Stage 2 Flashcards
Define sampling
A subset of items taken from the whole population (all the data available).
When is sampling used?
It is used when:
•The population is too large to study each individual
•Time is limited
•Collecting data is destructive and you are trying to minimise damage
What are the 4 features of a good sample?
- Unbiased
- Precise
- Appropriate size
- Easily collected
Features of a good sample:
unbiased
A sample where the mean and standard deviation of the sample is neither larger or smaller than the values for the whole population
Features of a good sample:
precise
A sample that provides an accurate estimate of the population characteristics. It does not distort patterns/trends and does not exaggerate anomalies
Features of a good sample:
Appropriate size
A sample that is large enough to produce conclusive results in terms of statistical significance (a sample size of over 12 is considered large enough for making reliable conclusions using statistical testing e.g. Spearman’s Rank), but small enough to be done well.
Features of a good sample:
Easily collected
Can be easily collected within the time/ equipment and researchers available
What should a plan of a geographical investigation include?
- Sampling frame
- Sampling number
- Sampling Design
- Plan of timing/sampling/ division of labour
Components of a plan of a geographical investigation;
Sampling frame
Sampling frame which clearly defines the outline of the area you will be investigating and a justification for choosing this area
Components of a plan of a geographical investigation;
Sampling number
Sampling number which is sufficient to provide a reliable sample but practical for collecting within the time scale and with the number of researchers available
Components of a plan of a geographical investigation;
Sampling design
Sampling design which clearly describes and justifies the choice of sampling technique employed.
- Sampling type could be grid, point or transect
- Sampling strategy could be random, systematic, stratified or pragmatic
- It is advisable to draw a sketch of the area to show the locations of the sampling sites
Define: sampling frame
The study area or group from which the sample is selected (e.g. area on an OS map, electoral register, telephone directory, every person entering a shopping centre)
Define: target population
Total set of measurements relevant to a study (e.g. all possible measurements of pebble size from a beach which are useful for the investigation)
Define: sampled population
Measurements for the area which is accessible to the investigation (e.g. pebbles which are exposed on the surface of the beach)
Define: sample size
The number of measurements in a sample. For further statistical analysis this is usually a minimum of 30
Define: population parameter
A true summary measurement of the characteristics of the target population (e.g. mean size of pebbles on a beach). If the parameter is based on a sample, it is only an estimate
Define: bias
This occurs when the sample measurement over- or under-estimates the population parameter. May be introduced unconsciously (e.g. including certain places because they are more accessible or interviewing a certain age group/gender of shoppers because of the time of day/day of the week)
Define: sampling error
The difference between a sample estimate and the population parameter
Define: representative sample
A sample that minimised bias
What are the three main types of spatial sampling?
- Point sampling
- Line sampling
- Area sampling
spatial sampling - What is point sampling?
a selection of specific points within an area
spatial sampling - What is line sampling?
sampling along a line through (e.g.) a city or across a sand dune
spatial sampling - what is area sampling?
when an area is sub-divided into a selection of sample squares
Sampling Type : Random
What is the method for random sampling?
Use number tables to generate random coordinates
Sampling Type : Random
What are the advantages of random sampling?
- Least biased because every individual has equal chance of being selected
- Relatively quick and easy to select sample types
Sampling Type : Random
What are the disadvantages of random sampling?
- May not take samples from a whole survey area
- Important variations may not be sampled and therefore you may not obtain a truly reliable sample
- Theoretically the same individual could be selected more than once
Sampling Type : Systematic
What is the method for systematic sampling?
Samples are regularly spaced e.g. every 100m along a transect
Sampling Type : Systematic
What are the advantages of systematic sampling?
- Quick and easy method
- Samples taken from all parts of the survey area so the whole area is covered
- No individual is selected more than once
Sampling Type : Systematic
What are the disadvantages of systematic sampling?
- The regular interval may miss an important variation e.g. the top of the dune or a dune slack
- You may sample and evenly spaced characteristic which means the sample is unreliable because it over emphasises the importance of the regular feature
Sampling Type : Stratified
What is the method for stratified sampling?
- sampling frame divided into sub-sets
- number of samples is taken from each subset in proportion to it’s relative importance
(e. g. sand dunes is taken at a change of slope.)
Sampling Type : Stratified
What are the advantages of stratified sampling?
- Can provide greater precision than a random sample
2. It requires greater precision, so often requires a smaller sample than for random sampling. This saves money.
Sampling Type : Stratified
What are the disadvantages of stratified sampling?
- May be difficult to get background data to allow for reliable sampling
- Can introduce bias because researcher selects predetermined factors as being important e.g. land use or age/gender of population
Sampling Type : Pragmatic
What is the method for pragmatic sampling?
Samples are only taken from areas which is safe or accessible
Sampling Type : Pragmatic
What are the advantages of pragmatic sampling?
- Quick, easy and SAFE
Sampling Type : Pragmatic
What are the disadvantages of pragmatic sampling?
- Will not take samples from the whole survey area so important variations may not be sampled
- May not obtain a truly reliable sample
- Bias is introduced
What is a risk assessment?
identification and evaluation of the risks involved in undertaking research and the strategies that can be taken to minimise these risks
What is a physical risk?
risks of harm or injury to the researchers caused by the environment
What is a statistical risk?
A risk that the data collected does not accurately represent the whole population because of bias in the sampling technique or errors in data collection.
When are statistical risks introduced into an investigation?
Statistical risk is introduced into an investigation when:
- Insufficient samples are taken or the sampling technique is not well planned so that not all the population is represented in the same sample
- Data collection methods can cause risk when the equipment is inaccurate or the researchers lack precision
Name 5 methods for reducing statistical risks :
- Sufficient samples need to be taken and the samples must cover as much of the sample frame as possible
- The sample size must be large enough for statistical testing to be used with confidence. A sample size of at least 12 is needed
- A careful balance between the number of samples needed and the time/researchers available must be made
- Highly accurate equipment must be used
- Researchers must be trained so that they can use the equipment accurately
Pilot Survey
What is a pilot survey?
A prior visit to the investigation site or practising your questionnaires with friends/family prior to the real investigation
Pilot Survey
What are the advantages of a pilot survey?
- To assess the accessibility of the site
- To assess the physical risks of the site
- To practise data collection methods to ensure more accuracy
- To make sure that the sampling techniques and data collection methods selected will result in the required data being collected for the investigation
- To help refine data collection sheets
- To identify any unforeseen risks both physical and statistical
Pilot Survey
What are the disadvantages of a pilot survey?
- Time consuming
- Costly
- Might be hard to access location