Contingent valuation is a survey-based economic method used to estimate the value of goods or services (often non-market goods) such as environmental benefits or public goods. The approach works by asking individuals to state their willingness to pay (WTP) for specific services or their willingness to accept (WTA) compensation for their loss. This method is particularly useful for valuing resources that do not have a market price, as it creates a hypothetical market scenario where respondents can express their preferences and valuations directly. In healthcare, this can also be helpful for understanding willingness to pay for hypothetical treatments or for treatments where actual payment is largely covered by insurance.
However, contingent valuation approaches also have a number of potential biases researchers need to address. Some common biases to consider include:
- Hypothetical Bias. Hypothetical bias occurs when there is a discrepancy between what people say they are willing to pay in a hypothetical scenario versus what they would actually pay in a real-world situation. This bias often leads to overstatement of willingness to pay (WTP) values, as respondents may not face actual financial constraints in a survey setting.
- Anchoring or Starting-Point Bias. Anchoring bias arises when respondents’ WTP is influenced by the initial value or range of values presented in the survey. For example, if a survey suggests a starting bid, respondents may anchor their valuation around this figure, leading to skewed results. This can be particularly problematic in double-bounded dichotomous choice formats, where initial bids can heavily influence responses.
- Information Bias. Information bias occurs when the information provided in the survey influences respondents’ valuations. If the information is misleading or incomplete, it can lead to inaccurate WTP estimates. This bias highlights the importance of providing balanced and comprehensive information to respondents.
- Opportunity Cost Bias. Opportunity cost bias can occur if respondents do not fully consider the trade-offs involved in their stated WTP. They might not account for the opportunity cost of spending money on the good being valued instead of other goods or services, leading to inflated valuations.
- Part-Whole Bias: Part-whole bias, also known as embedding effect, occurs when respondents’ valuations of a part of a good are not consistent with their valuation of the whole good. This can lead to insensitivity to scope, where the WTP does not change proportionately with the size or scope of the good being valued.
- Payment Vehicle Bias. Payment vehicle bias arises when the method of payment suggested in the survey (e.g., taxes, donations) influences respondents’ WTP. Different payment vehicles can evoke different emotional or cognitive responses, affecting the valuation.
- Strategic Bias. Strategic bias occurs when respondents deliberately misstate their WTP to influence the outcome of the survey. For example, they might understate their WTP to avoid paying higher taxes or fees if they believe their response could lead to the implementation of the proposed policy or project. Conversely, they might overstate their WTP to ensure that a valued good or service is provided.
- Embedding Effect. The embedding effect (or part-whole bias) occurs when respondents give similar WTP amounts for a specific part of a good or service as they would for the whole bundle. For instance, respondents might state a similar WTP for providing insurance coverage of one rare disease as they would for coverage of all rare diseases. This can result in an overestimation of WTP for treatments for specific diseases for the part or an underestimation for the whole.
- Framing Effect. The framing effect is when the way a question is worded or presented influences the respondents’ answers. For example, presenting a scenario as a gain (e.g., preserving a natural resource) might lead to different WTP responses than presenting it as a loss (e.g., losing a natural resource).
- Scope Insensitivity. Scope insensitivity occurs when respondents’ WTP does not significantly change with the scale or scope of the good or service being valued. For example, there might be a new cancer treatment that improves survival by 12 months and WTP may be similar to another treatment that increases survival by 24 months. Respondents are not fully accounting for the extent of the benefit, leading to potential inaccuracies in the valuation.
- Non-Response Bias. Non-response bias arises when certain groups of people are less likely to respond to the survey, potentially leading to a sample that is not representative of the broader population. If, for example, higher-income individuals are more likely to respond, average WTP may be overestimated; if lower income individuals are more likely to respond, average WTP may be underestimated.
If you are designing a contingent valuation study or any stated preference survey, these are biases to take into account and try to address with survey design and administration.
