Question 1

What is public opinion research?

Accepted Answer

Public opinion research has been defined in many ways, but these can be summarized as the collection of enough data from a defined population to be confident that you have valid and reliable facts about what and/or how they think. In particular, to distinguish it from market research, this population does not have to be defined by their past, current, or potential use of, purchase of, or influence on particular products or services. Typical methodologies include surveys conducted by phone, mail, intercept, web, or face-to-face methods, as well as focus groups and IDIs. The essential goal is to be able to extrapolate findings from the research sample to the more general real-world population of interest with confidence. Typical topics include preferences (e.g., “which candidate do you like best,” “do you favor the new health care bill”), attitudes or beliefs (e.g., “people generally get what they deserve,” “the country is headed in the wrong direction”), values (e.g., “social empathy,” “fiscal responsibility”), or intentions (e.g., “I will use the new park once it is built”). They may also include questions about knowledge (e.g., “who is your state representative,” “have you heard of Proposition 513”) or past behaviors (“how often have you been to the park”). Although past behavior is not an “opinion” in the strictest sense, it is sometimes considered to be an indication of what someone thinks.

Question 2

Why is public opinion research important?

Accepted Answer

Public opinion research is important because, without it, the decisions of leaders would be based solely on their personal preferences and values, the world-views of citizens would be mostly based on rumor and innuendo, and those without a voice would remain voiceless. If that sounds dramatic, consider the extent to which people are affected by consensus opinion. Whether their perception of the “consensus opinion” pressures them to conform or incites them to rebel, it makes a difference in how people think and in what they will do. The true consensus opinion – or the diversity of opinions – can only be known via public opinion research. In order to fulfill this noble mission, public opinion research must convince the audience that it has accurately measured opinions, that the opinions are projectable, and that they are reported in an unbiased way. This is why original reports of public opinion polls always include the sample size, details about the population, timeframe for the data collection, and the confidence interval around results. Others who cite the findings may leave these details out, so you will have to look up the original report or else trust these parties to have done the due diligence to ensure that the results accurately reflect the opinions of the target population. Without these details, doubts can be raised about the meaning of the research results. Perhaps no example of public opinion research demonstrates its importance as directly as political polling. Those of us who live in democracies are familiar with the so-called “push polls” that political parties and candidates sponsor which include leading questions and may even include obvious attempts at persuasion. These initiatives are intended to change public opinion, not to measure it. However, these parties and candidates also have to measure public opinion accurately via so-called scientific polls if they are to make effective campaign decisions because they need to know if a majority or plurality of likely voters will support them. For example, if they need to attract significant numbers of both affluent and disadvantaged voters, they need to have accurate data regarding how their positions and policies are received by each so they can recalibrate as needed. Ironically, however, their use of push polls may serve to undermine the credibility of scientific polls, exacerbating a challenge for legitimate public opinion researchers.

Question 3

How is public opinion research different from “market” research?

Accepted Answer

Public opinion research has many procedures and success criteria in common with “market” research, and the main differences concern how integral the methodology is to the reporting and the context around the work. Generally speaking, the methodologies used in public opinion research are also used in market research, both rely on random and representative samples, both require unbiased measurement instruments and data collection techniques, and both have similar ways to assess data quality and accuracy. In the reporting of public opinion research, however, the methodology tends to be a major point of emphasis while the audience for market research reports are often less interested in the “how” of the research. On one hand, market research is often conducted by private enterprises with their own researchers or suppliers who can be trusted once they have a track record or held accountable for any failures. On the other hand, end users of public opinion research are acutely concerned with how accurately the results reflect the state of mind of the population of interest and need to have the veracity proven out in each report. Of course, the results of market research also need to accurately reflect the target population, but the proof of this is not often the leading headline. Context is another key difference. Market research is ultimately concerned with resolving business issues related to a target population of current or potential customers while public opinion research is concerned with measuring the attitudes, beliefs, values, and/or behaviors of a population that is not necessarily defined by their relationship to certain product or service categories. Because of this difference, the unique perspectives of providers of public opinion research may offer differential value in comparison to market research firms. Although some of their work may involve issues that are also concerns of market research, public opinion researchers are likely to encounter a broader set of subjects that affect politics, public policy, and other non-marketing issues. They may have more knowledge in these subjects and more expertise in designing and executing research particular to them. Whereas the end users of market research may tend to be businesspeople, the end users of public policy research may come from other walks of life, and these researchers may communicate with them more effectively and fit better culturally than a typical market researcher. Also, in many cases, the deliverables from a market researcher might be business recommendations, while a public opinion researcher may be charged solely with providing facts to be debated and interpreted by others. Because of this difference in focus, a public opinion researcher may be more disciplined about maintaining the objectivity that is more thoroughly scrutinized in their arena. Although not wholly different from market research, public opinion research can benefit from experts who understand a broad range of issues that affect the public, the best ways to measure them, and how to deliver results that gain the trust of its end users. They may also have special databases that are more effective for public opinion research, and they may have analytical tools or models that overcome challenges or can complement the core research with other data.

