Cross tables are extremely useful while analyzing results. They allow you to observe the relationship between the two questions asked in the survey. The results are presented in the form of a table in which columns and lines form two selected questions with possible answers. Each cell of the table indicates the number of respondents giving a specific combination of responses.
Example of use:
If we want to check how the question was answered by women, and how by men, how answered people living in the countryside, and how in the city, etc. – it will be useful to reach for the cross table. Crossing questions is possible only for closed questions (single and multiple choice).
How to use cross tables?
- Click on the Results tab (1), and then click on the Cross tables (2). This move will expand the menu regarding this function and the form boxes for the crossed-out questions will appear next to the replies.
- Then select „Create a new cross table” (3).
- Choose two questions from the list of questions (4).
- To create a cross table, click Create a cross table (5).
- In the table created in this way, selecting the % from row field or % from column field will display the percentage value of the given answer in relation to the whole row or column (6).
- Clicking on Rotate data (transport) (7) will replace the question selected for the row with the question selected for the column and vice versa.
- After hovering over the result of the cross table, we will see the option Add to report (8) in the upper right year, which allows us to add a cross table to a new or existing report.
- We can export the created cross table in one of the three formats to choose from (XLSX, CSV). To do this, click Export (9) and then select the desired format.
Cross tables can be defined only for single-choice questions and multiple-choice questions. Thanks to the results filtering function, you can also get data from the intersection of other types of questions.
To learn how to create standard cross tables, go to the rest of the article. If you want to learn more about filtering results, please read the article Results filtering.
What is it about?
First, look at the format of the results depending on whether we use cross tables or filter results to process them. In the example below, we’ll use the gender question (1). At the beginning, we will focus on women.
The result obtained thanks to the cross tables (2):
The result obtained by filtering results:
As you can see in the above screenshots, that regardless of the method, in both cases the results are identical.
It can also be seen that depending on the chosen criterion – gender or age – a different number of filtering is required. In the case of a single choice question, eg a question about sex, only two filters are needed: for a woman and for a man.
In the case of the age question, filtering should be done as many times as possible to answer, e.g. how many age ranges have been included.
Although this type of analysis is quite time-consuming, it allows you to cross any type of question.
In the further part of this article, we will use the example of a short questionnaire containing two single-choice questions – a question about the age and gender of the respondent and a matrix question.
How to cross single-choice questions with single-choice matrix question?
- First, go to the „Results” tab (1), Cross tables (2), select the „Filter” option (3), and select the previously created filter or create a new one. If you do not want the crosstab to be created based on a filter, skip this step (3) and select „Create a new cross table” (4)
- In the first field (5), enter the title of the table. We select the first question, in this case the question of single choice about sex (6). In the next window, choose one of the matrix questions (7).
- After selecting the questions to the table, click „Create a cross table”.
How to cross a data segment with a question
The data segment is called the labels of respondents that can be created in the Contacts or Respondents tab. You can read about how to create labels in this article. Data segments can be crossed with the same questions as when crossing a question with a question.
- To create a cross table comparing a data segment and a question, go to the Results tab (1) and then the Cross tables (2).
- At the very top of the page, we can give your table a name (4). Next, we choose what we want to include in the table – in this case Data segment (5). From the drop-down list, one of the labels (data segments) previously created for the survey (6) is selected, and below is the question with which we want to cross them (7). Among the supported questions are: single and multiple choice questions, NPS, list and evaluation questions, drag and drop questions (DnD) and single-choice matrix. The whole is saved by clicking the button: Create a cross table (8).
- After generating the table, it will appear on your screen. To add it to the report, in the upper right corner, click on the option: Add to report.
You should remember that the segments are filtered – only those that have a maximum of 200 unique values will be displayed in the list (ie those that have any sense in the analysis). Also:
- Fills that don’t have a given label are aggregated to the „Not Specified” item.
- Segments are another, but not the last stage in developing the functions of cross tables in Startquestion. Another solution currently being proposed is the possibility of crossing questions with time periods.
Worth to remember:
- Crossing the one-choice questions with other types of questions can be very useful in situations where only a specific pool of respondents is needed, eg at a certain age or with a specific level of education.
- If we cross with each other questions that haven’t been completed the same number of times (eg optional questions, questions that the respondent could have bypassed through the filtering question, questions added after collecting the answers), there may be differences in totalized column and line items.
- It is possible to intersect only single-choice and multiple-choice questions and single-choice matrix questions.