SURVEYS NOT MEASURING POLITICAL CONCEPTS:

risk-taking measured by objective tests that discourage guessing:
Cronbach 1946
Sherriffs and Boomer 1954

the possibility that DKs tap psychological states unrelated to the phenomena of interest:
Francis and Busch 1975
Rapoport 1979, 1982, 1985

Confident test-takers offering answers rather than admit they don't know:
Gritten and Johnson 1941
Sheriffs and Boomer 1954
Stanley and Hopkins 1972

individuals understand the 'same' question in vastly different ways: Brady 1985

Zaller is more social cognitive --- he suggests that most people have multiple considerations (facts, beliefs, to be considered) potentially relevant to most survey items.

What varies, then, is which considerations are accessible at the moment an attitude must be provided
Higgins & King 1981

"behavior is often driven by automatic mechanisms, which leads self-report of mental processes to be notoriously unreliable and susceptible to many forms of contamination":
Bem, 1967
Wilson & Brekke, 1994

a reader's own characteristics may determine his or her perceptions of the media's message
Becker and Kosicki 1995

Incorrect and "don't know" responses to political knowledge survey questions may not mean the same thing; results indicate that the two perform differently when disaggregated.
--MONDAK 2000

Aggregated public opinion results may be a poor reflection of true public sentiment. Specifically, individuals who harbor "unacceptable opinions" may do so with a 'don't know' response. Feldman and Zaller's view that individuals do not possess 'true attitudes' on issues is limited; "it does not account for response effects arising from the social nature of the interaction in the survey interview" (1210).
--BERINSKY 1999

The use of knowledge batteries in surveys can have consequences for subsequent analysis when 'don't know' responses are encouraged. Encouraging DKs produces data that underestimates the effect that knowledge has on a respondents opinions; corrections are possible but results are not as reliable as when DKs are not encouraged. Results indicate that knowledge has a more significant effect on opinion is indicated by surveys which encourage DK responses.
--MONDAK 2001

"The greater the demands a question places on memory, the less accruate the respondents' answers and, all else being equal, the less accurate the survey estimates derived from them" (62).
--TOURANGEAU, RIPS & RASPINSKI 2000

Responses are not static and can be manipulated to obtain the 'best' recall possible -- attention to question wording and increasing the number of cues available to help assist respondent in answering more accurately.
--TOURANGEAU, RIPS & RASPINSKI 2000

There are ways of manipulating survey questionnaires that increase the likelihood of more accurate responses.  If recall of the specific number of events is important, questions should direct the response of a shorter/closer time period than if information is more biographical.  Bounded recall -- individuals are better able to recall accurate information from past events if they are pormpted or 'cued' by significant events in the past. The significance is important in that question wording which best responds to these political cues can be enhanced by keeping these issues in mind.
--TOURANGEAU, RIPS & RASPINSKI 2000

Information necessary to answer questions of opinion or long-term memroy are most prone to variations, and that responses come not from one indication of preference, but instead as a culminationa of a range of preferences, which are processed to find the 'correct' a nswer on a survey questionnaire.
--FELDMAN 1995

survey questions are inherently noisy measures...'Better questions' will never eliminate this variability" (266).
--FELDMAN 1995

"Researchers have generally been content to use survey responses as direct indicators of unobserved psychological constructs, recognizing that survey questions may be affected by random and systematic measurement error, but only occasionally attempting to explicitly deal with the consequences of these errors for statistical estimation" (249).
--FELDMAN 1995

Attempting to use a "one size fits all" approach to analyzing opinion survey results is inherently incorrect. Given research on cognitive processes and how individuals retrieve and utilize information, the author argues that "response format can have an effect on the distribution of responses" (260).
--FELDMAN 1995

The idea that "knowledge per se" is not responsible for opinion formation/changes is the first start of a process whereby we can also conclude that exposure to media and the questions utilized to "measure" this exposure are not the best ways to measure knowledge levels.
--DRUCKMAN & NELSON 2003

Current survey correction methods do not account for DIF (subjective response variation) between respondents. Instead the authors recommend interpersonally comparable measurements using answers to vignette assessments that have actual levels of a variable that are the same for every respondent, thus adjusting for self-assessments that create the DIF. These are statistical corrections that systematically link related vignette-measured variables pairwise by respondent, essentially providing a unique "anchor point" for each respondent that can be transposed to multiple response sets. The authors explain: The perception of respondent about the level of the person described in vignette j is elicited by the investigator via a survey question with the same K1 ordinal categories as the first self-assessment question. Response consistency is thus formalized.
--KING, MURRAY, SALOMON 2003