Thread: Political or Social Scientists

Do we have any of them in here? I could use some help.

Or if you're familiar with the Statistical Package for the Social Sciences (SPSS) program, I might be able to use you.

My class conducted an exit poll during the March primary here in Texas and we entered the data into SPSS. Now I'm running cross-tabulations of certain questions. My objective is to develop hypothesis and then test them using the data. I've got the raw data (everyone's info) then the data from just the republican primary, then the data from the democratic primary. Problem is that the number of African American respondents we got was around 10.4% of the total vote, which isn't congruent with the average of around 5.5-7%....I'm just brain-fogged and can't figure out the times it's imperative to use the weighted data and when not to.

Anyone have any experience with this?

Boy, it's been a loooooong time since I last used SPSS...

Ummm... let me see if I am understanding this, you didn't give too much information.

Of the people who responded to the exit poll (who voted in the Primary in Texas) 10.4% where African American. However, this is higher then the expected (historical?) number of around 5.5-7%.

So what you want to do is extrapolate the sample responses over the entire population of Texas or whatever. And you aren't sure when to use the 10.4% number and when to use the expected number. Right?

In a nutshell, you've got it.

Originally Posted by Rooster050

This just go's to show you how meaningless Polls actually are. Even using them to make a suggestion on who to vote for has failed.. The numbers will speak for them selves on the day of importance. Just like the surprise that Bush beat Kerry, The polls suggested otherwise.

Real political scientists don't use polls to predict a winner, but rather, to explain why someone won. Further, the reason our African American numbers were so high is because we simply polled them more than we did other ethnicities to ENSURE we got a representative sample. Weighting data is better than creating data.

ALERT!

Someone has tried to register using the IP Address 41.190.16.17 which is MATCHED IN THE RBL DATABASE of the proxies.dnsbl.sorbs.net RBL.

This registration attempt has been allowed and the account has been permanently banned.

Registration Details: Williamisksf ( cheapestwowgoldyai@wowhackgold.com )