POSC318 Public Opinion EXAM 1

The Electoral Connection

Elections are mechanisms for holding elected officials accountable because voters can evaluate elected officials and respond to their performance through elections

Classical Democratic Theory

A theory of democracy emphasizing citizen participation

Democratic Elitism

A theory of democracy in which political decision-making is concentrated among elected elites who are trusted to determine and pursue the common good

Elite Leadership

A system in which elected officials and political elites have substantial decision-making power because citizens delegate authority to them through elections

Trustee Model of Representation

A model in which elected officials use their own judgment to make decisions they believe are in the public interest

Delegate Model of Representation

A model in which elected officials act as representatives or mouthpieces for their constituents and primarily follow what their district wants

Trustee vs. Delegate

Trustees rely on their own judgment about the public interest

Democratic Pluralism

A theory in which government is influenced by many competing groups and interests rather than by one unified public will

Participatory Democracy

A system of governance in which citizens have direct and ongoing influence over political decisions and policies rather than participating only through periodic elections

Qualitative Research

Research that uses primarily non-numerical data to understand people's experiences

Qualitative Research Examples

in-depth interviews, open-ended questions, data are textual, thematic analysis, inductive / exploratory

Open-Ended Questions

Questions that allow respondents to answer in their own words rather than selecting from predetermined response choices

Non-Probability Sampling

A sampling approach in which participants are not selected through random probability procedures

Qualitative Research Disadvantages

can be time-consuming and labor intensive

Quantitative Research

Research that uses numerical data

Logic of sampling from a population 

If the subset or sample is representative of the population, observations can generalize to that population

Probability Sampling Advantages

tend to produce representative samples and allow researchers to quantify uncertainty in estimates

Probability Sampling Disadvantages

can be expensive

Non-Probability-Based Approaches

Sampling approaches in which participants are selected without random probability procedures

Non-Probability Sampling Advantages

often more feasible, fast, and budget friendly

Non-Probability Sampling Disadvantages

Participants may differ systematically from people who are not selected, limited generalizability, inability to quantify uncertainty

Convenience Sample

A non-probability sample consisting of people who are easy for researchers to access (

LIRR Interview

An example of convenience sampling in which researchers interview people encountered on the Long Island Rail Road rather than randomly selecting from the entire population

Self-Selected Sample

A sample in which people decide for themselves whether to participate

Snowball Sample

A non-probability sampling method in which researchers find one or two eligible participants and ask them to recommend other participants

Quota Sampling

A non-probability method in which researchers divide a population into subgroups and recruit a predetermined number of participants from each subgroup without randomly selecting them

Expert Knowledge Sampling

Selecting participants based on researchers' or experts' knowledge about which people are particularly relevant to the study

Systematic Differences in Non-Probability Samples

People who enter a non-probability sample may differ in consistent ways from people who are not included

Sampling Error

The difference between a sample's results and the true characteristics or opinions of the population

Random Error

Unpredictable or idiosyncratic errors that tend to cancel one another out when observations are aggregated and therefore usually have modest effects on population estimates

Systematic Error

Consistent errors that push results in a particular direction and can create significant bias in estimates of population characteristics

Random Sampling

A process in which each person in the population has a known chance of being selected

Coverage Error

Error caused when the sampling frame is incomplete and does not adequately cover the target population

Sampling Frame

The list or source used to identify members of the population who can potentially be selected for a study

Population-Sampling Frame Gap

The difference between the entire target population and the portion of that population represented in the sampling frame

Sampling Bias

Systematic differences between the people included in a sample and the larger population the researcher wants to study

Sampling Error vs. Sampling Bias

Error is random while bias is systematic

Nonresponse Error

Error that occurs when people selected for a study do not participate and those who respond differ systematically from those who do not

Reducing Nonresponse

Researchers can use follow-ups

Strategies for Reducing Sampling Error

Researchers can improve the sampling frame

Guidelines for Writing Good Survey Questions

A good question should produce a truthful and accurate answer without influencing the respondent and should be straightforward

Closed-Ended Advantages

The questions are quick to administer

Closed-Ended Disadvantages

The available response options may fail to represent a respondent's actual opinion or may force respondents into an inaccurate category

Open-Ended Advantages

allow respondents to express opinions that researchers may not have anticipated and provide more detailed responses

Open-Ended Disadvantages

labor intensive to code and interpret and are more difficult to convert into a simple aggregated result

Measurement Error

The difference between the concept researchers intend to measure and what their survey or measurement actually captures

Question Wording Effects

The specific wording of a question can influence how respondents interpret the issue and therefore change their answers

Social Desirability

The tendency for respondents to give answers they believe are socially acceptable rather than reporting their actual attitudes or behaviors

Question Order Effects

Earlier questions can prime certain considerations and change how respondents answer questions that follow

Questionnaire as a Whole

Researchers should evaluate the entire survey because individual questions can influence responses to questions that appear later

Salient Events

Recent or highly important events can temporarily affect people's opinions and therefore influence survey responses

Aid to the Needy vs. Welfare Programs

Different labels for similar policies can produce different public reactions because wording influences how respondents understand an issue

List Experiment

An indirect survey technique used to measure sensitive attitudes by comparing responses to lists with and without a sensitive item

Unobtrusive Measure

A measurement in which the research purpose or specific behavior being measured is less obvious to participants

Nonreactive Measure

A measurement in which participants do not change their behavior or answers because they know what researchers are trying to measure

Expressive Responding

Respondents give answers that exaggerate or misrepresent their actual beliefs in order to express identity

Group Signaling

Respondents use survey choices to support their preferred "in-group" or show hostility toward an "out-group” (aka partisan cheerleading)

Psychological Comfort

People do this for internal emotional or psychological gain rather than reporting what they truly know or believe

Partisan Inflation

It makes public opinion polls look much more divided or partisan than people's actual underlying views

Multi-item scales 

uses a group of related survey questions or statements to score a single abstract concept or latent construct

Cognitive Interviewing

The participant narrates their thought process without interruptions, probing techniques

Quality Controls to avoid measurement error

Interviewer training, early launching, cleaning (malicious responding/ bots), good translations

sensitivity analysis

tests how independent changes in your assumptions, models, or data structures impact your final results

Margin of sampling error

the amount of wiggle room or uncertainty in a survey's results. It tells you how close your survey's findings are likely to be to the real truth about the entire population

Self-selection bias

when individuals choose for themselves whether to join a study or group

Best practices for journalists writing about survey results

• Be clear about who was surveyed 

• When the survey was conducted 

• Stay faithful to question wording

• Pay attention to margins of error

• Good to provide context not causality 

Public comment periods 

Cheaper but not generalizable, self-selection bias

exit polling

A survey of voters leaving their polling place on election day

• Two stage process – cluster sample

• Voting locations (precincts) are selected at random

Disadvantages of poll aggregation

Advantages of poll aggregation

Poll aggregation

A collection of polls from many sources

Non-attitudes

People providing an opinion when they don’t really have one