Wisdom of crowd or wisdom of the majority according to Wikipedia expresses the notion that the collective opinion of a diverse and independent group of individuals rather than that of a single expert yields the best judgement. Wisdom of crowds does not mean crowds are more intelligent than experts, rather it works best when the crowds are diverse, well informed and able to give independenct opinion without being influenced or biased.

For example, an expert may predict that a new feature will attract 20,000 users based on the industry experience. However, if a large number of diverse potential customers independently estimate their likelihood of using the new feature, the combined estimate may provide a more accurate estimate of the actual usage.

Why Do Group Estimates Work?

Group estimates can be useful because individual errors often balance each other out. When many people independently give an estimate, some estimates may be too high while some may be too low. When these estimates are combined, both estimates balance each other out, making the final result closer to what is expected.

• Different perspectives: People make decisions drawing from different views, experiences, knowledge and information.
• Large numbers: The more people involved, the more estimates available to create a stable overall result.
• Diverse knowledge: People may pay more attention to details to things that others do not or might have overlooked.
• Error cancelation: The high and low estimate can balance out each other when combined.
• Independent judgement: When each individual within a group makes their estimate without influencing one another, their mistakes are less likely to be repeated.

How Averaging Survey Guesses Improves Accuracy

Averaging accuracy guesses improves accuracy because it combines the estimates of many respondents instead of relying on just an individual estimate. Individual estimates are often affected by limited knowledge but when many independent estimates are combined, some of the errors balance each other out.

• More responses: A large number of estimates provides more information.
• Different perspectives: Respondents draw from their different experiences, knowledge and information when making their estimates.
• Balanced errors: High and low estimates help balance each other when combined, ruling out errors.
• Less reliance on one person: Individual errors have less effect when combined with many other responses.
• Independent responses: People are more likely to contribute their true opinion when they make their estimate independently rather than imitating others.

When Crowds Beat Experts

Crowds beat experts when the estimation involves prediction, estimation or uncertainty and there is no adequate information available. A diverse group brings together different information, views and experiences, making the combined estimate more reliable than an expert’s opinion.

• The crowd is diverse: People tend to make their decisions based on their views which are built through different views and information. Different viewpoints, experiences and backgrounds can help reduce the risk of everyone repeating the same error.
• Independent responses: People make their estimate without being influenced by other respondents or expert responses or views.
• People have different information: People do not drink from the same fountain of knowledge or pay attention to things at the same rate, each respondent may know or notice something others do not know.
• The expert has limited information: No one is an island of knowledge. Even an expert with many years of experience may not have access to all the information available across a large group.

The Role of Diversity and Independence

The importance of diversity and independence cannot be overlooked, in fact they are two crucial conditions that make the wisdom of crowds work. The estimate is more likely to be accurate when the group is diverse, it makes the judgement reliable without being influenced by one another.

Diversity means it is not one man’s view but a collection of different experiences, knowledge, viewpoints and background. It means the question is approached differently and they are less likely to make the same mistake. Independence means the respondents make their judgement by themselves without imitating other people.

When diverse people make independent judgement, their individual errors will likely balance each other out when the responses are combined. This is what allows and makes the combined estimate sometimes more accurate than any individual response, even over that of an expert. Researchers should not just be focused on getting many responses rather should create a room where respondents can provide diverse and independent opinions.

When the Wisdom of Crowds Fails

The wisdom of crowds does fail when diversity and independence is not included. When these two conditions are missing, the crowds tend to repeat the same error and end up producing an invalid and misleading result. Wisdom of crowd can fail due to the following reasons listed below:

• Lack of diversity: If respondents have shared experiences, backgrounds or knowledge, they may end up making the same mistakes.
• Groupthink: Respondents may follow the opinion of others instead of sticking to their own true opinion or judgement.
• Social influence: Seeing or having an idea of another person’s answer can cause respondents to alter their response, reducing the independence of their opinion.
• Poorly designed questions: Unclear, leading or loaded questions can tilt many respondents towards the same incorrect interpretation.
• Biased samples: If the people included in the survey does not fully represent the target population, the combined estimates may end up being inaccurate and produce misleading results.
• Shared misinformation: If a large part of the respondents relies on the same inaccurate information or interpretation of the questions, combined estimates will not balance out the errors.

Using the Wisdom of Crowds in Surveys and Market Research

The wisdom of crowds can be quite useful for researchers in their survey and for businesses in their market research because a combination of the estimates of many respondents would help produce reliable and valid insights. Instead of relying on just one expert estimate, researchers can collect independent diverse estimates and look for the average pattern to get the result.

• Estimating quantities: Respondents can estimate such things as how much they may spend on a product, use a product or service.
• Forecasting outcomes: Survey participants can be asked to predict things like how well a product will perform, what customer wants and how much change it will bring to the market. Their responses can help predict the outcome.
• Predicting demand: Researchers can ask people or potential customers how they are likely to purchase a product or use a service. Combination of their responses can help estimate future demand.
• Understanding consumer preferences: Every consumer choice is not always the same, consumer A may prefer previous products to the updated one. Bringing diverse independent estimates together can help reveal what consumers want, do not want or value.
• Testing new ideas: To avoid incurring big risk or loss, businesses can ask a group to rate or predict the potential success of a product or service before investing deeply into it.

Best Practices for Getting Better Crowd Estimates

To get a reliable and valid result using the wisdom of crowds, using a large number of participants is not enough; rather researchers should design the survey in a way that it encourages diverse, independent and true responses.

• Use a diverse group: Ensure to include people with different backgrounds, experiences and levels so as to avoid getting the same response.
• Ask clear questions: Avoid using ambiguous, leading or loaded phrases that could cause them to tilt towards a particular response or interpret the questions differently.
• Keep respondents independent: Allow respondents to give their responses without being influenced by other participants’ responses.
• Collect enough responses: A large number responses makes the combined estimates more stable and reduces the influence of unusual estimates or guesses.
• Use a representative sample: Ensure the people participating in the survey reflect the demographic being studied. A large but unrepresentative ends up producing a biased and unreliable result.
• Choose the right aggregation method: Use an average, median or more appropriate method to combine the estimates. A median can be helpful when a few very high or low responses might throw off the average.
• Compare with other evidence: Use the crowd as one source of information, then compare it with expert opinions or actual data to see how reliable and valid it is.

Conclusion

The wisdom of crowds proves useful when there are diverse independent responses or opinions, when these two important factors are missing, getting a reliable estimate becomes hard. By combining survey estimates, individual errors balanced each other out, producing a reliable and valid result. Researchers should not just be focused on getting a large crowd only but ensure they are a diverse group that represent the demographic being studied, when this is achieved, getting a reliable and valid data becomes an easy peasy.