I'd love to see what types of charities households in different divisions (quintiles, whatever) tend to donate to. In particular, I wonder how charitable giving to one's religion differs -- if practices like tithing being more or less common explain some of this difference. But other interesting trends might pop out, too.

I think it might also be interesting to look at this trend over time, as well as the dataset as a whole. I think it would be interesting to segment by donation % of total income, and see what the average income came out to be, I wonder if there are variations within the quintiles that could allow us to draw some insights into donation patterns as well.

Let's not think the upper income people are too stingy. There is a positive correlation between income and both relative and absolute taxation. Taxation is a bit like charity, except you get thrown in jail if you don't pay it.

I'm too poor to worry about it, but I think if I gave 40% of my income to the government (which will mostly be used for various social programs ) I might not be particularly inclined to give that much additional money to charity.

George, I think if you were giving 40% of your income to the government, you would be making enough money that giving a little to charity really wouldn't hurt that much. Unless, of course, you just really need that new boat.

Why is before-tax income used, instead of after-tax, which is to say, actual, income?

To me this graph shows a minimum bound on giving.

That is, even lower classes buy a few candy bars for the schools or a box of girl scout cookies or throw money into the church collection box.

There's still a negative correlation, certainly, but not as pronounced as implied by the lowest quintile's data point.

Also, as someone else implied, with all the social programs tax revenues fund, I would imagine many in the highest quintile consider a good portion of my highest quintile taxes to be effective donations.

More about the tax thing:

In fact, looking at the graphed data there seem to be two step downs, the 4th and 5th quintile give at ~2%, while the 2nd and 3rd quintiles give at around 2.5%, and the lowest quintile gives at a higher rate (again, I acknowledge that this is crude data, and thus to find any real results beyond what's given)

I note that the first two (and the two largest) bumps in marginal tax rates (10-15% and 15-25% of income) happens at $16,700 and $67,900 of income, exactly where those two "steps" occur.

Clearly not definitive of anything, but interesting to consider.

I'm with AR (and Frank) - I think this graph is deceptive, and the trend will we eliminated if we look at actual income, after tax.

I'd like to point out, that the George who commented above is not me. I have a new name now, ever since George Harrison besmirched mine on the orange juice post, so there. I am the George who changed their name from Jenna in the comments. I notice that Original Will has the same impersonation/to common a name problem, too.

Signed,
Formerly "George"

interpreting a correlation from this data may not be fruitful. A confidence band around these estimates might show little differences among groups.

From my perspective, an interesting conclusion is that giving takes place across all levels. Additionally, we seem to have a baseline level of giving around 2%.