re: Study: Same-Sex Parenting Harms Children

quote:

---

I'm curious if I would consider 20 to be a large enough sample size to truly consider it scientific, especially considering that only 3 were gay couples.

---

20 wouldn't be a sufficient sample for any single random variable randomly drawn from a population (e.g., flipping a coin 20 times). The rule of thumb from the Law of Large Numbers is a minimum of a sample size of 30 as a lower bound to approximate to the variable's probability distribution. And that is under those perfect conditions, for a single variable to approximate to the likelihood.

When you start talking about variability and statistical significance then the sample size will need to be much larger. Then considering imperfect sampling and modeling multiple known variables, then the need increases. And since they are likely to be a number of other variables that can't be controlled for, it needs to be even larger.

In other words, the sample size would need to he factors larger to even consider it a representation of population phenomenon and drawing conclusions.

In my opinion, this Study should never have been published in a peer-reviewed journal, even if it was studying something trivial, let alone something that is based on something that is already controversial, especially since it was the sole focus.