Charity, November 12 at 9:04 pm Link to this comment
(Unregistered commenter)

I too would love to know why these people are so desperate to prove a point!? I’m sick of gays trying to rally and come together to show the world it’s “okay”. Who cares, just go be gay.

The point, since you obviously missed it by that much, is that people today still cannot simply “go and be gay”. Or adopt, or gain elementary rights all straight folks enjoy.

rico, suave, November 10 at 11:48 pm Link to this comment

Wow. 78 households. Now how can you quarrel with those stats?

Why are these people so desperate to prove a point?

Do you even read what you post? Condescension is about the best you could hope for posting crap that hides homphobia behind a thin veneer..“these people” indeed!

ardee:

Thanks for the condescension. The article was about the instance of child abuse in lesbian homes, not about whether or not it’s a good idea that gays be parents. It’s a fine distinction, but one that should not have been lost on you.

I don’t care if lesbians want to be parents. In fact, let me stipulate for you that I think it’s a wonderful idea.

I do care that someone draws conclusions from a “study” which samples only 78 subjects.

rico, suave, November 10 at 11:48 pm

Oh rico, the point is that people are people despite the stereotypes placed upon them. With so many children in need of good homes and people that will love and nurture them it is damn silly to deny a couple the chance to become a family.

rico, suave, November 11 at 1:41 pm Link to this comment

Agree. If you only have 78 in your sample, you had better be able to defend you sampling protocol pretty vigorously as well.
****************

That should be the rule for ALL scientific investigation. 

If you ever look at drug testing protocols, Phase I and Phase II studies are about that size—Phase II is about 50 to 100 subjects.  Phase I is much smaller numbers.

There are two keys:

1) Is your sampling sizing and stratification and selection process correct?
2) Is you sizing sufficient to produce potentially statistically significant results? (ie, it can be determined in advance if your sample can meet the 95% or 99% Confidence Interval threshold)

It’s not very different than selecting voters to poll.  The votes at the booth are the proof of the methodology, but the concepts are the same.

I’m not a statistician but I’ve had some course work and have worked closely with statisticians for decades.

ITW:

Agree. If you only have 78 in your sample, you had better be able to defend you sampling protocol pretty vigorously as well.

I agree that its not necessarily invalid just because of the small sample size.  It proves nothing but is interesting and worthy of further inquiry.  I wonder if the stats would hold true for children raised by homosexual men as well. 

One could theorize that Lesbian women who have children really want to have children, and parents who strongly desire children are far less likely to be abusive.

On the other hand it could be that the adolescents questioned in the study have had a lifetime of having to defend their parents and would minimize abusive behavior.  In short the kids could be defensive of their families, which could inject reporting bias.

What you both need is to see how the sample was selected.  That’s the key to sampling science.

There usually is a protocol and that has what’s called a selection criterion.  There are inclusion and exclusion criteria spelled out.  If done properly the protocol is peer-reviewed before actual selection begins.

A properly done sample can indicate a great deal from a moderately small sample.  15 is usually the least.

Wow. 78 households. Now how can you quarrel with those stats?

I’m pretty sure I can come up with 78 households run by a junkie dad and a hooker mom, who are too distracted to beat up their six kids who would fit that description, too.

Why are these people so desperate to prove a point?

This may be true or not…statistically significant results must drawn from
randomized samples of thousands, not dozens.

Most studies in the US are done to prove a hypothesis, not to test it. There is
usually money or a political motivation behind lab studies. That makes unbiased
scientific data hard to come by.