Surveys - How Accurate Are They - How Conclusive

Surveys - How Accurate Are They - How Conclusive

There are so many things that can go wrong or right with surveys.  And so many ways that surveys can be used beyond their scope.

1.  how many people are questioned - from how many countries
2.  what age range, gender, social standing in society are included
3.  what locale are they from
4.  what are the parameters of the questions, are they very general or very precise

All of those issues and more determine how the study can be used. These parameters determine what kind of conclusions are reasonable to draw from the questionnaires.  What might be a correct analysis for one gender, one age group, in one social level, in one city might not reflect accurately the other gender, or other age ranges, in other social standings, in other cities.

If that isn’t enough to question how people use surveys, the questions themselves are of great importance.  For instance, when spouses are asked if they have ever hit their partner, are they also asked when it is in self defense, when it is in playfulness and not harmful, or when it is not instigated by their spouse and out of anger, and when it results in harm (serious or mild)?  Most of the time, no,  these clarifying questions are not asked.  Which to the wise, enormously limits the conclusions.

Here is a survey of Boston youth from a particular area. 1000 youth were questioned.

http://www.boston.com/news/local/massachusetts/articles/2005/12/27/majority_of_boston_public_school_students_witnessed_violence/

What use is it?  Well, it does inform that violence is on the rise in that area.

Here:
http://www.sheridanhill.com/batteredmen.html

A writer claims that because of a study done by Strauss and Gelles, in which the parameters of the Survey were NOT spelled out, The National Family Violent Council concludes that "The fact that women had higher mean and median rates for severe violence suggests that female aggression is not merely a response to male aggression.”

But how do we know that if we do not know the nature of the questions?

Furthermore, where were the people from that filled out the survey?  What genders, what social standing?  And how do we know that they were answered truthfully?  It does not surprise me at all that men are in a clamor to have women viewed as equally abusive as they are in order to diminish the importance of men’s violence that has been going on for eons.

How does one validate a claim that women supposedly are equally as violent as men when we see the abuse that we see men promoting world wide.  Balance American men’s claims that women are just as violent with the situation of women in Africa, India, and Moslem countries, for instance.

http://www.nytimes.com/2005/12/30/international/africa/30africa.html?th&emc=th

“Women's rights legislation has also been enacted. Swaziland's new constitution, adopted this year, makes women the legal equals of men, able to own property, sign contracts and obtain loans without the sponsorship of a man. Zimbabwe this year allowed women to inherit property from their husbands and fathers. Liberia passed a stiff statute against rape, and president-elect Ellen Johnson-Sirleaf, the first woman in modern Africa to be elected a head of state, pledged to enforce it.”

Does that sound like women are as equally violent as men in Africa?  And when you think about it, American women were going through this same fight only upwards of 200-150 years ago.

So, I ask again, just how accurate are all these surveys?  And even if they are accurate for one area, one city, one state, what makes anyone think they are accurate for all of womankind?

TL

Comments

[Maureen]

I work for Statistics Canada - recognized world-wide as the 'cadillac' statistical agency, bar none. Everything from the wording of the questionnaire, to the sample selection, the training of the interviewer, the data capture and finally the analysis, can introduce a bias if precautionary measures aren't taken. Hence the qualifier "these statistics are accurate within x.x%". But, in some cases in the end, between you and me, torture numbers long enough and they'll say anything you want them to. Maureen

[Galina]

I'm hopelessly cynical about most "statistical results". People with an agenda have ways of making their statistics line up with their preconceptions.