Question 4

What is instrument bias in public opinion research?

Accepted Answer

Instrument bias, such as how a survey is written, means that the design of the instrument influences how the respondent answers beyond their pure opinions on the topic. As with any survey, the researcher must strive to ensure that each question and set of answer choices are understood the same way by everyone who answers them. Achieving this is, of course, impossible, because public opinion research often targets populations that are very diverse and are likely to interpret the questions differently. The same terms can have different meanings for different people, some terms might have no meaning for some people, and relevant knowledge varies from individual to individual. Attempts must be made to use the most common language possible, knowing that what is common for one person may not be common for another (including the researcher), especially where familiarity might vary due to factors such as language, regional colloquialisms, ethnic heritage, socio-economic position, occupation, and so on. Where resources permit, multiple instruments may help minimize this bias. In some cases, respondents cannot express their opinions accurately because they lack knowledge about a particular subject. For example, if you want to measure support for particular legislation, you need to consider whether the respondent has heard of it, what they have heard about it, and, possibly, any misconceptions they have. If you are measuring support and not awareness, you want them to respond based on a fair understanding of the legislation, and this may require that you provide some information. This is tricky because you need to translate into common language while revealing enough to capture the essential features without over-explaining. This can be problematic to achieve without introducing bias. The translation may be imprecise, it might introduce language that is pejorative or approving, and it might leave out elements that would prove to be important to the respondent. Of course, you could just offer the legislation verbatim as is done in some referendums, but the literal version might be very confusing to a non-lawyer. Public opinion research on subjects such as emerging technology may be difficult because the jargon is not yet widely used and the concepts may be unfamiliar to many, inscrutable for some, or require a lengthy prolog. The instrument can introduce bias in other ways. Double-barreled questions such as “do you believe doctors are competent and dedicated?” are common but uninterpretable because you don’t know which responses are driven by perceptions of “competence,” “dedication,” or both. Undefined terms can perplex some respondents or trigger inconsistent interpretations. Establishing a premise, such as “if the park is built,” may be necessary in order for the respondent to answer in the proper context, but other kinds of premises, such as “given the sad state of the industry,” may be completely biasing if the respondent does not agree with the premise or does not normally link it with the subsequent question. For closed-ended, multiple response questions, having an unbalanced scale, such as three “positive” responses but only two “negative” ones, can bias the responses in one direction or another, while asking attitudinal questions, such as “I like the trees in our neighborhood” with only “yes/no” options can result in inaccurate readings because, in reality, the respondent does not have a binary feeling about them. There are other types of instrument bias as well as interviewer bias, respondent bias, and sample bias, and an expert in public opinion research can help you to identify them, minimize them, and account for them in the results. Research always contains biases, but it is important to minimize them and to know how to account for them.

Top Public Opinion Research Companies

What are public opinion research companies?

Featured Experts in Top Public Opinion Research Companies

Compare Public Opinion Research Companies

EurekaFacts

EurekaFacts

Provoke Insights

Provoke Insights

The Harris Poll

The Harris Poll

Discovery Research Canada

Discovery Research Canada