One way to make the article better is to have better racial classification systems in place. I would prefer the simply "white" be used in the chart, instead of white-non hispanic. Some Hispanic are from Spain in Europe, while others prefer to use the white classification. In addition, the black catagory includes black hispanics, and people who are mixed race, 80% of whom are partially white, are denied this part of their hertiage in the tabulation of city (and national) census data. Also missing are the people of two or more races, even though this is a census catagory. A good OP-ED piece appeared in the New York Times on May 8, 2001 on this matter, which serves to fan racial and ethnic tensions. It was written by Harvard University sociologist Orlando Paterson and was called "Race by the Numbers". Also, the ethnic group "carribean" is not an offically recognized census ethnic group, so unless there is another one, the Irish have to be moved into the top 5.

Wikipedia disambiguates West Indian to Caribbean. This doesn't mean that the "West Indian" group in the the cited reference doesn't exist; rather, it means that we should do a better job of making the disambiguation explicit on this page. I attempted to do this with my own edit, which reverted your changes. Please check the cited source; I've edited the link in the article to point directly to it, rather than to an index page.

Your comment about the use of "White (non-Hispanic)" in the census data table is interesting, because it shows just how many different categories of racial distinctions are available in the Census Bureau's data (spoiler : quite a lot, including both "White" and "White (non-Hispanic)"). Looking at the cited sources on this page (and on the Demographics of New York City page), I'm not sure why we're using one categorization over another -- what we need is an authoritative source that can tell us which racial divisions, per the Census Bureau, are appropriate for an overview such as this one. I suspect that a demographer would use "White" and "White (non-Hispanic)" preferentially in different contexts, and one is more useful for general overviews such as this. As I'm not a demographer, I can't say which one should be used here. However, personal preference isn't a good reason for changing this. Let's look for authoritative sources (Census Bureau outreach / publication guidelines, rather than op-ed pieces) and find out. A good place to start might be the Census Bureau's American Community Survey homepage.

Best, -- Docether 21:02, 14 August 2007 (UTC) reply

The Post recently released an article on Cenus Demographics http://www.nypost.com/seven/08102007/news/regionalnews/the_white_lie_regionalnews_bill_sanderson.htm This is a news article

The issue is the there are a lot of ways to crunch the census to bring about different results. There is no real authoritive way to do it, as far as I know. Let's put both in the chart.

This is the Census fact sheet for NY State: It is from the Census.gov. http://factfinder.census.gov/servlet/SAFFFacts?_event=Search&geo_id=01000US&_geoContext=01000US&_street=&_county=&_cityTown=&_state=04000US36&_zip=&_lang=en&_sse=on&ActiveGeoDiv=geoSelect&_useEV=&pctxt=fph&pgsl=010&_submenuId=factsheet_1&ds_name=DEC_2000_SAFF&_ci_nbr=null&qr_name=null&reg=null%3Anull&_keyword=&_industry=

National Census.gov site shows http://factfinder.census.gov/servlet/SAFFFacts

The offical city data from census.gov: http://factfinder.census.gov/servlet/SAFFFacts?_event=ChangeGeoContext&geo_id=16000US3651000&_geoContext=01000US&_street=&_county=New+York&_cityTown=New+York&_state=04000US36&_zip=&_lang=en&_sse=on&ActiveGeoDiv=geoSelect&_useEV=&pctxt=fph&pgsl=010&_submenuId=factsheet_1&ds_name=DEC_2000_SAFF&_ci_nbr=null&qr_name=null&reg=null%3Anull&_keyword=&_industry=

The offical site shows different numbers than appear on Wikipedia.

Warmest, Rock2003--14:00 August 14, 2007

The "Census 2000 Demographic Profile Highlights" format only presents a subset of the full data gathered. So, for example, there is in fact a "White (non-Hispanic)" data point presented in the full DP-1 (Profile of General Demographic Characteristics) for New York City -- it's listed as "Hispanic or Latino and Race" --> "Not Hispanic or Latino" --> "White Alone". But that's nitpickery on my part ... in any case, the real issue is that the full Census 2000 data contains far too many data points to show in this article. So how should we resolve this?

I propose that we follow the limited set of data points presented in the "Census 2000 Demographic Profile Highlights" documents. The Census Bureau obviously intends these documents to provide a high-level overview of demographic data, which should be sufficient for the casual reader of this article. Additionally, we can link from the table to the Census Bureau document, which allows the interested reader to drill down further into the dataset (the current footnote links to a strange quasi-unrelated document). Since the set of data points used in the "Census 2000 Demographic Profile Highlights" format seems to be a standard at the Census Bureau, we'll be relying on an authoritative primary source both for the data and for the choice of data points to be used in a high-level presentation. This should allow us to head off later questions about "why did you show this data point and not that other one?"

We should restrict the data points in the table to those referenced the article's text. The "demographics" section mentions total population size, population density, and cultural diversity in the first paragraphs. Additionally, median household income and the rate of home ownership in NYC is compared to that of the country as a whole, so these data points should also be included. I suspect that we can drop the "bachelor's degree or higher" data point, as it's not referenced in the section's text. I haven't found a Census Bureau source for the "percentage population change" figure, so I'm going to tentatively drop it as well, though I'll admit I haven't looked very hard.

A sketch of the modified table might look like this :

Obviously, the formatting is a bit odd, but that can be fixed. Also, each column's header should have a footnoted link pointing directly to the appropriate "Census 2000 Demographic Profile Highlights" document at the Census Bureau website. Comments from other editors appreciated. Best, -- Docether 19:07, 15 August 2007 (UTC) reply

You are correct, and according to the link above (2000 Census Demographic Highlights[ http://factfinder.census.gov/servlet/QTTable?_bm=y&-geo_id=16000US3651000&-qr_name=DEC_2000_SF1_U_DP1&-ds_name=DEC_2000_SF1_U&-_lang=en&-redoLog=false&-_sse=on full DP-1 ) there are two or three ways to look at it: the races alone or in combination with another race, what I will call "single races", and single races that do not contain any Hispanic orgion...i.e. people who might be both White and Hispanic or Black and White. In some racial catagories there are difference in the 10%-20% range between these different "interpretations". However, it seems that there are at least three different ways to interpret the data, which is why I think for the casual reader, your chart would work very well. There is of course several ways to crunch the numbers...but I think yours offers the most amount of variation without being cumbersome to the reader. Thank you for creating this chart.

Rock2003 22:56 August 15, 2007