Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
For Columbus using the formula you just gave:
1990 with bachelor's, 25+: 94,664
2012: 177,354
Total growth rate 1990-2012: +87.4%
Overall city population 1990: 632,910
2012: 809,798
Total growth rate: +27.9%
87.4-27.9= 59.5
59.5% is 19th on the list you provided.
These numbers were all taken directly from the census, and the 2012 from the ACS. So what numbers do you actually have?
Also curious why you used 5-year estimates rather than 1-year, which would only be for 2012.
This is my data:
1990 = 94,726 with bachelors or higher
2012 = 164,477 with bachelors or higher
1990 population = 632,910
2012 population = 790,168
Total difference = 48.788%
I use 5 year estimates because they have more data, better averages, etc. 1 year estimates are fine but they may not be as accurate because 5 year sees the trend better than a 1 year does. And in the end, because I used 5 year, saying that Columbus for the 1 year would be 19th on the list is false, because you'd have to recalculate every other city's percentages as the data would be different for each.
Its a shame that vocational classes like auto shop and carpentry have been eliminated from so many high schools.
Not everyone was made to get a four-year degree and not every degree holder is guaranteed a good living-and might I add, many great paying jobs dont require a degree.
It depends on the individual, their abilities, interests and acquiring the skills sets best suited for them.
I agree with this completely. It's really a shame how little people look at trade jobs, just about all of them offer the same middle class life you can get with a college degree (without having to pay so much student loans too as well). There are too many students who shouldn't be in college now in college who can't even find a job while trade careers are having a tough time filling many positions.
For Columbus using the formula you just gave:
1990 with bachelor's, 25+: 94,664
2012: 177,354
Total growth rate 1990-2012: +87.4%
Overall city population 1990: 632,910
2012: 809,798
Total growth rate: +27.9%
87.4-27.9= 59.5
59.5% is 19th on the list you provided.
These numbers were all taken directly from the census, and the 2012 from the ACS. So what numbers do you actually have?
Also curious why you used 5-year estimates rather than 1-year, which would only be for 2012.
This is a simple misunderstanding. I didn't do the original calculations, nor quite 100% understand the way the OP is putting the formula into words, so I can't copy their exact formula...but understanding the topic and the way the rankings fell I have a good grasp of it and I can use your numbers to show in detail where you are going wrong.
First Lets calculate what percentage of Columbus has a bachelors degree in 1990:
Degree holders / total population
94664 / 632,910 = 14.95%
Then 2012
177,354 / 809,798 = 21.90%
This is an increase of 6.95 percentage points. It should be noted... unlike Cleveland the city grew overall, including non-degree holders. So while it has a strong growth in total degree holders, the chances of knocking on a random door and finding a degree holder isn't increasing at as fast of a rate.
The 6.95% points represents a 46.57% increase from the 1990 value of 14.95%, however the total increase of degree holders is diluted by the increase of non-degree holders too.
This is a value much closer to marothisu's calculation trying to do the same thing with the 5-year data.
This number isn't a pure increase in the number of degree holders, rather it is a measurement of how much degree holders represent the area. The OP's measurement are better at showing the changing face of US core cities becoming more educated. Now as mentioned from my previous post it is impacted by how boundaries are drawn and isn't showing us what metro regions (to better capture the economic influence of a city) are growing more educated quicker.
I'd calculate out some of the other cities on the top 20 list as an example, but I just have the time to dig up the data. If you use your source to dig up the data, we can go over a few more for comparison.
I agree with marothisu that the 5 year survey estimates will be more exact, but can't show year-to-year changes. However, that shouldn't cause large differences in values in most cases. The year-to-year estimates are based on data and findings from the 5 year ACS.
The issue is the 5 year ACS is designed to get detailed data and more exact survey based estimations, but they collect the data needed over a 5 year period.
This means the data can't be used for year-to-year changes with the same degree with accuracy/statistical probability of being accurate, but they can use it to make less exact estimates year-to-year on some research subjects.
1990 = 94,726 with bachelors or higher
2012 = 164,477 with bachelors or higher
1990 population = 632,910
2012 population = 790,168
Total difference = 48.788%
I use 5 year estimates because they have more data, better averages, etc. 1 year estimates are fine but they may not be as accurate because 5 year sees the trend better than a 1 year does. And in the end, because I used 5 year, saying that Columbus for the 1 year would be 19th on the list is false, because you'd have to recalculate every other city's percentages as the data would be different for each.
