Students

IMPORTANT GUIDELINES FOR DATA COLLECTION

Data can be collected in four broad ways:

  1. Documentary sources
  2. Observations
  3. Surveys
  4. Tests, measurements and experiments.

DOCUMENTARY SOURCES

  • Documents can be used to set an issue in a historical context or as the basis for an entire study.
  • A wide variety of documents can be used, e.g. the Census of Population, available from the Central Statistics Office.
  • Personal documents, used judiciously, can be useful in providing information.
  • Try to ensure that whatever documents you are using are the most current available.
  • Photographs and maps may also be used. These are available from the Ordnance Survey Office, Phoenix Park, Dublin. LoCall: 1890 674 627 or Email: custserv@osi.ie

OBSERVATIONS

  • This is one of the primary methods of collecting data, but care must always be taken to ensure that these data are observed in an unbiased way.
  • The observer's senses may not be able to record everything. Also, if the observers are watching people, animals or other organisms whose behaviour changes because they are being observed, the results may be invalid.

SURVEYS

If you are doing a survey read the following carefully:

  • Questionnaires, interviews and schedules are some of the techniques used in conducting survey work.
  • Questionnaire design merits great attention.
  • It is very important to think through how you are going to analyse the results you will get.
  • Your questions should be clear, concise and should gather the relevant information
  • Test your questionnaire in advance on a small section of the population - this is called a pilot survey. This will identify the questions that need changing and it will lead to a more effective questionnaire.
  • Good interviewers do not influence the answers given during an interview. Work from prepared questions.
  • If you are recording any type of behaviour by animals, plants or humans, it is advisable to use a schedule to record your observations.

TESTS, MEASUREMENTS & EXPERIMENTS

  • These should only be used if they are relevant to your research and if you are capable of doing and understanding them yourself. Particular attention should be given to the design of experiments, the requirement for controls, sufficient replication and repeat experiments where appropriate.
  • Ensure that any testing or experimentation you undertake is not dangerous i.e. it does not put yourself or others at risk of injury or disease.

GUIDELINES ON SAMPLING

  • Remember to use a representative sample.
  • A random sample means that every member of a population had an equal chance of being chosen, e.g., pulling numbers from a hat.
  • A systematic sample takes every nth member from a population.
  • Stratified sampling uses the idea of groups or classes within the population being analysed.
  • Any group which shares similar characteristics and has boundaries may be termed a population.

Therefore, it is perfectly acceptable to refer to plant populations.

  • Quota sampling means that if you want to interview, for example, 200 people about shopping, you could go to a particular part of town where you could meet shoppers. You may have pre-set guidelines, such as age group and numbers of men and women. However, it is not statistically random and people on the street may not be fully attentive.
  • When sampling a population, you may also need to use a control group. If, for example, you were testing the effects of a particular experience on a group of people, you would need a control group of the very same type of people, who have everything in common except the particular experience.
  • Case studies, which look at a small number of individuals and a particular context in depth, may be useful in helping us understand how a particular process works. They will help lead towards a better way to formulate a hypothesis for testing with a large sample.

GUIDELINES ON STATISTICS

What techniques can you use to analyse data?

There are three main procedures you might use:

  1. You could summarise your data.
  2. You could try to explain patterns which emerge, using comparison techniques.
  3. You could carry out a significance test, e.g. a t-test.

SUMMARISING DATA

  • This procedure means what it says. It is a way of reducing the bulk of data to a more manageable size, as well as seeing some patterns emerging. You can put data into groups or classes. You can also measure typical values, such as the mean, mode and median.
  • Some data, of course, will not be accurately described by these statistics. What is then needed is a technique to measure movement away from the average.

This technique measures deviation from the mean.

COMPARING DATA

We can compare data in the following ways. Firstly, we could compare the similarities and differences between the data. Secondly, we could use statistical techniques to compare the data. These techniques are widely used to compare variables.

SIGNIFICANCE TESTS

When you have made your comparisons and conclusions, you need to know if they are really significant. Significance tests are used to make sure that results from comparing one data set with another are not the result of chance.

« Back to Student home page