Deboer v Snyder Day 8 – Update 1

Day 8 of the trial.

Witness:  Douglas W. Allen, Ph.D., called by the AG.

Full Professor at Simon Fraser University, Burnaby, B.C. Canada, Research Fields: Microeconomic Theory, Law and Economics.  Douglas Allen received his B.A.(hons) (1983) and M.A. (1984) from SFU, and his Ph.D. (1988) from the University of Washington where he studied under Professor Yoram Barzel. He was an assistant professor at Carleton University in Ottawa before moving to SFU in 1990.

His field of study is the economics of transaction costs and property rights, and he has applied this methodology to understanding institutions like marriage and divorce, welfare, the church, farm organization, homesteading, and the military. He is currently writing a book called “The Nature of the Farm” with Professor Dean Lueck.  The bulk of professor Allen’s teaching is in microeconomic theory and law and economics.

Some Recent Publications Include

  • The Institutional Revolution: Measurement & The Economic Emergence of the Modern World (Chicago: The University of Chicago Press, 2012)
  • “The Evolution of Criminal Law and Police During the Pre-Modern Era” (with Yoram Barzel) Journal of Law, Economics, and Organization 27(3) October 2011: pp. 540–567.
  • “The Ancient Olympics As a Signal of City-State Strength” (with Vera Lantinova) Economics of Governance  14(1), 2013: 23–44.
  • Nontraditional Families and Childhood Progress through School: A Comment on Rosenfeld” (with Joe Price and Catherine Pakaluk) Demography 50(3) June 2013: 955–961.

(Taken from the Simon Fraser University website)

One of the Circle of Experts for the Ruth Institute, an organization that promotes families at college level, and is opposed to same sex marriage.

Has published a paper in Demography based on Rosenfeld’s research, a paper based on the Canadian Census of 2006 relating to high school graduation rates among children raised by same sex couples, and a volume published by the Witherspoon Institute that “extends Rosenfeld’s study”.

He “worked on” four cases involving same sex marriage.  The Ontario case regarding same sex marriage, Halpern, an “Irish Case” where Irish citizens sought to divorce in Canada, the Perry case and this case.

He is a member of the Christian and Missionary Alliance Church.

He states: He studies economics of institutions, and because marriage is an institution, his work also applies.  He will testify as to 1) Literature since 1995 relating to same sex marriage, 2) Rosenfeld’s study and why there is more to it, and 3) his high school graduation rates paper.

He was certified as an expert in Economics with specialty in applied economics, empirical economics and the family.

On direct examination, he started with his literature review.  The article is titled More Heat Than Light, a review of literature from 1995-2013.  He looked at 60 studies involving same sex marriage and parenting in different household types.   Of those he stated that 54 were questionable because the results were not generalizable, they were small samples, biased samples (he explained this meant convenience samples) and that they were  “difference of means test with no power”, that there was researcher bias, they used “soft” measures and they were unable to replicate (because the research did not publish the data or make the code available).  He explained that “power” was a term of art that meant that the results had a “narrow confidence band” and that the studies had too much “noise”, another term of art that meant that it had a wide range or band of outcomes.

In a graph showing the articles he reviewed, 18 of them were dated 2001 and before.  He did not explain how he chose the studies he included.  Other witnesses state that there are about 150 studies published.

He testified that there were three studies he looked at published since 2010, the Rosenfeld study, the Regnerus study and his own study about high school graduation rates.  All of these were random sample studies, all had hard content, none were time series data, and all were gay & lesbian studies.  The Rosenfeld study had a gay sample of 3502 participants and a comparison group size of over 700,000.  The Regnerus study had a gay sample of 248 participants and a comparison group size of 2,988.  The Allen study (2013) had a gay sample of 8,632 and a comparison group size of 1,189,833.  Allen states that his 2013 study was comparable to Rosenfeld.  He cites to five other literature reviews, one by Steven Nock covering studies up to 2001, a review by Loren Marks covering studies up to 2005, and three by Walter Schumm dated 2005, 2011, and 2012.

Dr. Allen calls the 54 studies “preliminary” studies that can’t be generalized to inform public policy decisions.  He also stated that “counting studies” and finding a large consensus that “no differences exist is not a valid exercise.  Further, “standard practice” doesn’t resolve the problems.   He states that the research is just getting started.  When asked if the census studies should be considered as evidence, he states ”yes, because it is all we’ve got”.

In his critique of Rosenfeld’s study that examined the progression of children through school in relation to their household type, he criticizes Rosenfeld and quotes him as saying there is “no difference between children raised by same sex couples and opposite sex couples.  In fact, Allen states, there is “no statistical difference” and per Allen this is a different outcome.  He criticized Rosenfeld for not including a standard error factor.  He states that when the opposite hypothesis is tested, the results are different as between the heterosexual married couples and all of the rest of the categories (such as heterosexual co-habitating, separated/divorced/widowed women, never married men, etc.)  than those measures are when the same sex couples are measured against the other groups.  He states that the results are not robust to sample selection change because, adding back in those removed from the study as not the household head’s “own children” made a difference in the data outcome, that addition back in the participants removed because they did not meet the 5 year residency control made a difference in the results, and that the gender of the same sex parent and the gender of the children makes a difference.

His solution was to add back in the “own child” control and the “5 year residence” control  to increase the size of the study, and then tell the computer which ones experienced a divorce, which ones moved, which ones were foster, step or adopted children, and correct for other factors affecting outcome after the data is processed.  He states that it is only a dispute about how the calculation is done and the variables controlled for in the study.

Allen went into some discussion about how the Canada census differs from the US census and why that data lends itself to this study.  In his study he was able to separate the gender of the parents and the genders of the children and concluded that boys raised by gay men had a 69% greater odds of graduating from high school.  He concluded that “it looks like gender combination matters.  If shouldn’t matter if household types don’t matter.”

He admits that his work was criticized in blogs and by Rosenfeld when it was published.  The biggest concern was that the children were age 17-22 at the time of the census, and there was no information about what could have happen prior to age 12 that could have influenced the child’s odds of graduating from high school.  Also, they would have still had to be part of the household at the time of the census, and there was concern that children of opposite sex couples vs same sex couples may be more or less prone to living at home longer.

He again stated that there is not enough data and time to form any define conclusions. Power comes from sample size and from studies of different variables.

Opinions:  He stated verbally:  “Of evidence available there is no hard evidence that child outcomes are not different.  Must be cautious in making such a fundamental change to such a fundamental institution.”  However, the power point slide showing at the time he was saying this stated:

“Children in gay and lesbian homes do not perform as well as children in intact, opposite sex married homes.”  “Social Science is a long way from stating anything conclusive”  “Any conclusive statements are premature – not made based on any solid evidence since that evidence does not yet exist.”

{Editorial comment – There appear to be significant internal inconsistencies in these conclusory statements}

Cross will be posted later this evening.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s