• Primary research is used when existing secondary information does not provide the specific information required.
• The organisation may carry out the research itself or employ a professional research organisation.
• The most common primary research method described is the survey.
• Main survey methods:
• Personal interviews
• Telephone interviews
• Postal questionnaires

8.1 PERSONAL INTERVIEWS

• A sample of potential users is selected and questioned face-to-face.
• Advantages:
• Allows detailed questioning
• Interviewer observations can supplement answers
• Disadvantages:
• Can be expensive
• Can be time-consuming

8.2 TELEPHONE INTERVIEWS

• Cheaper than personal interviews and can reach many people across a wide geographical area.
• Disadvantages include:
• People may not react favourably to long interviews
• Personal questions may be poorly received

8.3 POSTAL QUESTIONNAIRES

• Relatively cheap and can allow more questioning than a telephone survey.
• Disadvantages include:
• Usually low response rates
• Potential response bias – respondents may not represent the wider group
• No interviewer is present to prevent misunderstanding of questions or instructions

8.4 QUESTIONNAIRE DESIGN

• State the objective of the survey early.
• Make the questionnaire long enough to obtain the required information, but not so long that respondents lose interest.
• Use appropriate language.
• Avoid confusing question construction and unnecessary technical language.
• Decide whether demographic information is needed, such as age, sex or income group.
• Direct income questions can be difficult; occupation may help build a useful picture.
• For passenger satisfaction surveys, it may be useful to distinguish regular from infrequent users.
• Allow space for additional comments or suggestions.
• Decide whether names and addresses should be collected to permit follow-up questioning.

8.5 ADMINISTRATION OF THE SURVEY

• Decide whether data will be collected by company staff or an outside agency.
• If company staff will administer surveys, consider the training and facilities they will require.
• It may also be useful to obtain staff advice when designing the survey.

8.6 TESTING – PILOT STUDY

• Every survey method should be tested before the full survey.
• Conduct a pilot study using a small sample drawn from the target group.
• The pilot can reveal misunderstanding, unclear wording or misleading questions.
• The questionnaire can then be redesigned before the main research takes place.

8.7 ANALYSIS

• Plan how collected data will be analysed.
• Factual questions can often be analysed statistically to identify significant trends.
• Qualitative questions require other forms of analysis to extract useful information.
• Sample size matters: a very small survey has less statistical significance than a much larger sample.
• The material illustrates this by comparing a sample of 10 with one of 1,500.