Thursday, August 23, 2007

Unbearable confusion of two-sided matching... or Forbes Strikes Again

I honestly would ignore this... but I feel obliged. We have been through this before. Yes Forbes has again come out listing us near the bottom on its "best places for singles" list.

Some years ago I had a message from someone at Forbes magazine. If I understood it right they needed someone to confirm that Pittsburgh was an 'old' region. Either I was busy, didn't realize they were on deadline, or couldn't comprehend they needed someone to "check" that statistic and I didn't get back to them. The next week I think was when the first Forbes "Best Places for Singles" list and we were ranked 40 out of 40. One of the key causes was clearly the demographics. I doubt anything I would have said would have dissuaded them from their list at that point, but I should have called them back. Might have earned a half sentence of explanation... but no... the angst machine has its ration for the day.

I just am not going to take the time to parse the entire Forbes story.. But it clearly places some weight on a metric defined as the percentage of the population that has never been married. They even say they double weight that metric. What does that mean? Here is the perspective that gets lost. Yup, no denying it, lots of old folks around. and guess what, most of them have been married in the past. It must be just a horrible place to date as a result right?

Here are some quick numbers. For the population age 15 and over in 2000, 26.2% of Pittsburgh's has never been married, which is indeed lower than 27.1% for the US population. Even if you think it matters that's less than a percentage point. But there is a bigger point. The percentage here is clearly impacted by our larger elderly population. If you look within age cohorts you get a different picture. Among those age 20-24, 83.8% of the Pittsburgh population had never been married, compared to 74% for the US as a whole. For 25-29 year olds, 50.6% of the Pittsburgh population had never been married compared to 43.7% for the US. In fact for every individual age cohort including seniors, the percentage of population that has never been married is higher in Pittsburgh than the US. Hell, it's even a better place for seniors to find a date than elsewhere. It's only when you aggregate the different (young and old) groups that you get Pittsburgh ranking below the US...

Maybe you think: so what?... old is old and unmarried is unmarried. Depends on what the point is to begin with. If there is a point that is. If you throw a dart at a map of Pittsburgh, you are indeed less likely to have it land on someone who has never been married compared to elsewhere. Yet if you go bar-hopping on the South Side one evening, among the younger population likely to be out, the probability that the person on the barstool next to you is unmarried is actually much higher than the US as a whole. So unless you are scamming for a date in a senior center you are not impacted much by this notional lack of singles here...

Update: Amos points out that a table would be clearer, and possibly exlain some of the confusing numbers. Here is what I pull from the 2000 Census on the marital status of US and Pittsburgh populations by age. You will see that the "never married %" is higher for all Pittsburgh age cohorts even though the overall percentage is lower.. again, an artifact of just having more elderly here.

United States Pittsburgh
Never Never
Married Total Married Total
15-17 11,626,885 11,869,522 98.0% 89,107 90,368 98.6%
18 and 19 7,279,799 8,041,530 90.5% 54,737 59,095 92.6%
20-24 14,085,200 19,025,980 74.0% 110,380 131,751 83.8%
25-29 8,394,149 19,212,244 43.7% 66,988 132,440 50.6%
30-34 5,247,870 20,365,113 25.8% 43,058 152,340 28.3%
35-44 7,167,680 45,905,471 15.6% 65,323 376,930 17.3%
45-54 3,313,475 37,578,609 8.8% 34,770 339,529 10.2%
55-59 757,864 13,383,251 5.7% 8,320 120,468 6.9%
60-64 518,607 10,787,979 4.8% 6,142 103,598 5.9%
65-74 798,277 18,501,149 4.3% 12,152 208,472 5.8%
75-84 519,442 12,317,262 4.2% 9,981 160,625 6.2%
85+ 204,122 4,160,561 4.9% 3,165 48,712 6.5%

Sum 59,913,370 221,148,671 27.1% 504,123 1,924,328 26.2%


Anonymous Anonymous said...

