On Discover Magazine’s Gene Expressions blog, Razib Khan takes a look at how different groups of Americans perceive the number of Asian Americans and other minorities. He concludes that Americans consistently overestimate the number of minorities and underestimate the number of whites. He also states that despite 31% of Asian Americans marrying a non-Asian American in 2008, Americans overestimate the extent of interracial marriage, especially for Asian Americans. Are the concerns raised by one of Asian American blogs’ favorite topics totally overblown? How can Khan possibly justify his conclusions?
Let’s look at a snippet of the data Khan uses. The real percentages from the 2000 census are in the top row. Each following row is the average of the estimates of percentage of each racial group racial by given the group in the first column, from a study also done in 2000 (since this each item in each row is an average of estimates, the perceived percentages do not add up to 100). The last two columns show the difference between estimates of whites and minorities compared to the census numbers. You can clearly see how whites are underestimated and minorities overestimated.
How does Khan conclude that Americans overestimate the extent of Asian American interracial marriage? He says that they underestimate that amount of homogamy, or marrying within a group. If you look at the 2000 census numbers for Asian Americans, you will see that 3.6% of the population is Asian American. If interracial marriage was done a purely random basis (a person has the same chance of marrying someone without regard to race), then the chance of two Asian Americans happening to marry would be 1.3% (3.6% times itself), or an interracial marriage rate of 98.7%. Compare that to the 2008 interracial marriage rate for Asian Americans of 31%.
What’s the cause for these findings? Regarding the overestimation of the percentages of minorities, Khan guesses that the mainstream media, produced mainly in New York and Los Angeles, reflects diversity in those locations that are not present in much of the country. I’m highly skeptical of that, as I suspect as are many readers of this blog. Some studies instead suggest that mainstream media is actual becoming more white. Perhaps Americans are buying into “minorities are taking over” sensationalism.
Regarding homogamy in Asian Americans, he seems to have more reasonable answers. He mentions that there is a great deal of residential segregation, which increases the likelihood that people with marry within their group. Indeed, some cite the rise of Asian ethnoburbs as one of the reasons for the slowing of Asian American interracial marriage rates.
NOTE: 8Asians.com is a community, and we thank you for being a part of it. While we welcome and appreciate differences in opinion, if you're rude or you're promoting spam, we have a right to edit or delete your comment. Read our comment policy for more information.
If you see a comment that violates the 8Asians.com comment policy, you may flag the comment by mousing over the comment and clicking "FLAG."
The post says "Each following row is the average of the estimates of percentage of each racial group racial by given the group in the first column, from a study also done in 2000 (since this each item in each row is an average of estimates, the perceived percentages do not add up to 100)."
That doesn't seem correct. Imagine you had three types of people, X, Y, Z (e.g., asian, black, and other). If the first person's estimate is (x1, y1, z1) such that x1 + y1 + z1 = 100, the second estimate is x2 + y2 + z2 = 100. For n respondents, if you add all of the estimates up, then you'd get n*100. Rearranging you get (x1 + … + xn) + (y1 + … + yn) + (z1 + … + zn) = n * 100. We can divide each group's estimate by n to get averages, but thh sum of the averages would add up to 100. (x1 + … + xn)/n + (y1 + … + yn)/n + (z1 + … + zn) = 100.
Did the original estimates add up to 100%? Were the respondents taking into account people with multiple races or ethnicities?
thejrnf In the original article, the author writes "Why? Because I’m averaging the responses of individuals, and they aren’t talking to each other and figuring that you can’t get more than 100 percent as a collective whole."
I'm going to interpret this as the fact that they didn't ask for a full decomposition of the population from each person surveyed (i.e., the asked for x1 from person 1, but not y1 or z1) so the totals don't have to add up to 100. Though Ican't figure out why they wouldn't ask that...
thejrnf "lies, damned lies, and statistics"
it depends on how you add them fuzzy numbers together
Actually, the probability of two Asian people marrying isn't 3.6% squared. That's the probability of some random marriage you pick from the total population of all marriages being between two Asians. The probability of any Asian marrying another Asian given a random selection process is still just 3.6%, which is what you should compare the 31% to.
FenZhao Good math find! (although I think mathematically speaking we need more info to find out the probability of marriage in a total population being between Asians--what if Asians <i>never</i> married despite being 3.6%.) Nationalized, you're right--if we don't break down by gender, it should be around 3.6%.
However, I think the math should be localized rather than nationalized. If you are Asian living anywhere, chances are that your circle is much more than 3.6% Asian. Even those in the <a href="http://www.bigwowo.com/2010/06/asian-femal-celebrity-unio/">celebrity club</a> have lots of friends who are Asian, maybe even most. If you live in a very Asian area like San Francisco, no matter what your race, your circle probably includes lots of Asian people...some people's percentages are around 100%!
bigWOWO Indeed, one has to sort of make some vague assumptions about the homogeneity of the racial distribution over the whole population vs of marry-able individuals. There's likely some deviations of racial percent versus age that makes this not true.
The math should be localized in some way, but there you run into some different problems. If the ultimate question is about racial preferences in marriage, then doing it locally confounds the issue a bit. Take an extreme dummy case where you have a population of Asians living in a community that is 99% Asian. If the interracial marriage rate is 1%, then you would say that homogamy is representative, and make the conclusion that there's no slant to preference. But the very fact that you have a such a dense racial population population already indicates racial preference. Not sure what a better metric would be though; time to call a statistician.
FenZhao Hey, good catch of a subtle distinction that I missed. Just excellent! Minor correction, we should be comparing the 3.6% to 69%, which is the Asian to Asian marriage rate.
I have to add that I am just delighted (in a perverse way) to see an 8asians discussion centered around math!
Stats can be kinda tricky. Like even though when most Americans mentioned the word Asian, in most minds, it's Chinese, Japanese, Korean, Vietnamese, etc. the East/Southeast Asian groups. However, on the census, it expands to all groups who come from the Asian continent. Also, in some Asian groups, White, sometimes Black and Asian overlaps (like some will use white instead of Caucasian or in some rare occasions, they will use Black to describe other features rather than the Black African-American we Americans use when we say Black).
So, I mean interracial marriages might be exaggerated in some sense. Might be.
I think in the case of mainstream media, I think it's changed slightly to reflect diversity, though not as much or as fast as we all hope to be. Like for the most part, it's hardly changed and even if it did, it's very small. That's the perception I'm getting, not necessarily getting more white. Though I could understand why some people might see that way.
hmmm, i'm not sure if the mainstream media is becoming more white. you can look at major news networks on down to the local channels and they seem to have more anchors of different races.
Oct 14: (Seattle, WA) From Fields to Family: Asian Pacific Americans and Food
Feb 3: (Philadelphia, PA) Marvels and Monsters: Unmasking Asian Images in U.S. Comics 1942-1986
Feb 9: (Los Angeles, CA) East West Players presents THREE YEAR SWIM CLUB
Feb 17: (Los Angeles, CA) All My Sons
Feb 24: (Toronto, ON) SNOW, Opening at the Cumberland Theatres in Toronto
Feb 25: (Los Angeles, CA) Past Present I Future Imperatives: Queer Space Time
Mar 3: (New York, NY) Vong Pak’s ‘Electric Shaman’ Concert
Apr 30: (Sacramento, CA) California Asian Pacific Islander Policy Summit 2012: iAdvocate
