# Calculating Incidence Rates

Simply defined, incidence is the frequency of something occurring in a given population, for example, x percent of adults in the United Stated chew gum.�In the marketing research world, the incidence rate is used to measure the level of effort needed to reach respondents that are qualified to take a particular survey (pass the survey screeners).�

As defined by the marketing research industry, "net effective incidence" is a summary measure of the frequency of individuals who meet all the qualifying criteria.�The "net incidence" goes one step further and is the frequency of individuals who meet all qualifying criteria and agree to take the survey.�These two calculations play a key role in the cost of market research because they describe how much resources (time and money) will need to be dedicated to the data collection process.

Most list and panel companies have subscribers profiled to varying degrees and are therefore able to estimate final survey incidence numbers. As a test of those assumptions, a soft launch with 10 percent of the anticipated total sample can be conducted to better estimate the net and net effective incidence. Greater care must be taken when using lists or working with panels that have not yet profiled their subscribers by the particular area of interest of your survey.�

This is especially true of business-to-business research.�If, for example, you purchase access to a group of subscribers who have maintenance titles, you cannot be sure they purchase the particular cleaning product of interest.�Further, assumptions must be tested carefully, as maintenance personnel at larger companies or in different parts of the country may have distinctly different needs and buying patterns.

The differences are significant.�As an interesting example, if you wanted to collect 400 completed surveys and the expected "net incidence" is from 80 to 90 percent, you could expect to contact between 450 and 500 potential respondents.�If the "net incidence" is from 5 to 15 percent, you could expect to contact between 6,000 and 8,000 respondents.�So, as the "net incidence" drops down the percentage scale, the difference has much larger pricing consequences.�For this reason, wherever possible, researchers should try and obtain the most targeted sample list possible while maintaining random stratification within the group.