In 2008, California citizens passed a controversial amendment to the state constitution, Proposition 8, defining marriage as between one man and one woman.

Recently, the US Supreme Court has begun hearings on Prop 8, and for the article "Court Weighs Gay Marriage" in the Wall Street Journal Online Justice Kennedy is cited as being concerned about the children of gay partners:

Justice Kennedy showed sympathy for the argument that it was too early to know whether gay marriage could be harmful. No state authorized same-sex marriage until Massachusetts in 2004. "We have five years of information to weigh against 2,000 years of history or more," he said. At the same time, he continued, denying marital status to gay couples potentially harmed at least one group: their children, of whom there are nearly 40,000 in California. "They want their parents to have full recognition and full status. The voice of those children is important in this case, don't you think?" Justice Kennedy said.

Question: Are there 40,000 children of gay couples in California? Is there a census of this sort of thing, is it taken from school records, how do we know? Or is this just a statistic pulled out of thin air?

Political arguments involving children seem automatically suspicious because the occasional use of children to make emotion based political arguments in the USA is well known. The Daisy girl ad of the 1964 presidential campaign is an excellent example -- vote for us or this little girl might get nuked. Another more recent example, the Child Online Protection Act was a bill unpopular among internet activists who claimed it would have hobbled the internet in order to make it safer for children. Proponents argued, of course, that children shouldn't have to be exposed to all the garbage that is online.

I suspect in this case the Justice is repeating something brought to him in court, and if so, it should be possible to trace this claim to its proponents and perhaps some references that can be evaluated.

  • " occasional use of children to make emotion based political arguments in the USA is well known." not just in the US, and far more than occasional... Can't say about the number, seems high. But a lot would depend on how you define things, as always with statistics. And of course it should not influence what the definition of "marriage" is in the first place, not in a country where over 50% of all children live with a single parent.
    – jwenting
    Mar 27, 2013 at 7:03
  • @jwenting Do you have a source for over 50% of all children live with a single parent? That sounds incredibly high (would also be interesting to see how it compared to other countries).
    – gerrit
    Mar 27, 2013 at 9:03
  • @gerrit can't remember, read it time and again on blogs though. Could be it's specific for certain population groups or areas, it's bound to be worse in places like Detroit or Chicago than in small farm towns for example.
    – jwenting
    Mar 27, 2013 at 9:20
  • 2
    While researching this, found 2.5% of children under 18 living in US households are adopted. Head exploded. I had no idea it was that high. (Table 1)
    – Oddthinking
    Mar 27, 2013 at 11:29
  • @Oddthinking adoption of a "poor child" from another country was the ultimate fashion statement for the fashionable couple wanting to show off to the world how charitable they are. Kids as pets and status symbols, yes. Same in Europe, but the rules for adoptions are probably stricter here. I wonder what the ratio of imported babies vs. children adopted that were abandoned or removed by child services is.
    – jwenting
    Mar 27, 2013 at 14:46

2 Answers 2


Maybe but very unlikely as that estimate is dependent upon the what data you are using to make the estimate. Based upon newer data that number is likely invalid though.

The United States Census Bureau can help us in answering this question as they provide a break out of data on same-sex couples. In the 2010 ACS Brief on Same-sex couple households reported that nationwide, 115,064 same-sex couple households of which 94,627 households had only their own biological children present.

While interesting this does not furnish us with definitive information. This is where the Supplemental Tables from the 2010 Census come into play. Here we find the following data for California:

2010 Census summary file counts 
Total:                                  125,516
  Male householder:                      64,625
    Same-sex unmarried partner:          39,901
      With own children under 18 years    2,504
      No own children under 18 years     37,397
    Reported same-sex spouse:            24,724
      With own children under 18 years    7,812
      No own children under 18 years     16,912
  Female householder:                    60,891
    Same-sex unmarried partner:          34,159
      With own children under 18 years    7,121
      No own children under 18 years     27,038
    Reported same-sex spouse:            26,732
      With own children under 18 years   10,392
      No own children under 18 years     16,340

Which gives us a total of 27,829 same-sex households in California with their own biological children present.

This raises a problem though since 27,829 is less than the 40,000 claimed. However you will note that this is being reported as households which implies that there are at least three people present (both spouses plus at least one child). In this case we can get a fair estimate of how many children there are by finding out the average children per household. Here I had to use Families and Living Arrangements data from 2010 where California is listed as having an average of 1.95 children per household which now gives us an estimate of 54,266.6 children in same-sex households.

