INTRODUCTION
The following manual is intended to provide an understanding of the issues involved in survey design. It provides the key issues to be considered when designing surveys and potential survey designs and covers the advantages and disadvantages of the various methods available.
By reading these guidelines you would not be expected to be a "sample design expert". You should however be able to:
- formulate survey objectives to discuss with consultants,
- identify appropriate survey techniques to collect data,
- design and test simple questionnaires.
The information provided in this document is by no means exhaustive, and you are encouraged to refer to relevant books and journals to follow up areas of particular interest.
DEFINITIONS OF CONCEPTS USED WHEN COLLECTING DATA
The aim of this section is to enable you to become familiar with some of the more basic concepts involved in survey design. We will start off by providing a few basic definitions.
Survey
Surveys are used to collect information in order to answer a question or to make a decision. More formally, they are used to put a value onto some indicator or measure. Surveys measure one or more characteristics of a population. These characteristics may be measured by surveying all members of the population, or a sample of the population. A sample survey is a survey of a subset of the population.
Census
A census is an enumeration of the whole population. (See Samples and Censuses for more details.)
Sample
The sample is the set of observations which is taken from the population for the purpose of obtaining information about the population. (See Samples and Censuses for more details.)
Unit
A unit is the base level at which information is sought. There are two types of units: selection units and reporting units. A selection unit is a unit that is selected in the survey and a reporting unit is a unit that reports the information required. The reporting unit may not be the same as the selection unit (eg. in a household survey, the selection unit may be a household, but the reporting unit may be any responsible adult living in that household).
Population
The population is the complete set of units from which information is to be obtained and about which inferences are to be made. There are two types of population: target population and survey population. A target population is the population about which information is to be sought and a survey population is the population from which information can be obtained in the survey. The target population is also known as the scope of the survey and the survey population is also known as the coverage of the survey. (See The Set-up Stage of a Survey for more details.)
Frame/Sampling Frame
The frame is the list of all selection units in the survey population. (See Frames & Population for more details.)
Population Parameter
A population parameter is a summary measure of a population, the value of which helps to describe a population. An example of this is average weekly income, which is often used as an indicator of well-being or of spending power in the community.
Sample Statistic
A sample statistic is a summary value of a variable calculated from the sample.
DEFINITIONS OF CONCEPTS USED TO PRESENT DATA
From a sample of selected observations, we calculate certain statistics. Based on these statistics, we then make inferences about the population. For example, wine tasting is a form of sampling. A small sample of wine is taken, and from that sample, inferences are made about the contents of the whole vintage.
Mean
This refers to the average of a set of sample or population values. For example, the mean of 6, 7, 12, 15 and 20 is; (6 + 7 + 12 + 15 + 20)/5. The 5 corresponds to the number of values added together.
The mean therefore equals 12.
Variability
The spread of the values of a variable.
Variance
A measure of variability which measures the spread of the values around the mean. A larger variance indicates the data is more spread out around the mean.
Standard Deviation
Standard Deviation refers to the population variability based on the square root of the population variance.
Standard Error
Standard Error refers to establishing confidence levels in the data, based on the variability of a sample statistic.
The mean, variability and related topics are discussed in more detail in Analysis.