This blog is about public opinion surveys.
Most public opinion surveys are like research on nutrition or health. They are inherently contradictory and there will soon be a more recent publication to contradict the latest findings.
I read that a recent survey by the CDC shows women in the 18-25 age group have gone from some lower number up to 14% who have had a lesbian experience. Men report a much lower number of same sex relations. From my personal experience and countless earlier studies, the number is much higher for both groups if experience starting at age 10 is included in the survey data. But that is irrelevant.
I think there has probably been an explosion in lesbian relations among women in the 18-25 age group for one reason:.....
.... an explosion in lesbian relations among women in the 18-25 age group for one reason: there is a word/phrase that has entered the language to describe it. The word/phrase is L.U.G. Lesbian until graduation.
Now let’s get to understanding survey data. Each question asked in a survey is a probe. For each individual or group of people with common views many probes are needed.
Why many probes? Think of probing an orange. The results from random probes would be 75% orange juice, 7% cell walls, 7% seeds, 7% skin pulp, 4% pungent skin surface. Only the whole structure tells you what an orange is about.
If all the probes are in the same area, or in several poorly chosen areas, you will get a wrong idea about the orange. Same is true with individuals and groups of people with common views. You need to know who you are sampling and the structure of their thought.
Let us use the sex sample given above. There is clearly one group of college women who are avoiding the problem of dealing with men while they are in college. Ask them if they are most comfortable talking about sex with women and you will get 98% yes. Are they physically comfortable with other woman naked or dressed 85% yes. Have they fallen asleep with another woman nearby in the past three months, 80% yes. Have they had mutual oral sex with another woman in the past 
year, 10% yes. Have they had hetero-sexual sex in the past three years, defined as genital penetration, 40% yes. Now what would a single question to members of that group, 'Have you had a sexual relation with a woman in the past year,' tell you? The answer is not predictable because the term sex is not clear (it was not clear to most young people even when Bill Clinton said it wasn't clear to him either.)
There are other groups of women in this age group who are married, some groups who are married with children, some divorced with children, some who have never had sex of any kind including masturbation. There is no meaningful way to describe or predict what their response to this question might be. Nor could the meaning of their answer be understood without many probes related to each group.
No possible single probe can give a sense of what is happening in the sexual lives of American women 18-25 unless the survey is done in a very different way than a national cross-sectional questionnaire with simple reporting of responses to one or two questions. Even a few demographic breakdowns doesn't add much.
A good survey is based on personal multi-hour face to face interactions with many small groups of fairly homogeneous people that generates a model of their thinking and value structure. Once that exists a larger composite can be created that is useful and meaningful.
A few people have done such work on particular issues; one good example was done by my friend Charlotte Mayerson Goin' to the Chapel, a study of the relations between success in marriage and pre-marriage dreams.
No studies other than Alfred Kinsey have been done on larger broader subjects.