You are currently viewing Analyzing your survey results

Analyzing your survey results

Before you dive into the data you have collected after your survey has closed, let me remind you about the types of questions that would dictate the kind of analysis we can use to analyze your data.

First, we have categorical questions, AKA nominal questions. These are the type of questions that respondents select from a list of categories for their response, such as male or female, yes or no answers.

Secondly, we have metric questions.  These are questions that a number can answer.  They can be similar to categorical questions but with some sort of order between them. For example, if we ask our respondents their age in categories like 18-25, 26-40, 41-55, this is a metric type.

And lastly, we have open-ended questions.  These can be answered by a number, for example, asking our participants how many webinars they have attended in the last year. Or it could ask to explain what makes a vacation memorable.

 

It is fundamental to get our questionnaire correct to determine what type of test we will use to analyze the results.

Once we have collected all our data, the first step to code the questions, this means that we look at each question and allocate a number for each response.

The next step is to transfer the data into an Excel spreadsheet or a statistical package such as SPSS.  Once all the data has been entered and cleaned, we need to determine a strategy for the analysis.

This strategy must be linked to the research question.  If the plan is to do a quantitative study, then the research question must reflect that.  Once we have summarized and described each question, we can determine where the data points to (central tendency) and how apart those points are (variability).

Finally, we find meaningful differences in the results.  We must ask if there is a fundamental difference obtained by statistically significant tests, such as descriptive analysis tests, T-tests, and Anova.  These Statistically significant tests will tell us the probability that the relationship exists, and it is not just an anomaly in the data.

Leave a Reply