Your population data for Columbus is wrong. 2012 had the population at 809,798. In any case, 5 year estimates are not necessarily the most accurate, as it seeks to find an average number rather than an actual one. For example, a city is estimated to have grown the following in a 5-year period: 1,000, 2,000, 6,000, 10,000, and 11,000 in consecutive years. That would be a total of 30,000, with an average growth of 6,000. The 5-year estimate uses that average rather than the final year's actual growth, which in this example would be much lower. Your methodology is flawed because it seeks to compare a single year's data, 1990, with the 5-year average of 2007-2012, not single-year 2012's actual estimate. Whether or not you'd have to recalculate all the other cities probably goes without saying, and I'm sure just using 2012 will change the order and Columbus may still not be anywhere close to the top 20. I just don't know why you started with a single year and don't end with a single year if you are trying to do a true end to end comparison.
This is a simple misunderstanding. I didn't do the original calculations, nor quite 100% understand the way the OP is putting the formula into words, so I can't copy their exact formula...but understanding the topic and the way the rankings fell I have a good grasp of it and I can use your numbers to show in detail where you are going wrong.
First Lets calculate what percentage of Columbus has a bachelors degree in 1990:
Degree holders / total population
94664 / 632,910 = 14.95%
Then 2012
177,354 / 809,798 = 21.90%
This is an increase of 6.95 percentage points. It should be noted... unlike Cleveland the city grew overall, including non-degree holders. So while it has a strong growth in total degree holders, the chances of knocking on a random door and finding a degree holder isn't increasing at as fast of a rate.
The 6.95% points represents a 46.57% increase from the 1990 value of 14.95%, however the total increase of degree holders is diluted by the increase of non-degree holders too.
This is a value much closer to marothisu's calculation trying to do the same thing with the 5-year data.
This number isn't a pure increase in the number of degree holders, rather it is a measurement of how much degree holders represent the area. The OP's measurement are better at showing the changing face of US core cities becoming more educated. Now as mentioned from my previous post it is impacted by how boundaries are drawn and isn't showing us what metro regions (to better capture the economic influence of a city) are growing more educated quicker.
I'd calculate out some of the other cities on the top 20 list as an example, but I just have the time to dig up the data. If you use your source to dig up the data, we can go over a few more for comparison.
I agree with marothisu that the 5 year survey estimates will be more exact, but can't show year-to-year changes. However, that shouldn't cause large differences in values in most cases. The year-to-year estimates are based on data and findings from the 5 year ACS.
The issue is the 5 year ACS is designed to get detailed data and more exact survey based estimations, but they collect the data needed over a 5 year period.
This means the data can't be used for year-to-year changes with the same degree with accuracy/statistical probability of being accurate, but they can use it to make less exact estimates year-to-year on some research subjects.
I think towards the end you're getting a sense of my problem with the numbers. A 5-year average end and a single-year starting point are not the same. 1990 was a census year, 2012 is not, but using only 2012 makes more sense in this case because 2012's numbers are not going to match the rest of the 5-year average just as 1990 would not have matched 1985 or 1988.
I think towards the end you're getting a sense of my problem with the numbers. A 5-year average end and a single-year starting point are not the same. 1990 was a census year, 2012 is not, but using only 2012 makes more sense in this case because 2012's numbers are not going to match the rest of the 5-year average just as 1990 would not have matched 1985 or 1988.
Well the biggest problem you had with the numbers was calculating it correctly.
Admittedly, the OP didn't state the formula very well.
The census source being used is a minor difference at best. (ie. my calculated 46.57% based on your numbers vs. their calculated 48.788%). We're talking small differences.
The reason they have 2012 estimates is that is the most recent American Community Survey released for 1, 3, and 5 years. (most accurate publicly available). For 5 years estimates this is data that was collected during the 2008-2012 time period.
Your population data for Columbus is wrong. 2012 had the population at 809,798. In any case, 5 year estimates are not necessarily the most accurate, as it seeks to find an average number rather than an actual one. For example, a city is estimated to have grown the following in a 5-year period: 1,000, 2,000, 6,000, 10,000, and 11,000 in consecutive years. That would be a total of 30,000, with an average growth of 6,000. The 5-year estimate uses that average rather than the final year's actual growth, which in this example would be much lower. Your methodology is flawed because it seeks to compare a single year's data, 1990, with the 5-year average of 2007-2012, not single-year 2012's actual estimate. Whether or not you'd have to recalculate all the other cities probably goes without saying, and I'm sure just using 2012 will change the order and Columbus may still not be anywhere close to the top 20. I just don't know why you started with a single year and don't end with a single year if you are trying to do a true end to end comparison.
