# Kinda confusing Stats Problem

#### Juice Box

##### Diamond Member
CNN/Time conducted a nation-wide survey on the legalization of marijuana last October. A sample of 1007 adults, 18 and older , were selected by random digit dialing. 34% of the sample said they favored legalization of marijuana. See http://www.pollingreport.com/drugs.htm

Q: 34% of the sample believe marijuana should be legalized. Assuming the sample is a simple random sample, what is the standard error of the percentage of the population that believe marijuana should be legalized?

I know you need the SD to find the standard error....but I dont know where to find the SD...

#### Soccer55

##### Golden Member
The first thing I can think of is if maybe you need to look at using the normal distribution to approximate the binomial distribution. After a quick glance, I think this poll will meet the criteria for being a binomial experiment (where your 2 categories are "favor legalization" or "not favor legalization"). Note that I said "not favor" instead of "against" since "not favor" could include the few people that answered "not sure".

Anyway, if this does meet the criteria for a binomial experiment, you could probably use the normal distribution as an approximation, in which case, the SD for the binomial distribution would be the same as the SD you're looking for.

Hopefully this is in line with the type of stuff you're doing in your stats class right now. This could be wrong, but it's the first thing I thought of.

-Tom

#### Syringer

##### Lifer
Yeah..isn't SD n*p*(1-p) or something..which would be 225?

#### Random Variable

##### Lifer
You need to create a confidence interval for a proportion. There's no one correct answer.

#### Random Variable

##### Lifer
Originally posted by: Syringer
Yeah..isn't SD n*p*(1-p) or something..which would be 225?

I don't think he's talking the standard deviaton of the data.

EDIT: You forgot to take the square root, BTW.