This number leaves us in an awkward situation though because we have a fairly wide range of 27,829 to 54,266.6 children that are in same sex households. If we split that down the middle we end up with 41,047.8 which is close enough to the claimed 40,000 that I would speculate is how that number came about. If we wanted to be strict about the numbers than 27,829 would be the safest number o use; however, it is reasonable to assume that some of those households will contain more than one child.

On 27 September 2011 released a report which called into question the results of the 2010 Census and thus the data reported there in. In the report it was found that as many as 28% of the same-sex couple households are likely to be opposite-sex households. The report indicates that that the form design for the 2010 Census is likely to blame for some invalid data collection.

This means that the 2010 American Community Survey is being recommended for use as a source of data instead of the original 2010 Census data. The updated figures look like the following for California:

American Community Survey 2010 estimates tabulated as reported
Total:                                   90,023
  Male householder:                      47,691
    Same-sex unmarried partner:          33,115
      With own children under 18 years    1,383
      No own children under 18 years     31,732
    Reported same-sex spouse:            14,576
      With own children under 18 years    2,231
      No own children under 18 years     12,345
  Female householder:                    42,332
    Same-sex unmarried partner:          28,014
      With own children under 18 years    5,685
      No own children under 18 years     22,329
    Reported same-sex spouse:            14,318
      With own children under 18 years    4,037
      No own children under 18 years     10,281

If we add this up we come up with 13,336 households with children and if we use the 1.95 children per household from before that gives us an estimate of 26,005.2 children in same-sex households. This is a substantially lower estimate than the previous high estimate of 54,266.6 children in same-sex households.

Thus, our conclusions are largely driven by the data that we are looking at. While the 40,000 children in same-sex households appears to be an extremely fair estimate based upon the 2010 Census data, when the estimate is updated based upon the revised 2010 American Community Survey data the estimate is quite high. A better estimate using the revised data would be somewhere around 20,000 children in same-sex households.

  • You might want to use the more accurate figure of 15698 instead of the dubious 27,829. The census revised its estimates after discovering reporting errors of 29% higher counts.
    – user1873
    Mar 28, 2013 at 1:46
  • Sure, you already have the link to the table, just use the 2010 Census preferred estimates at the top of the table.
    – user1873
    Mar 28, 2013 at 4:36
  • The scale of the errors reported in the first version of the census is astonishing.
    – Evargalo
    Apr 24, 2018 at 12:32

Based on my research, I have not shown that there are 40,000 children living with same-sex couples in California.

However, I am able to show that figure is consistent with US census data.

According to the 2010 US Census:

0.1% of 116,716,292 households have a same-sex partner with their own children. [Table 3]

(Also, this is an estimate to account for believed mistakes by householders in filling in the form.)

California accounts for 11,506,870 of all US households. [Table 4]

However, it has disproportionate number of same-sex households. 0.8% of CA households have a same-sex couple compared to the national average of 0.6%. [Table 6]

So, if we make the entirely unsubstantiated assumption that same-sex couple with children households are spread evenly amongst the same-sex couple households, independent of state, we can estimate that California will have:

11,506,870 Households in CA 
* (0.1% Same-Sex-With-Child-Households) 
* (0.8%/0.6% same-sex bias in California) 
= 15,342 same-sex-with-children households in CA.

Now, 0.1% is the smallest increment they show, and many of these estimates are close to that, introducing significant error. We need to be careful with precision. We can't go past the first significant digit.

So, the estimate produced is 10,000-20,000 CA same-sex-with-children households.

However, this is an significant underestimate of the number of children, as each household may have more than one child.

Conclusion: I have not proven that the number of children to same-sex couples is 40,000 in California.

However, back-of-the-envelope calculations, coming direct from US Census data, suggests that the number 40,000 is not an unreasonable figure. It is likely to be in the ball-park.

  • If someone can find a better reference, where this or a similar calculation was performed in a peer-reviewed setting, please post it, and I will delete this answer.
    – Oddthinking
    Mar 27, 2013 at 12:00
  • 1
    More importantly, can you find the Margin of Error and Confidence Interval of that statistic. This study used ASC and CHIS numbers in addition to Census data to come up with there calculations, but they don't list the MOE or CI.
    – user1873
    Mar 27, 2013 at 14:12
  • 1
    @user1873: Whoa! That reference is near perfect. Why don't you base an answer off it, and I can delete mine?
    – Oddthinking
    Mar 27, 2013 at 14:22
  • This is a good answer. I like the approach. The answer by Rob Z cites more direct data.
    – Paul
    Mar 27, 2013 at 21:55

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .