Originally posted by Merge:
Asking someone if their parents had a same sex relationship before they were 18 is A LOT different than asking of they were raised by a gay couple. Sorry, that study appears to be garbage.
Out of a sample of 18,000 people they were able to narrow it to 248 people who's parent had a gay relationship at some point, though we do not know if they were raised by a gay couple.
We really just don't have enough data available at this point to suggest any of the conclusions Regnerus is reaching towards. He used a set of the population born between 1972 and 1992, an era where being gay was really not accepted. We are just entering an era where it will be common to see a child raised by a same sex couple. In 20-30 years we will be able to get a much better snapshot of the data.
As Mark Twain said, there are liars, damned liars, and statisticians!
If a study is not random, then it is not representative. That said, the science of statistics and sampling, while -- like all science --- not perfect, is the 'best' science can do when the rules are followed.
As Seton75 suggested --- and I concurred --- in the other 'marriage' thread, people
almost always have some sort of an agenda. If one actually
reads the Regnerus
study linked above, it should be patently clear that Regnerus has no identifiable agenda and that he far exceeded
all of the previous 'scientific' studies on this topic in terms of breadth and randomness.
In isolation some sample numbers can seem very small. But have you ever seen the limited sample size of the highly reliable Presidential Election exit polling? Their 'universally accepted' (or at least TV Network accepted) conclusions are based on much smaller samples.
In the discipline of statistics the size of samples is scrutinized and almost always necessitate "error" controls via multiple 'statistics' (e.g. degrees of freedom, deviations, Type I and Type II errors). That's why Presidential polls are almost always quoted with ranges. It is only when randomly sampled results have deviations which are
beyond the accountability of these control statistics that the results are categorized as ''
significant''. All of the categories I listed above were deemed to be ''
statistically significant''.
Almost nothing in this world is perfect. But modern science does set the bar high in the scrutinized discipline of statistical sampling. Many lay people and columnists can cherry-pick numbers out-of-context from the study and then try to draw unscientific intuitive conclusions of the validity, but these are
not scientifically based opinions. That said, I am aware of no scientist or statistician who has challenged the methodology, the sample size or the analysis, but I have not read more than a few pages from the Google search lists.
IMHO it is 'unscientific', personally biased, and strictly subjective to characterize a validated study with arbitrary and insulting adjectives such as 'garbage'.
Someday, if the tide of relativism continues to flow, there might be larger populations of what Merge posted as 'same-sex couples' to sample. And we might get a more variegated 'snapshot', but that does not mean it will be ---as Merge posted---'better', nor that the conclusions will be any more 'normal'. Personally, I will always try to base my opinions on the best science available---now or in future.
Linked is Regnerus answers to some of his vicious critics as published in the Dallas News.
This post was edited on 6/12 12:53 PM by Old_alum
Regnerus Defends his study
