You've designed and administered a survey and you have some results. Now what? You could spend hours looking at each person's responses, but this is very time-consuming and you may lose sight of the bigger picture. The goal of effective survey analysis is to draw accurate and meaningful conclusions about the survey results without straying too far in the weeds of the data.
Basic survey analysis follows three steps:
1) Organize the data
Depending on the extent to which your survey was distributed, you may have dozens, hundreds, or even thousands of survey responses. The first step involves consolidating and organizing the respondents' answers to the survey questions.
One common method of organizing data is to enter the information in Excel spreadsheets or tables created in Word documents. The respondents' names or some other identifier such as their initials (which preserve the respondents' confidentiality) are placed in rows and their answers to the survey questions are placed in columns.
For example, in the following online survey administered by NGC, the employees of a company are asked 3 questions:
Let's imagine that 10 employees responded to the survey (hopefully you'll have more responses to your survey!). We would then use a data organization tool, in this case Excel, to list the employees (by their initials) and their answers to each of the questions.
Looking at the Excel document, we can see that respondent #1 (“JS”) is a 62-year-old woman who selected “agree” for the life-work balance statement.
2) Describing the Data
Once the data is organized, we can perform elementary data analysis techniques and answer preliminary questions about the respondents.
One common way to begin analysis is to calculate the number and/or percent of respondents within each category of a question’s answer choices (this is called univariate analysis). This type of analysis allows us to answer questions such as: What percent of survey respondents are female? What percent of survey respondents are older than 40? What percent of people agree that their employer provides life-work balance?
Using the data in our Excel spreadsheet, we can create the following tables to summarize the data (these tables are called frequency distributions):
We can see that 6 respondents (60%) are female and 4 respondents (40%) are male.
In the table above, the response categories for the life-work balance question are listed in order of agreement (i.e., from “strongly agree” to “strongly disagree”). Focusing on the first two categories, we can see that 6 respondents (60%) either “agree” or “strongly agree” that their employer provides life-work balance.
For ease of analysis, the respondents are classified as either under 40 years old or 40+ years old. We can see in this table that 6 respondents (60%) are less than 40 years old and 4 respondents (40%) are 40+ years old.
3) Looking at Relationships
A second and more complex way to analyze data is to look at relationships between the answers to the survey questions (this is called bivariate analysis or crosstabs).
One of the most common ways (but not the only way) to look at relationships is by sorting answers according to characteristics such as gender, race/ethnicity, income level, age, years of education, etc. These characteristics have been shown to influence peoples’ attitudes, opinions, and behaviors. Thus, sorting data according to these characteristics can be very enlightening.
To continue with our example, we can construct tables (i.e. crosstabs) showing level of agreement with life-work balance by gender.
Using a basic bivariate analysis technique, we can compare the percent of women who strongly agree to the percent of men who strongly agree. Whereas 50% of the men strongly agree that their employer provides life-work balance, none of the women strongly agree that their employer provides life-work balance. These results suggest that there are gender differences with respect to life-work balance.
Now let’s look at level of agreement by age.
The table reveals that 25% of people 40+ years old strongly agree that their employers provide life-work balance, but only 16.7% of people younger than 40 strongly agree that their employer provides life-work balance. These results suggest that there are age differences in life-work balance.
Now that you’ve analyzed your data, you may be interested in other “how to” topics, like How to Build a Kick-Ass Survey.