But in this case, if I had to predict, I would say if the sample were expanded to be representative of the general population then the conclusions wouldn’t change much.

But then again one can be a non tatooed, non pierced, bald and dumpy guy in his 50’s and still “be prone to deviant behaviour, which includes excess drinking, smoking pot, (fantasise about) being promiscuous and being willing to cheat.” – especially at Monopoly :)

Jay

In the case of the two motorcycle gangs, I would predict that the number of tatoos or piercings would not correlate with drug use or criminal records. This would be because of sample error. Both groups diviate from the normal population in a significant degree, but share certain sub-cultural behaviors, such as tatooing and body modification, clothing and recreational transportation, but not others, such as criminal behavior and drug use. In that regard, college students are a sample much closer to the general population than the two motorcycle clubs.

College psych experiments and surveys often contain errors in methodology and samples because they are learning exercises for the psychology graduate students. I remeber my days a social pyschology grad student, running an experiment for my advisor. It was a opinion change study. One of the things I had to do was write the persuasive communication which would effect the changes we were measuring. I did pretty good at writing the script; too good. All the subjects who read it were convinced, no matter which experimental or control group they were in! The professor and the grad assistants who continued the experiment had to water down my arguments three times before they could get a significant difference between experiment and control groups.

Freemann

You are correct anecdotal evidence is always suspect. Doesn’t mean it is wrong though. Let me phrase it a different way.

If you gave the survey to a christian motorcycle gang (which my friend belonged to) and hard core a criminal motorcycle gang, I predict you would get different results. I could be wrong but I doubt it. Would I use either sample to predict the population at large? Probably not, because I don’t think either sample is indicative of the general population.

Which I think is the crux of the matter. What was the sample distribution in the survey? I do think it is difficult to infer general behavior from a narrow sample of college students.

Doesn’t mean the survey was wrong, but more broader samples are needed. My professors taught me that sampling is where most studies fail. The guys who predicted Dewey would beat Truman made correct inferences from their data. They just took poor samples. My experience with college psych experiment/surveys is that they do a poor job or writing the surveys, giving the surveys, taking samples and post analyzing the sample data.

You also made a good point about how the term deviant is used. I think there is a negative connotation to the word. As you pointed out it just means so many standard deviations from the mean (assuming a normal distribution). It does not necessarily mean they are good or bad people, although I think people do infer that.

Mike

By the definitions of social science, in a bipolar distribution of does/doesn’t, if 1/3 does, and 2/3 doesn’t, the doesn’ts are the norm, and the does’s behavior deviates from the norm. I’m sorry if this seems judgemental to you, but that is the language of sociology and social psychology.

Moreover, if a second measure is compared to the first , say the number of a specific thing the subject has, and we find that posession of zero things is associated with a ratio of 20% does to 80% doesn’t, while possesion of 5 things is associated with 1/2 does and 1/2 doesn’t, then we can say, to a certain level of confidence, that possession of those things is correlated with the deviant behavior. It doesn’t matter what the ‘things’ are or what the does/doesn’t behavior is. Nor is crrelation causation.

And Mike, my professors were always warning me, anecdotal evidence proves nothing.

Freemann

There are countless examples where surveys at colleges are later proven wrong. Like the previous post said college sophomores are not indicative of the population. A lot of the research is done by first or second year grad students with little experience. Also people fund research expecting certain results. If they don’t get the results they want they move on to someone who will. Finally they find errors in the statistics all the time. It has been my experience that psych majors suck at statistics.

I remember at my University they got a grant for $100k to see if good looking people have an edge in life. It turns out they do. Who would have thought!

My daughter is a tribal belly dancer. Tattoos are the norm. I have found these young women to be the nicest sweetest people in the world. People are constantly assuming certain behaviors because they are dancers and tatted up. People look for answers they want to believe.

The nicest kindest most religious guy I know has a pony tail half way down his back, wears leather, is tatted up all over his body. He doesn’t drink, smoke or use profanity. But I constantly see people pointing at him and talking about him. People never get to know him because he looks scary to them.

Mike

Uh, NOT using marijuana. Not breaking the law, or at least, not telling researchers from the school’s sociology department that one does break the law.

BTW, lots of research has shown that persons who commit behaviors that deviate from the norm, and are willing to admit it, tend to vastly overestimate the numbers and proportions of people who do likewise.

Freemann

We had a saying in psychology: University social/psychological research is valid, if you assume that white lab rats and college sophmores are representative of the population as a whole.

Freemann

I think you hit the nail on the head: did this study measure behavior – or merely the willingness to admit to certain behaviors? It should come as no surprise that a young adult with more than the usual number of tattoos or piercings might be more willing to admit to certain behaviors than those whose appearance is more conformist.

BTW: categorizing marijuana use among college students as “deviant” make you wonder what the researcher considers “normal”