Surveys only become useful when responses are analysed and trends are discovered. This provides evidence that can inform future action. Analysing survey results involves observing the incidence of responses, spotting patterns in charts and noting results that seem unusual. Ensuring your survey analysis is rigorous is the key to making it meaningful and worthwhile.
- Skill level:
Note the frequency of responses. This relates to the number of times responses occurred. For a paper-based survey, manual counting is the way to arrive at frequency totals. For a survey conducted via Microsoft Excel -- or similar application -- you can simply look at the frequency count boxes, if included in the survey form. The response with the largest frequency count indicates the most popular response. For example, if most people selected “online,” from the available options to the question “How do you normally renew your car insurance?,” you can ascertain that this is the most frequently used method.
Use graphical methods of representing tables, graphs and statistics to make analysis of survey results easier. Where a particular response occupies the largest section of a pie chart, it is clear that this is the most popular choice. Where a bar chart or column chart has an empty row or column, this shows that nobody selected that particular option. Likewise, where a sudden spike on a line chart is visible, this indicates a phenomenon that needs further investigation. It may indicate a new trend. Alternatively, it may mean that data input errors have crept into the survey results.
Work out what has gone wrong when survey results are erroneous. This is a way of ensuring your analysis is reliable. Sometimes, questions are framed incorrectly. For example, if the question is: “How many times a month do you go to the gym?” and the available answers are “1 to 4,” and “4 or more,” this will cause confusion. Some people who go to the gym once a week will select the first option and some will select the second option. Likewise, if you fail to include an applicable option, people may fail to give an answer or choose a best-fit response. Further, if you allow people to enter their own responses, rather than providing them with a set of choices, errors can result. Different responders may spell the same word in several different ways, leading to an apparent conclusion that there are many different responses when in fact there is just one. To analyse survey results rigorously, you must trap these errors and process them appropriately.
- 20 of the funniest online reviews ever
- 14 Biggest lies people tell in online dating sites
- Hilarious things Google thinks you're trying to search for