Seeing as how most singles want to marry eventually, would a city that has a large % of people who've never married show indications that it isn't a good place for sigles?

If the single population in a city stays single longer, that may indicate that the city is lousy at connecting singles with their future soulmates. A city where people marry faster may be better, if you are actively looking for Mr. or Ms. Right (as opposed to Mr. or Ms. Right-Now).

Thursday, August 23, 2007 3:41:00 PM  
Blogger Amos_thePokerCat said...

Ya, we went in circles about this never married thing last year. This time you did a very good job of explaining it with some detailed numbers. Of course, a table would have been short and to the point.

There is one thing. You say: "for every individual age cohort including seniors, the percentage of population that has never been married is higher in Pittsburgh than the US. ... It's only when you aggregate the different (young and old) groups that you get Pittsburgh ranking below the US...".

How is that mathematically possible, for each cohort the UNmarried% > married%, yet for the aggregate, suddenly married% > UN%? That does not seem possible.

Thursday, August 23, 2007 8:27:00 PM  
Blogger C. Briem said...

its a decent point questioning whether the singles statistics should be interpreted as cause or effect. I had thought about that a bit. I figure that in this case the things impacting high or low singles counts... mostly schools I bet, are not reflecting that a region is bad for singles.

amos is right.. I should put up a table which I will do when I have a chance. But many people question how statistics like that work. It is exactly the logic behind our educational attainment statistics. We can be more educated than the nation at every individual age cohort, but because we have so many of an older generation (who typically did not go to college, graduate school or even finish high school at rates comparable able to today) our overall educational attainment stats are not as high. If that still is confusing, hold on until I put up a table for the unmarried stats and maybe it will be clearer.

Friday, August 24, 2007 12:17:00 AM  
Blogger Amos_thePokerCat said...

I understand that one cohort can be larger than others. In PIT that clearly is the case with the older cohorts. Obviously, that can skew the average. ... but if each and every cohort is above the national average, how can the total be lower? Not that it will be not as high, but that it goes from above to below. It sounds like the sum is less than the parts?

On another note, I do not find it unreasonable that when formulating a list of cities favorable to singles, that you would want to weight the factors in favor of a city with a lot of singles, i.e. double weight that metric. Of course, they do tell you they double weight that factor which is more transparent than the self published "Places Rated" list.

Oh, and don't blame your self for not calling back Forbes. I doubt if they would have changed their selection of statistics, or weighting. I think he just wanted a pithy quote.

The other key cause for our fair region's consistently poor ranking is future job growth. Again we discussed this last time. Still not doing very good with the job growth here.

Of course, all of this really does not matter to a true Steeler mascot like Yinzer McMullet. The important thing is that we did better that CLE.

Really, it could be a comedy routine. You just might be a yinzer, ... if when presented with statistics that make PIT look bad, the first thing you do is make sure that CLE did worst.

Friday, August 24, 2007 4:25:00 AM  
Anonymous Anonymous said...

Some things just cannot be measured scientifically, but Forbes keeps on trying anyway.

Honestly, how on earth can you measure something like how good a place is for dating? Their effort looks well thought out at a glance but obviously means little, as you have aptly demonstrated.

I used to know a guy who took this ranking quite seriously though. It made my head hurt.

Friday, August 24, 2007 9:18:00 AM  
Blogger Amos_thePokerCat said...

Great table. Actually, it is an artifact of two things. Having more elderly, and having relatively less younger people. All the kids, GenX/13ers and Millenniums, of the Baby Booms that left during the 70's and 80's.

The "reason" why the pieces do not add up is the cohorts are different portions of the total. Less younger people, who are the most likely to be unmarried, and more elders who may be windowed, but not likely to be single.

Of course, the whole city accepted the bogus self published placew rated list seriously.

Friday, August 24, 2007 7:13:00 PM  

Post a Comment

<< Home