I'm interested in using this Statistical Testing Tool (https://www.census.gov/programs-surveys/acs/guidance/statistical-testing-tool.html) to compare '10-'14 to '15-'19, but for percentages rather than for estimates. The examples all show estimates only. If both the numerators and the denominators are statistically different, does that mean the percentage would be? (My intuition would be no, since population in general has grown over that period, so the raw numbers for both would be higher.)
The tool works for both number and percentage estimates. You do not have to separate the percentage into a numerator or denominator to test the significance. If the numerator and denominator are statistically different, it does not guarantee that the percentages would be significantly different.
Thank you so much Mary!