94% of all people end up in a member of a "Pair Bond" (One male, one Female). This is considered the only accurate statistic because it does NOT depend on people telling the truth about themselves, all you have to do is see who is living with whom. Once you get away from this statistics you have to rely on people who have good reasons to lie. It may be ego (I had sex when I was 12 for people wanted me) to status (I was a virgin till I was married and NEVER thought about sec till them).
When Polls are taken, most pollster tend to be middle age women, for the Polls have found out people are least willing to LIE to their mothers (and middle age women are more like people mothers then young women). This brings with it its own set of "lies" i.e. will you tell you mother you had sex? or didn't have sex? if what your mother wanted to hear the later? Any study has to consider this, the Kinsey report had a problem, while its initial reports were "sexy" its final observation was boring (Most people had sex only with their spouse). Thus the final report was not issued till the 1990s even through the date had been collected by 1972 (Through as early as 1953 it was concluded that you could NOT have a true random sample, thus the Kinsey report is good for what it is, but the statistics used in it and derived from it are unreliable).
The number of Homosexuals is the classic problem. It is a common belief that the Kinsey Report found 10% of people are Homosexuals, but that goes against another study that found that 94% of people are in paired bonds (Through just because you are in a Paired bond does not mean one member is also NOT having sex with people of the same sex, but such actions do NOT explain the 4 percentage points problem). Kinsey actual report found 34% of men had had homosexual relations. This is questionable, through Kinsey tried to make his report as reliable as possible, even by the mid 1950s it was found to be unreliable. Efforts to correct this were made, but the corrections could NOT solve the problem of the inability to get a true random sample. Thus statisticians dismiss Kinsey's numbers (Through the overall report is considered good, but you can NOT say the numbers he used are accurate).
The reason is that Kinsey started with a bias, having a huge number of his surveyed people being prisoners or male prostitutes, when these are removed, the number seems to hold up, but again the remaining people surveyed were also NOT random (i.e. with the same bias but not listed). This bias was caused by the fact most people do NOT talk about their sex life. The question for Kinsey and other people who study sex is how do you get a random sample when people feel uncomfortable about talking about the subject with a total stranger? If you can get random people to talk about it at all will often lie, both positive and negative depending on what they believe the survey taker wants to hear. Given these two sets of "lies" (Refusal to answer, or if answer give what they believe you want to hear) there is no way to get a random sample of people, and the results are unreliable. The basic problem is there as there is no way to determine who is lying (or what people who refused to say, would say) and thus no way to get reliable data.
All of this Leads to the question is what questions were asked and by whom? What is the error rate? How do the poll taker determine if someone is lying? If the Poll taker says someone is lying, do they mark the opposite answer or just go to the next subject? If you go to the next subject are you excluding people who would give opposite results to what you are finding?
The second set of problems is are people using the same definitions of Sex? If the poll taker understood that to mean intercourse but the person being polled understands that to mean including just flirting with someone of the opposite sex, does that make the result valid? Sexual banter, dancing, kissing, holding hands, going out on the town together etc is even encouraged among the most conservative Religious leaders in this country. It is part of being Human. The line drawn is actual sexual intercourse. If this survey ignored this problem of definitions, the results are invalid IF THE RESULTS SAY TEENS SHOULD HAVE INTERCOURSE, but may be valid that teens should look at each other as sexual beings and treat each other as sexual beings but stopping one step short of intercourse.
One last comment, this is about twins, yet the article did NOT mention any difference between twins. Did one twin have early sex and the other did not? If true, why? If they were in the same household then the difference must be the sex. But such Twin to Twin comparison is NOT mentioned in the article, it is a overview of the whole group. If one sets of twins turn out "all right" and another set of twins do not, that is NOT a valid twin to twin test. In a Twin test you compare a twin that turned out "bad" to one that turned out "good". That is a valid Twin to Twin test (and more valid if it is consistent among most of the twins). My problem such results are NOT made, the results mentioned is only as a group (i.e. it may be one set of twins had sex early and turned out good, another set had sex late and turned out badly, the problem that is NOT a Twin to Twin test but a survey of teens, whose bad or good behavior may be caused by something else in their lives.
I am just pointing out three problems with this study, a study I suspect was influenced by what the 24 year old wanted to find. She then felled into a Trap common to people who wants to find a result, you find want you want to find. The results may be valid, or may not be, but the above three problems must be addressed, and I do not think there can be given people's attitude to sex (i.e. most people do NOT brag about their sex lives), various definitions of Sex (Which I may add from the Article seems to include the kissing, holding hands, dancing etc more than actual sexual intercourse) and that this seems to have been a twin study that did NOT compare twins to twins.
On the Kinsey Report:
http://en.wikipedia.org/wiki/Kinsey_reporthttp://www.jackinworld.com/library/articles/kinsey.htmlhttp://www.kinseyinstitute.org/