.
"If George W. Bush runs for re-election in 2004, in general are you more likely to vote for Bush, for the Democratic Party's candidate, other, or not sure?".
Bush Democrat Other Not Sure .
% % % % .
11/01 27 20 6 47 .
In order to better compare my sample to national statistics, I also noted the gender and age group of each respondent, inquired about his/her political identification, and asked how often he/she votes. .
Why my results differ from CNN/USA Today/Gallup poll's results:.
Target population/Attempt to Randomize.
My target population included all eligible voters, and I attempted to randomize by simply interviewing the general public. In spite of this, certain groups were inaccurately represented.
Those in the 56-65 age group constitute for 11 percent of all eligible voters nationwide, and those in the 65+ age group constitute for 12 percent of all eligible voters nationwide, but, by coincidence, my poll did not include anyone in these two age groups. In addition, the 26-35 and the 36-45 age groups were under-represented, while the18-25 and 46-55 age groups were over-represented.
Females were over-represented, constituting for 67 percent of my survey while only 51 percent of eligible voters nationwide are female. Inversely, males were under-represented, constituting for 33 percent of my survey while 49 percent of the nation's eligible voters are male.
Reasons for misrepresentation of some groups may be due to the location of where I conducted my poll "a small Wisconsin town of primarily Caucasian, middle to upper-middle class residents. In addition, my sample was volunteer-based, and certain groups of people were generally are likely to volunteer than others.
.
Question Wording.
Question wording plays an important role in survey research. Questions should be valid and reliable, and should avoid factors such as bias, middle alternatives, double barrelism, and double negatives.