Surveyonics - Online Survey Software
Principles of Online Survey Design & Delivery

For most organizations, it is not possible to survey all individuals of interest. For example, a company with 10,000 employees may wish to survey all their employees on their job satisfaction. Both the cost and logistics of administering the survey to the 10,000 individuals would make it prohibitive. This is especially the case when a company surveys its employees periodically in an effort to improve retention and productivity. Analyzing a large number of surveys can also be time-consuming and costly as well.

Organizations use a sample rather than an entire population in gathering information for a number of reasons as follows:
  • A sample is less time-consuming to survey than a population. Exceptions to using a sample occur when a company is conducting quality assurance processes and would need to certify all professionals of a certain designation. For example, if a company requires that all employees are certified in safety practices, then the entire population is surveyed to make certain that employees are up to date. Other instances occur when government offices conduct a census on an entire population. However, in most cases, an organization can only survey a carefully-selected sample.
  • Testing a sample is cheaper than testing a population.
  • Data from a sample is easier to analyze statistically than a population.

  • To gain more information about clients and customers without incurring high costs, organizations choose to use samples. It is important to note that results from a sample is only useful if it is representative of the larger population. For example, if a company wishes to develop a new dish-washing product and expects that 80% of their customer base would be stay-at-home mothers, their sampling strategy must reflect this demographic. Choosing a sample that is comprised of 20% or 30% stay-at-home mothers would not be representative of the population that the company wishes to target. To avoid inaccurate and/or biased results, an organization has to make an effort to choose a representative sample.


    Before an online survey is prepared, a sampling frame has to be identified. A sampling frame is a listing of all elements within a population. A university’s database of all students currently registered is considered a sampling frame. The current students registered in the university represent the entire population. Another example of a sampling frame is a local government’s full-time and part-time employees. Other sampling frames could be geographic maps and directories.

    A sampling frame can provide inaccurate or biased data if:
    • The frame does not include certain elements or individuals in the population necessary for the information to be gathered.
    • Certain elements or individuals under-represent the desired demographic that the organization wishes to target.


    Choosing a representative sample from a population is a multistep process that ensures the information received is useful. In the sampling process, the following steps must be conducted:

    1. Defining the population
    – In this step, a population is defined for surveying. If an organization is interested in the purchasing behaviors of college students in a particular city, then all students in that city are considered a population. For some survey studies, the population is simply defined as the consumers (e.g. Internet users or mall shoppers). However, marketing strategies focus on specific demographics to survey in a population. If the manufacturer of specialty rugs is interested only in the buying preferences of upper middle class residents, people who make a certain amount of money and above (e.g. $250K per year) would be the population. A clear definition of a population is important for the accuracy of the remainder of the steps.

    2. Developing a sampling frame – As described, a sampling frame provides a source or a listing of all elements or individuals within a population. In the example of the specialty rugs manufacturer, a sampling frame of upper middle class individuals could be public records that show tax and income figures. Since those records reflect all high income earners in one city, they are considered the sampling frame for the survey study. In sales and marketing, a sampling frame is not as easy to obtain as customer lists may not be available. For many organizations, sampling frames are usually previous customers’ lists or those purchased from other companies.

    3. Determining sample size – Once a sampling frame is identified, a sample size is determined. The size of a chosen sample depends on a number of factors: the number of questions in the survey, the type of questions, and the purpose of the survey. Sample sizes can range from 30 to several hundred depending on the availability of time and cost.

    4. Specifying sample method – This final step in the sampling process determines the sampling methodology. For instance, a survey may require only answers from experts in a field. Another survey that is informal may be given to any customers that frequent a business without regard for the population from which it is drawn.


    In the final step of the sampling process, a particular methodology is chosen and applied. This methodology depends on the type of sample that is surveyed. Samples are divided in probability and non-probability samples:

    1. Non-probability samples – Samples chosen out of a target population without worrying about their probability of occurrence are considered non-probability. There are four types of non-probability samples:

    • Convenience sampling – In this type of sampling, customers are pulled out of the population because it is the easy and convenient thing for the organization. For instance, a shopping mall manager may survey the first 200 customers who appear for shopping in a Saturday morning. The goal would be conduct the survey with a minimal investment in time and cost.
    • Judgment sampling – In this type of sampling, a sample is chosen based on their specific expertise or knowledge. For instance, a sample of doctors may be chosen by a medical equipment company to find out their preferences and requirements.

    2. Probability samples – Probability samples are chosen in a population based on known probabilities of occurrence. For instance, in a sample of 300 employees of whom 50 will be chosen, the probability of each worker being chosen is 1/6. Probability samples can be as follows:

    • Simple random – This type of sample means that each element or individual in a sampling frame has an equal chance of being selected. If a promotions campaign for a company is surveying customers based on a previous list, then every customer has the same chance of being selected for the sample.
    • Stratified – In a stratified sample, the population is divided into ‘strata’ based on a common quality such as demographics or earnings. For example, the specialty rug company owner can divide customers to those earning $250-$500K and those earning above $500K. The two earning brackets are considered strata. In some samples, a company can have upwards of 5 or 10 strata.
    • Systematic – In a systematic sample, the population is divided into groups with differing characteristics from which one sample is chosen.
    • Cluster – In a cluster sample, representative samples from the population are chosen based on clusters determined by those preparing the surveys. A large company can divide its employees into clusters based on the number of years they have worked in the company.

    Once a sampling methodology is chosen, the next step is to plan the creation of questions to be included in the survey.

    Previous                                                                                                         Next

Survey Features

User-friendly survey creation
Surveyonics online survey software gives you the freedom to create, analyze, and publish surveys in minutes. Use existing survey templates or create fresh surveys.

Choose how users can respond to your Survey questions
Define responses as single correct option, multiple correct options, text boxes, priority lists and various other formats.

Popularize your Surveys
Allow more and more users and friends participate in your survey. Easy sharing tools from Surveyonics help you link your surveys with your social network profiles.

See results as they pour in
Instantly view all the responses to your surveys in real-time. Surveyonics guarantees speedy reporting and uninterrupted reporting.

Customize Result Report
Smart survey result reports from Surveyonics help you to apply various filters and parameters to your reports.