Primary data
Sampling
a) use non-probability sampling methods including convenience and quota sampling
b) use probability sampling methods including simple random sampling and stratified random sampling
Sampling is the process of selecting a number of people to represent the whole population. It is an essential part of geographical investigation, as it allows the collection of data from a large number of people without having to interview or survey every single person in the population.
You will need to look at:
Where - to position the group members on the site to collect the sample e.g. at sites with more footfall so as to get sufficient responses.
How - to determine which sampling method to use so that the sample are representative and reliable.
Who - to determine the size of the sample e.g. to collect 20 responses at each site.
When - to determine duration and days to collect the sample.
Probability sampling
• Simple random sampling or stratified random sampling can be used. Stratified according to a known subgroup e.g. by genders
• Samples are randomly selected, without the researcher’s conscious decision, using a random number generator or die.
• This removes bias that may come from the choices made by the researcher.
• Has a greater chance of creating a typical representation of the population so it is used when you need to select a more representative sample
Example:
• To find out your schoolmates’ favourite place in the neighbourhood, you select students from your school using a random number generator, instead of selecting only schoolmates that you know.
Simple random sampling is normally used when there is little known about the population under study. Every member of the population is given a number. A random number generator is used to generate random numbers to select the samples
Another example of random sampling can be seen below on sampling air temperatures at different locations in the school. The investigator will first use a random number generator is used to pick the x- and y-coordinates from a map of the school, which will locate the various points for measurement on a map with grids, as seen below.
Stratified random sampling is used when the population clearly includes significantly different sub-groups. The researcher may then wish to select a sample that has a proportionate makeup to the population based on age and sex (or other categories such as ethnicities).
For example, a researcher may hypothesise that a person's sex is an important factor in the choice of activities. If 60% of the population is made up of females and 40% males, then the sample must also follow the same ratio. Hence, the researcher will have to randomly select 6 females and 4 males if he/she wants a total sample size of 10. The selection of cases by sex in each sub-group should be carried out randomly using a random number generator. Otherwise the sample becomes a quota one.
Advantages of stratified random sampling
Consider the diversity in the population so that it is better represented ie. selecting people of different genders and age groups
Free from researcher bias as every subject in the population has an equal chance of being selected.
Non-probability sampling
Samples are non-randomly selected, often using the researcher’s conscious decision such as convenience sampling and quota sampling.
This means that the researcher subjectively selects samples, such as family or close friends.
This selection may be biased.
Samples are unlikely to be representative as they are subjectively selected, making it hard to make generalisations about the population.
It is used when it is unnecessary or impractical to select a representative sample such as when:·
conducting interviews,
testing out the design of the questionnaire
conducting an exploratory research
Example:
To find out your schoolmates’ favourite place in the neighbourhood, you only select your friends in your class or CCA.
Advantage of non-probability sampling:
When sampling needs to be conducted quickly due to time limitation
Convenience sampling is carried out in an ad hoc manner that is most accessible to the researcher. For example, students working on a school project on local transportation systems may decide to give out a questionnaire survey to the first 100 people outside the train station on a Saturday morning.
Quota sampling takes into consideration factors, such as gender. For example, a researcher who is interested in investigating the preferences of various green spaces in the Punggol neighbourhood may select samples based on the gender, i.e. male or female, using convenience sampling. However, the samples selected may still be subjective. For example, the researcher may select only respondents from the same age as them. This will reduce the representativeness of the data collected.
2. Closed-ended questionnaire surveys
a) create pre-defined responses to questions that are limited to short phrases, single words or numbers
b) use rating scales to guide responses including the Likert Scale, frequency scale and ranking scale
Predefined responses
The predefined responses in the questionnaire survey could be short phrases, or a single word arranged in a series, or numbers.
Rating scale
Likert scale is a type of rating scale that presents a range of responses anchored by two extreme opposing positions.
Frequency scale is a type of rating scale which shows a range of responses based on the number of occurrences.
Ranking scale is a type of rating scale where items are compared with one another and ordered in preference by survey participants.
3. Mental maps
a) visualise experiences by drawing features and adding labels onto the base map of a study area
b) conduct semi-structured interviews with open-ended questions exploring features and labels added to the map
Mental maps can be used to capture these personal insights. Mental mapping provides a lens into:
how people produce and experience space,
how people think visually and spatially about their environment, and
the dynamic interrelationships people have with their environment.
Look at the mental maps of Singapore drawn by two respondents on how they feel global warming will impact on Singapore.
Can you spot the differences between the two mental maps?
• Map B shows more impacts of global warming. It shows the impact of sea level rising resulting in flooding of the low-lying coastal areas, more extreme weather in the form of more frequent and intense rain resulting in the urbanised areas such as Orchard Road as well as increased spread of insect-borne disease such as dengue.
• Map A shows inaccurate information of global warming on Singapore. As Singapore water is relatively sheltered with narrow straits between the island and neighbouring countries, there is very low chance of storm surge affecting the west coast as shown in Map A. This is especially so as storm surge are formed by tropical cyclones which Singapore is unlikely to experience due to its location near the equator which has weak Coriolis effect.
• Map B has more places in Singapore labelled more accurately. For example, Sentosa was drawn at which Jurong islands are in Map A.
• The outline map of Singapore of Map B is more accurate than Map A.
Why do you think the mental maps are different?
• The person drawing Map B have a better knowledge of the impacts of global warming on Singapore.
• The person drawing Map B is very familiar with the map of Singapore and could remember it well.
A mental map of my neighbourhood