only takes specific value in a given range often count in)
Qualitative data
Associated with non-numberical observations
Large data set places
from west to east then south to north Camborne = coastal Hurn = coastal Heathrow = inland Leeming = inland Leuchars = coastal
Jacksonville (NH)= coastal
Beijing (NH)=inland
Perth (SH)=coastal
things to be aware of for each large data location
- coastal or inland
- hilly or flat
- distance fish from equator
(need basic geographical knowledge e.g. coastal = windy, mountains = rain and changeable)
recording data
-reading not avaliable = n/a
-total amount of rain fall less than 0.05 = trace of rain
-relative hummdities above 95% = associated with mist and fog
-1 knot is 1.15 mph
-cloud cover = okras 0 = clear sky and 8 = overcast
-pressure = hectopascal
(sheet in folder)
statistic
random variable that is a function of sample which contains no unknown quantities/parameters
(solely/only known)
