Chart of the Day: US Cities According to Income Spread

Lately I’ve been thinking about how we compare Minneapolis to other cities around the Midwest, particularly when it comes to things like economic inequality. I wrote about it a week or two ago, and the common confusion between Minneapolis-the-metro and Minneapolis-the-city when thinking about social measures. So when I saw this chart on Twitter yesterday, it really caught my eye!

This shows the 95% income level for the top 50 US cities. As near as I can tell, this is measuring income on a strictly municipal level (i.e. not including the suburbs or Saint Paul):

incomes-US-cities-mpls-spread-ARROW

As you can see, compared any city with a similar median income, Minneapolis really jumps off the chart in terms of its “confidence interval.” Does anyone have any thoughts on why the confidence range might be so wide for Minneapolis, in particular?

7 thoughts on “Chart of the Day: US Cities According to Income Spread

  1. Cobo Rodregas

    Large confidence intervals in charts like these usually mean small sample size.

    But without any knowledge on the source of the graphic or how they did their sampling I wouldn’t speculate too much.

    Statistics are only as good as the data used to calculate them.

      1. Cobo Rodregas

        Good find. its a reputable source.

        I guess it makes sense since this appears to be within city only, Minneapolis has some of the poorest & richest neighborhoods in the metro area.

        I wonder how it would look if they used the entire metro (or preferably urbanized area). I love Minneapolis but its less than 20% of the whole, and just about everybody crosses city boarders every day.

  2. Paul Strebe

    Yup, I looked at the data and they are talking about Minneapolis proper, not the “Minneapolis” that is often used as a stand-in for the whole metro area. But it would be interesting to see the data on a more metro area for these cities, since more poverty has moved to the suburbs.

      1. Wayne

        My guess is small sample size and/or the presence of a few outliers. Without examining the data it’s difficult to say but generally the whiskers can denote the min and max (some forms of this plot exclude outliers from the whiskers and mark them as points beyond, but this doesn’t appear to do so).

  3. Chris Morbitzer

    When analyzing data myself, I almost always avoided the 1-year estimates for this reason. It’s likely that when it came to this particular data in the survey responses, that there was a lower response rate than in other cities or outliers that changed the confidence of the data. Its possible to download the raw data if you’re curious enough to take that time, but you might also be able to ask the Census Bureau and get a succinct answer.

Comments are closed.