Survey

Metadata
Category
Terminology
Recorded date
2005-10-13
Tags
tools

Definition

The term survey refers to a sampling, or partial collection, of facts, figures, or opinions.

Using surveys

Surveys are a good method for collecting information in a standardized form from groups of people using a questionnaire or interview. A small amount of information is usually gathered from a large number of people using set questions. Short categorized questions on key issues allow answers to be easily filled in, tabulated and analyzed. Surveys allow a statistically significant sample of information to be collected for comparison, to identify larger patterns and structures and to monitor change.[1]

The most reliable means of conducting a survey is face-to-face. In some cases or with regard to certain issues, surveys may also be conducted by telephone, through the mail, or via the Internet, though the use of indirect means for conducting a survey is highly context-dependent.[1]

Piloting questionnaires

It is usually advisable to test the design of your questionnaire through piloting. Testing your questionnaire among a smaller group of respondents can help determine whether the questions generate appropriate data for your research, and whether the questions are meaningful for the informants and appropriate to the context, especially in relation to cultural or linguistic issues. It can also help in deciding how the data will be recorded and analyzed, e.g. whether certain questions should be closed (a specific choice of answers) or open (where respondents can answer in their own words). Once the questionnaire has been tested, the design of the questionnaire can be adjusted or changed as appropriate.[1]

Sampling

The are several ways of deciding who should be the respondents for questionnaire surveys. These include the random ‘lottery’ selection, geographical stratification and theoretical sampling. The sampling method chosen depends on the objectives of that stage of the research. To ensure the data collected is representative of the population as whole, random sampling may be used or households can be randomly selected from a map of the community. However it must be remembered that certain social, geographical and temporal aspects may impact on those who actually respond to the questionnaire. For instance if household questionnaires are undertaken during the day it may be more likely that female members of the household are at home.[1]

If however the research relies on gaining information from key population groups based on characteristics of wealth, gender, age, tenure status etc. more selective sampling methods are necessary. One example of this is ‘theoretical sampling’ (Burgess, 1992, p.209), whereby the choice of respondent is identified as someone who may have a perspective upon the themes and aims of the research. Which groups will be approached for their opinions can be decided after using such tools as a stakeholder analysis or well-being ranking. It is often important to use an aspect of selective sampling when it is important to ensure that minority groups, such as female-headed households, are included in a wider study.[1]

Examples of questionnaires used in case study countries in the ‘Regulatory Guidelines for Affordable Shelter’ project can be accessed here.

Examples of questionnaires used to study regulatory frameworks using a Sustainable Livelihoods approach can be accessed here.

Sources

Further Reading

Source notes