Religious groups and marriage

In examining American economic and demographic research data, we noticed an intriguing pattern in the religious aspects of the data.  For example, see this chart below that shows the portion of high income people ascribed to major religious groups, and the marriage rate of those same people.  A 99% confidence interval is also displayed, and the size of the marker is based on the religious group's share of the United States population.  Based on some feedback, a small amount of added comment below.



A few further comments: While not shown, the inclusion of additional (less-populated) religions would more strongly evidence the pattern above.  For example, Mormons would have an above-average 20% high-income rate, and equally an above-average 66% marriage rate.  Now opening with the horizontal axis, this is fixed nominal categorical data type, and we tested inclusion of the next income category below it as well with the same result. Did not partition out the data at a certain arbitrary level to produce favorable (or unfavorable) results. There are some additional absorbing perspectives one could also explore, however the data didn't exist, and it's perfectly fine for anybody to therefore conclude that what's seen here is a throwaway.  Now as for the vertical axis, this is the portion of the religious group who are married combining multiple sources including the 2012 U.S. Census Bureau American Community Survey. While this was not our survey classification, it is jointly commonly accepted that the vast majority of people marry within their own religion, like my showing that >95% marry within their own race. Durable enough of a portion to not power the pattern shown.