A politician running for office called random phone numbers to see what proportion of the vote he could anticipate getting. From his sample of responses, he estimated that he would get 47% ± 4% of the vote with 90% confidence. When calling, some people refused to answer his question about whether they plan to vote for him, some people never answered the phone (even after repeated calls), some people do not have landlines (or are unlisted), some people may not have told the truth since the politician himself talked to them, and some were not sure yet whom they plan to vote for.
What biases are present that cannot be accounted for through margin of error? Check all that apply.
voluntary response bias
undercoverage
response bias
nonresponse bias
sampling variation