Survey Analysis

In the previous chapters Pimprenelle designed her online survey aimed at the other pets of the neighborhood and sent invitations to take the survey to potential respondents. Now, the first responses have been submitted and Pimprenelle is impatient to have a look at them. Here is how to do.

Open your Survey's Insights Platform Page

For each of your surveys, GA Sciences offers online statistical tools that will help you analysis the results. This Insights Platform is accessible as soon as the first respondent submitted his answers to your survey. To open it, once logged in, you just have to click on the My Surveys item in your personnal menu, accessible either under the dropdown item in the navigation bar on the top of any GA Sciences webpage or directly from the left side menu in your Dashboard.

My surveys item in Pimprenelle's personnal menu
My Surveys button in Pimprenelle's personnal menu on the upper right corner of GA Sciences webpages.

On the "My Surveys" page you will see a table listing all your GA Sciences surveys, as in Pimprenelle's example below.

Pimprenelle My surveys table
Pimprenelle's My Surveys table.

In the "Analyze" column you will see a pictogram for each row. Click on the one that corresponds to the row of the survey you want to analyze. This will open the GA Sciences Insights Platform for your survey.

Pimprenelle survey analysis welcome message
Pimprenelle's survey analysis welcome message.

As stated in the above message displayed as the welcome message of the Insights Platform of Pimprenelle's survey, the page is devided in three distinct parts

When clicking on the "Trends" tab you will automaticaly scroll to the Trend Analysis section. That's where you will monitor the responses momentum of your survey by looking at curves where the number of responses your survey received during a given period of time is plotted Vs time. So far these time curves are either available on a dayly basis for the last gliding week, or on an hourly basis for the last 24h. As shown in the example below, Pimprenelle received 9 responses to her survey on August 12th, and only 2 responses the following day.

Pimprenelle survey trend analysis
Pimprenelle's survey trend analysis.

Coming Soon It will be possible soon to take a look at the number of answers by locations on a map.

Question Based Analysis

The second tab of the GA Sciences Insights Platform will help you perform a statistical analysis on the answers provided to specific questions of your survey. It is mostly based on graphical analysis and descriptive statistics.

Depending on the question type the analysis and graphical representation will differ. In fact the answer to a survey question might be considered as a variable and depending on the question type, the answers fall into a given variable category. In statistics variables are usually divided in two categories:

  • quantitative variableis, are variables that are numerical
  • qualitative variableis, also called a categorical variables, are variables that are not numerical

Quantitative Variables

The answers to questions of types Integer Number, Real Number, Star Rating and Slider are condidered as quantitative variables.

As you will see in the four examples below, quantitative variables are displayed in the GA Sciences Insights Platform by splitting the view in two parts: the left part of the view is a histogram representing the variable's distribution and the right part of the view are descriptive statistics applied to that variable.

Note that depending on the question type histograms are slightly different: for Star Rating type histograms are horizontal, for Slider typethe number of bin is equal to the number of steps on the slider's scale and for Integer and Real Number types the bin width is determined based on the Freedman-Diaconis rule.

Pimprenelle integer number bar-chart
Pimprenelle's integer number bar-chart.

Pimprenelle real number bar-chart
Pimprenelle's real number bar-chart.

Pimprenelle star rating bar-chart
Pimprenelle's star rating bar-chart.

Pimprenelle slider bar-chart
Pimprenelle's slider bar-chart.

Qualitative Variables

The answers to questions of types Single Choice, Multiple Choice, Short Text, Long Text, Date/Time, Ranking and Contact are condidered as qualitative variables.

Short Text, Long Text and Date/Time Types of questions

Short Text, Long Text and Date/Time answers are simply displayed as a table so far. Below is an example from Pimprenelle's survey. You can see that only the answers of the three last respondents are displayed but you can add as many as you want by clicking the "Show more" button. Note also that you can inspect in details all the answers provided to other questions of the survey by a given respondent by clicking on "Show details".

Pimprenelle three last answers to qualitative question
Pimprenelle's answers table for qualitative questions.

Single Choice and Multiple Choice

Single Choice and Multiple Choice answers summaries are displayed in the GA Sciences Insights Platform as pie charts.

Pimprenelle single choice pie-chart
Pimprenelle's single choice question pie-chart.

Pimprenelle muliple choice pie-chart
Pimprenelle's multiple choice question pie-chart.

Also, as you can see on the image above, the right part is deoted to descriptive statistics.

Ranking

Ranking answers summary are displayed in GA Sciences Insights Platform as multiple bar charts: one bar chart per item to rank. each bar chart is made of as many bars as the number of items to rank, each bar reprsenting a relative weight of the given postion of the given item in the submited answers. For instance, in the example below, it seems that the item "A clean laundary basket" appears three times in position 1 and six time in position 6.

Pimprenelle ranking bar-chart
Pimprenelle's ranking question bar-charts.

Respondent Based Analysis

The last tabs of the GA Sciences Insights Platform is basicaly a table showing the answers of individual respondents. Initially it is just the ten last submitted responses to the survey but you can display as many as you want by clicking the "Show More" button next to the bottom left corner of the table.

Now let's tackle the way this table is structured: each row corresponds to one respondent. Initially only the ten last submitted responses are loaded but you can display as many responses as you want by clicking the "Show More" button next to the bottom left corner. They are sorted in reverse chronological order, that is that the last respondent appears first.

The first columns of the table are devoted to metadata, that is data that provides information about other data than the answers the respondent provided to the survey questions. In order you will find the date and time at which the respondent answered the survey, then the location (country and/or city) of the respondent (if available), the respondent's IP address (if available) and the type of terminal he used to answer the surey (e.g. Desktop, Mobile, Tablet) (if available). If the data is not available, the corresponding box in the table will simply remain empty or will be noted as "Unknown".

The next columns corresponds to the questions asked in the survey: one column per question in the survey, with one exception for questions of type Contact. Indeed, as you have seen in the Survey Design chapter, a question of type Contact is actually a set of sub-questions. In Pimprenelle's survey example she asks the respondent his name and email address. So instead of one single column every Contact question will be represented with as many columns as the number of sub-questions it contains. In the example of Pimprenelle's survey there will be two columns for the Contact question she asked: one containing the name of the respondent, one containing his email address.

Pimprenelle survey responses summary table
Pimprenelle's survey responses summary table.

As shown in the above example of Pimprenelle's survey, this table can be dense and hard to read. But you can export it and read the full table using your favorite spreadsheet software by clicking on the "Export" button next to the upper right corner. It turns out that exporting your surveys raw data is the subjet to the next chapter of this documentation where you will learn in details how this table is structured.