Wednesday, March 10, 2004

Quantifying the Roe Effect

Old Post: I originally talked about the Roe effect below.

It took me all day to do this, which explains why this is the only post today. I've been wondering about whether it's possible to quantify the Roe effect. According to Pia de Solenni's column in National Review:
Approximately 40 percent of American women under 45 have had at least one abortion. Twenty-five percent of all pregnancies end in abortion. Since the legalization of abortion in 1973, over 40 million abortions have taken place.

What I basically did was calculate what the 2000 population of each state would be if those 40 million had been born, then derive the electoral college representation for the revised population. Since I didn't have access to the complete statistics for abortions performed in each state, I used this source for the abortion rate per 1000 women aged 15-19. I took this rate to be representative, so that this rate times the total population of the state is proportional to the total number of abortions in the state over the last 30 years. (For state populations, I used the 1990 US census. Since the abortion rates were for the years 1988, 1992, 1996, and 2000, I used the rates from the year 1988 for simplicity.) This gives me a set of numbers proportional to the number of abortions per state, which I normalize to 40 million. I then reapportioned the House seats for each state given the new population. Any statistician can point out the multitude of problems in this analysis: taking one year to be representative for an entire three decades (which ignores changes in demographics and abortion laws), assuming the rate for teenage girls is representative of the rate for all women, not accounting for population migration, etc. A lot of these would be solved if someone could point me to a simple listing of the number of abortions received by the residents of each state since Roe v. Wade. In any case, with my limited information, I've calculated what the electoral vote would have been in the 2000 election using the Electoral college as apportioned by the 2000 census (NOTE: The 2000 election used the apportionment of the 1990 census, so this result differs from the electoral votes the candidates actually won.), and what it would have been if the 40 million citizens had been born. This is assuming that each candidate won the same states (unlikely, considering the millions of extra voters) and received all the electoral votes from those states (not all states are winner-take-all). You'll note that if the election had taken place under the 2000 Census, Bush would have won 278 to 260. Whereas with the revised population, it would only be 270 to 268. In the table below, those states that voted for Bush are in red, and those that voted for Gore are in blue.

State Population (thousands) Electoral Votes Revised Population (thousands) Revised Electoral Votes
Alabama 4447 9 4916 9
Alaska 627 3 703 3
Arizona 5131 10 5662 10
Arkansas 2673 6 2903 6
California 33872 55 42067 59
Colorado 4301 9 4767 8
Connecticut 3406 7 4096 8
DC 572 3 814 3
Delaware 784 3 902 3
Florida 15982 27 18420 27
Georgia 8186 15 9055 14
Hawaii 1212 4 1485 4
Idaho 1294 4 1356 4
Illinois 12419 21 14200 21
Indiana 6080 11 6583 11
Iowa 2926 7 3198 6
Kansas 2688 6 2931 6
Kentucky 4042 8 4336 8
Louisiana 4469 9 4821 9
Maine 1275 4 1408 4
Maryland 5296 10 6353 11
Massachusetts 6349 12 7504 12
Michigan 9938 17 11589 18
Minnesota 4919 10 5379 9
Mississippi 2845 6 2994 6
Missouri 5595 11 6151 10
Montana 902 3 972 3
Nebraska 1711 5 1866 4
Nevada 1998 5 2255 5
New Hampshire 1236 4 1409 4
New Jersey 8414 15 10095 16
New Mexico 1819 5 2011 5
New York 18976 31 22953 33
North Carolina 8049 15 9130 14
North Dakota 642 3 684 3
Ohio 11353 20 12572 19
Oklahoma 3451 7 3758 7
Oregon 3421 7 3864 7
Pennsylvania 12281 21 13745 21
Rhode Island 1048 4 1179 4
South Carolina 4012 8 4429 8
South Dakota 755 3 793 3
Tennessee 5689 11 6237 10
Texas 20852 34 22760 33
Utah 2233 5 2327 5
Vermont 609 3 684 3
Virginia 7079 13 8110 13
Washington 5894 11 6723 11
West Virginia 1808 5 1919 5
Wisconsin 5364 10 5825 10
Wyoming49435323


CandidateCurrent Elect. College TotalRevised Elect. College Total
Bush278270
Gore260268


Update: Talk about irony. I posted this on what La Shawn Barber notes is the National Day of Appreciation for Abortion Providers.

Update: It figures. I get a link from Best of the Web and it's on a post where I messed up the Old Post, New Post scheme. It's fixed now.

Update: One of my commenters says that unless I can show that women who've had abortions have less children overall (in other words, that they don't make up for the aborted children with children later in life), then my statistics don't hold. He has a point, but it would be hard to separate that from other correlations. I think the analysis is valid as long as the abortion is treated as a form of contraceptive. The Alan Guttmacher Institute argues that most women have more children than they want, and abortion is a necessary means for keeping that from happening. Doc Rampage argues that abortion is a symptom, not a cause, of lower family size with some groups. I'd be interested in seeing more data on this, if anyone wants to point me in the right direction.

No comments:

Post a Comment

I moderate comments on posts more than a week old. Your comment will appear immediately on new posts, or as soon as I get a chance to review it for older posts.