Data color coding is a function that automatically shades values that are statistically significantly higher or lower than the overall values in the data displayed in the crosstabulation table.
📌Make sure to know!
To enhance the precision of data interpretation, DataSpace’s color‑coding rule will change from “confidence‑interval overlap” to a “Z‑test‑based statistical difference test.” (Effective date: 26 June 2025)
◾Before the change
Used the overlap (or lack thereof) between confidence intervals to flag values as higher or lower than the total.
◾After the change
Statistically verifies whether the difference in proportions (or means) between groups is significant using a Z‑test.
Determines whether an observed gap is due to chance or reflects a meaningful difference.
The data themselves do not change, but the number of highlighted cells may increase or decrease because the interpretation criteria are now more precise.
Color coding notation criteria
Red : Displayed when the result of a quantitative statistical test shows that the proportion of a specific group selecting a particular response is statistically significantly higher than that of the overall respondents.
Blue : Displayed when the result of a quantitative statistical test shows that the proportion of a specific group selecting a particular response is statistically significantly lower than that of the overall respondents.
📌Make sure to know!
Color coding only applies to variable results with 30 or more responses.
How to change color coding notation
The criteria for color coding in cross-tabulation tables vary depending on the confidence level you set.
Based on Z-test statistical difference testing, color coding is applied when a group’s value is statistically significantly higher or lower than the overall average. The default confidence level is set to 80%.
If you prefer a more stringent analysis, you can adjust the confidence level to 90% or 95% via the right-hand sidebar.
📌 What is the trust level?
The confidence level is a number that indicates how much you can trust the results of a survey. It is a number that expresses how many times the same results will be obtained when the same survey is conducted 100 times. For example, if the confidence level is 80%, it means that the same results will be obtained 80 times out of 100 times when the same survey is conducted. If you would like to know more about the confidence level, please refer to the article below on the Open Survey blog.