Before you analyze your results—and even before you administer your survey—it can be helpful to create a “data analysis plan.” Forming a plan in advance can ensure your approach is appropriate for the question you are seeking to answer. Your goals are to understand the data you collected and answer the questions that led you to conduct research in the first place. When planning your analysis, you’ll want to consider the core of your research question. Namely, the gap you were trying to fill, and the research strategy that produced it.
Components of Survey Responses
Before diving into analyzing your data, you need to understand the different components. The building blocks of your data are variables, or units of the phenomena you’re studying. Once you have considered the relationship between your variables, you’ll want to consider the range of possible values, also known as a measurement scale. This can help you determine which analytical comparisons are feasible to make.
Analytical Options to Consider
| Types of Responses Received | Visualization & Descriptive Measures |
| Free Text, Short Answer | These questions can be the most challenging to analyze as there is typically very little structure to collected responses. As a researcher, you can attempt to create structure in participants’ responses by providing them with explicit instructions. If you successfully institute some common organization among respondents, it will become easier to observe patterns when using any of the various techniques for analyzing qualitative data. |
| Choose all that apply, Multiple choice | These types of questions offer respondents a series of categorical options to choose from as they see fit. In order to demonstrate what categories participants selected in their responses, researchers will commonly create a visual representation of the results. If you want further insight into how the population of interest–represented by your sample–thinks or feels, you can also use statistical tools. |
| Numerical, Scaled | These kinds of questions allow participants to respond with a number–commonly representative of a personal sentiment–or numerical category in response. There is disagreement among researchers about how scaled responses should be considered during analysis. Generally speaking, as long as you have a lot of responses (more than 100), you can treat scaled responses like numeric data. As such, this data can be summed up in charts. This is similar to categorical data, but the difference is that you can demonstrate responses using different visualization tools. |
| Mix of Categorical & Numerical | Using category and numeric information can help you realize a lot of interesting information representative of a population of interest. You can use tools such as pivot tables, summary measures by group or category, or more complex statistical tools. |
Personal Project
Remember, when choosing your approach to analyze the data, ask yourself key questions:
- What type of questions did you ask?
- What types of information did you collect?
- What is your purpose? Think back to your research question….are you interested in relationships between samples/variables?
