Who to ask? Edit

Sometimes it is impossible or too expensive to do a survey population. A survey population is where you conduct a survey to all potential market the firm is aiming at (total population). Therefore, it is necessary to have or select a ‘sample’ of the population. A sample is a group of people taking part in a market research survey selected to be representative of the overall target market, it is similar to a focus group but without long discussions.

In order to undertake primary research, a sample of the total potential market will need to be chosen. the larger the sample the more confidence can be given to the final result. The amount of people sufficient for a sample is 100 to 1000 people. Less than that then there is no confidence and there is no accuracy. A sample of 100 or 1000 will produce results that will reflect much more accurately the total preferences of the whole survey population. It will give you better confidence if the questions were divided into different age groups or income groups.

Though the idea of surveying 1000 people prevents all primary research. The two reasons are cost and time. The cost of the research gets more costly the bigger you sample size is, especially when a specialist firm of market research analysts is used. Surveying 1000 people will take a long time while managers are needed to make rapid decisions.

There are seven methods of selecting an appropriate sample:

  1. Probability Sampling
    • The selection of sample from population that has already had the probability or principle.It is time consuming and useually more costly than non-probability sampling. The selection of the sample is dones randomly thus making the result much more reliable and can be made about bothe the whole target market and the chances of errors occurring.
  2. Simple random sampling
    • Every member of the target population has an equal chance of being selected. To select a random sample you would need:
      • a list if all the people in the target population
      • sequential numbers given to each member of this population
      • a list of random numbers generated by computer
  3. Systematic sampling
    • Everynthterm item in the target population is selected until the desired size of sample is reached. The researcher must ensure that the chosen sample does not hide a regular patterns and the starting point must be done randomly.
  4. Stratified sampling
    • This draws a sample from a specified sub-group or segment of the population and uses random sampling to select an appropriate number from each stratum
  5. Quota sampling
    • When the population has been stratified and the interviewer selects an appropriate number of respondents from each stratum. This may work well, however the interviewer can be biased in their selection of people. This makes it less probability based than other methods.
  6. Cluster sampling
    • Using one or a number of specific groups to draw samples from and not selecting from the whole population, e.g. using one town or region. This method can help reduce the costs but it may not fully represent the whole population.
  7. Non-probability sampling
    • Non-probability sampling cannot be used to calculate the probability of any particular sample selected. It can’t be sued to make judgments or inferences about the total population. Result of this method must be analyzed carefully and filtered by the researcher knowledge of the topic being researched

Methods of non-probability sampling: Edit

  • Convenience sampling
    • Members of the population are chosen based on their relative ease of access
  • Snowball sampling
    • The first respondent refers a friend who then refers another friend … and so the process continue
  • Judgemental sampling
    • The researcher chooses the sample based on who they think would be appropriate to study
  • Ad hoc quotas
    • A quota is established and researchers are told to choose any respondent they wish up to the pre-set quota

What to ask? Edit

The questions you ask must be unbiased and unambiguous, it is essential if you want to obtain useful results. Do not ask too many questions in a questionnaire because the people might become suspicious or bored with so many questions. If it is essential to ask what age group they are or how much do they make on a daily basis it is best to avoid them because they would be reluctant to answer them. If it is a must then it is better to group them.

There are two types of questions you may ask in a questionnaire, they are: Edit

  • Open questions

those that invite a wide-ranging or imaginative response-the results will be difficult to collate and present numerically

  • Closed questions

questions to which a limited number of pre-set answers is offered

It is best to actually have no open questions at all but if it is essential then it is better to give out only one. The answers to an open question will be so varied  that it would be difficult to be presented statistically, it would be best if you can change it into a closed question.

Before actually conducting the survey it is best to do an initial pilot of the survey to test the quality of the questions. Other principles to follow include:

  • Make the objectives clear so the questions can be more focused
  • Write clear and unambiguous questions
  • Make sure that the questions follow each other in a logical sequence
  • Avoid questions that points to one particular answer
  • Use language that will be readily understood
  • Include questions that will allow a classification of results by gender, area lived in, occupation and so on

How to ask? Edit

How you ask the questions should be no problem, it could either be a face-to-face interview or could a be a telephone survey. It can be sent through mail or on the spot. Sending questionnaires by mail are cheap and t can cover a wide geographical area and there is no chance of someone distorting the results. However the response to both methods is nearly always very poor. The questions could easily be misunderstood and it can be biased in favour of the respondents who have the most spare time.

Direct interviews are done by interviewers usually on the street or in the respondent’s home. It would be best if interviewers avoid bias and interviewers would be able to explain the questions to the respondent well so they’d understand and maximum results can be obtained. The downside is it can be an expensive method but the interviewer would finish a job until the number of samples set is achieved, with the post method it is uncertain.

How accurate is it? Edit

Sometimes primary data can be not so reliable, there are three main reasons:

  1. Sampling bias
    • The only accurate method of primary research is when you asked the entire target population. Result from a sample may be different than the information obtained form the entire target population. If the selection of the sample is handled with not much care then the greater the degree of statistical bias will exist. Therefore no one can be 100% confident if the result obtained from the sample is accurate or not.
  2. Questionnaire bias
    • When questions lead respondents to one particular answer. This is called questionnaire bias, the results do not reflect what the people actually believe.
  3. Other forms of bias
    • Respondents not answering in a truthful way because he/she does not wish to admit something he/she is ashamed of

Ad blocker interference detected!

Wikia is a free-to-use site that makes money from advertising. We have a modified experience for viewers using ad blockers

Wikia is not accessible if you’ve made further modifications. Remove the custom ad blocker rule(s) and the page will load as expected.