Interpreting Survey Results

SurvSoft Survey Software

Survey Articles

Survey Analysis Articles

Interpreting Survey Results

By Jennifer Nunez
2007-07-27

Interpreting Survey Results

In the rush and excitement to get a survey up and running, it is sometimes easy to lose sight of how you will interpret your results once you have them. All too often, people and businesses will conduct surveys only to find themselves with a lot of raw data and the nagging question of “Now What?” lingering in their minds. Understanding how to work with survey results is the key to uncovering new insights or creating new action steps.

Making it Count
Decide how you want to tabulate your results. Are you interested in all respondents, or only those that answered one way or the other? Some companies that are interested in enhancing customer satisfaction often “throw out” those responses that were initially positive so that they can concentrate on the responses that identify areas for improvement. Other surveyors are more interested in measuring general perception, which can typically be found by comparing one group of responses against another. Once the desired type of data has been chosen, it is a simple matter of giving results an initial “sift” in which similar answers are compiled into distinct sets. If the available responses for a particular question were Yes and No, then there will be two sets to begin with. If a question had four possible responses, then there will be four sets, and so on. After counting up the total number of responses in each data set, some surveyors have all the information they need. However, there is still a lot more that can be done with the data.

Revealing Results
Percentages and ratios are among the most insightful bits of information that can be gleaned from survey results. If a survey was conducted of 100 individuals to find out whether or not they would purchase a new kind of remote control, and only 5 respondents gave a yes reply – it can be reasonably assumed that the remote control isn’t currently well received. In fact, it shows that 95% of those surveyed would prefer either their current or a different remote control to the new one being proposed. It might then be prudent to explore ways to lower this percentage. On the other hand, if results showed that 60 out of 100 respondents preferred the new remote control, it might be an indicator that the product is ready to go to market.

It is important however to understand that percentages and ratios can be misleading if interpreted incorrectly. While the 60% preference example above is a good market “indicator” it should not be taken as fact. The reason is due to something statisticians call sample size. Understanding that it’s nearly impossible to predict human behavior, sample size helps gauge a margin of error. If you survey 10 people you may have valid results, but you can’t assume that the same results will repeat themselves if you were to give the survey to 10,000 people. At the same time however, if you survey 10,000 people and find that 75% (7,500 people) prefer one brand to another, it would be reasonable to assume that close to 75% of 100 (75 people) would react the same way. Simply put, the more people that are surveyed, the lower the margin of error, and the more reliably results can be applied to larger populations.

Zero or Hero?
Although many are quick to discount non-responses as in-valid data, these can be extremely useful results in improving the quality of a survey. For example, if the majority of respondents leave a certain question blank, the question itself may be the problem. It could be that the question is confusing, doesn’t apply to the situation, or lacks an accurate choice of responses. The same can be said for respondents that don’t participate in surveys at all.

When it comes to getting real-world impressions, few things are as effective as surveys. The key lies in knowing how to use them to their fullest potential. Surveys are the net that catches all the data, it still has to be sorted and cleaned before it can become the fuel to a successful operation.

Article Pages:
» Interpreting Survey Results

Related Articles:
» Analyze, Interpret and Bank Online Survey Research