Survey: Americans Believe Video Games Cause Violence

Actually for a population of 313,914,040 you can get a confidence level of 95% and a confidence interval of 2.1% with a survey size of 2,178.

That means you can be 95% sure that between 55.9% and 60.1% of all Americans think video games cause violence.

You can get a 99% confidence level and a 2.8% confidence level with a survey size of 2,123. i.e. you can be 99% confident that between 55.2% and 60.8% of all Americans would think games cause violence.

The sample size of 2,278 is higher than either of those figures, so it should be reasonable to conclude this survey represents American opinion fairly closely.

This is all dependant on the sample being representative of the total population of course. That means there should be the same mix of ethnicity, education, wealth, etc in the sample that there is in the total population.

So maybe not so far out after all?

@DrkStr And then you looked at the much larger polls conducted all campaign season long and realize that polls are garbage--only a pollster and a statistician that lives and dies by percentages could love them.

(Also: duh? Americans are told *every day* for months by pols that games are bad bad bad, and imagine that--a bunch of people believe they're bad.)

@Ergo I'm not saying they're right, I'm saying the sample size is plenty big enough to draw the conclusions that the report does without taking a mocking tone from journalists.

No. The sample size in too small to make that assumption. You need to know the exact question asked, the area it was taken and the relation to events such as the various massacres in the US. Love the stats by the way, as a a Doctor of physics at University of Leeds, England, I think you have just made stuff up. Where's your calculation of standard deviation for instance. What is a confidence interval? It is not a statistical term. Glad to unmask you as a bullshitter. If you want to talk math as the yanks call it I am waiting. Would be interested in why you made stuff up apart from trying to be clever. I am not interested in that motive.

and no, confidence interval is not a valid statistical index before you harp on. It is a parochial and convenient disconnection between cause and effect in data collection. You might as well point to extreme parts of a Gaussian distribution and say they were cogent to your argument.