No, it's not wrong. I'm using the 5 year ACS and the population is from the 5 year ACS, which is different from what you'd find on wikipedia. Why would I use a population for a number not from the same data set? It's completely idiotic to do that considering they're using different samples and also same sizes.
Again, even if I used the 1 year estimates for all data, I'd be shocked if the number was 19th, but I can certainly go ahead and calculate this for you..
Admittedly, the OP didn't state the formula very well.
I didn't state it in its purest mathematical form but this is pretty simple
Quote:
Originally Posted by marothisu
The final calculation is again the difference between the percent change of people with Bachelor's degree or higher age 25+ and the total percent change of population between 1990 and 2012.
I also stated I'm using 5 year ACS data, not 1 or 3 year:
Quote:
Originally Posted by marothisu
I used the 2012 ACS 5-year estimates for the data.
Here is the recalculated percentages with the ACS 1 year estimates instead of 5 year estimates:
1. Henderson, NV - 392.926%
2. Las Vegas, NV - 151.327%
3. Portland, OR - 114.489%
4. Miami, FL - 112.208%
5. Austin, TX - 107.244%
6. Atlanta, GA - 100.156%
7. Tampa, FL - 98.583%
8. Fort Wayne, IN - 95.198%
9. St. Louis, MO - 86.295%
10. Chicago, IL - 85.495%
11. Charlotte, NC - 82.595%
12. Nashville, TN - 78.366%
13. San Jose, CA - 77.679%
14. Denver, CO - 74.332%
15. Pittsburgh, PA - 74.197%
16. Cleveland, OH - 72.265%
17. El Paso, TX - 72.157%
18. San Francisco, CA - 70.145%
19. Seattle, WA - 69.049%
20. Jacksonville, FL - 67.422%
...
27. Columbus, OH - 59.157%
Now, if you do the absolute value of ((2012 Bachelors or higher - 1990 Bachelors or higher) / (2012 population - 1990 population)) using 1 year ACS, then it's like this:
1. Chicago, IL - 4.106
2. Washington DC - 3.798
3. Newark, NJ - 2.895
4. Philadelphia, PA - 2.329
5. Minneapolis, MN - 1.854
6. San Francisco, CA - 1.551
7. Atlanta, GA - 1.494
8. Oakland, CA - 1.471
9. St. Paul, MN - 1.386
10. Kansas City, MO - 1.237
11. Boston, MA - 1.106
12. Seattle, WA - 1.092
13. New York, NY - 0.842
14. Los Angeles, CA - 0.82
15. Milwaukee, WI - 0.795
16. Memphis, TN - 0.728
17. San Diego, CA - 0.726
18. Miami, FL - 0.718
19. Portland, OR - 0.707
20. Long Beach, CA - 0.689
...
29. Columbus, OH - 0.466
So while you're right to assume Columbus should be higher, it's still not top 20 (or 25). But it's closer now at least..so that shows perhaps an increasing trend year over year lately (other cities also had that - most of the top 10 did for example).
No, it's not wrong. I'm using the 5 year ACS and the population is from the 5 year ACS, which is different from what you'd find on wikipedia. Why would I use a population for a number not from the same data set? It's completely idiotic to do that considering they're using different samples and also same sizes.
Again, even if I used the 1 year estimates for all data, I'd be shocked if the number was 19th, but I can certainly go ahead and calculate this for you..
I know what you were using, I just don't particularly agree with why you used it or why it makes sense to compare a single year with a 5-year average. You're not using numbers from the same thing, so I don't know why you're saying you are.
I don't really care that much what the ranking is. I just don't think you came up with the numbers in a way that makes sense to the question you were trying to ask. In any case, the new calculation seems like a better way to view them.
Please register to post and access all features of our very popular forum. It is free and quick. Over $68,000 in prizes has already been given out to active posters on our forum. Additional giveaways are planned.
Detailed information about all U.S. cities, counties, and zip codes on our site: City-data.com.
Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
