Naturalness Concerns And Clean Meat Acceptance: A Faunalytics Study

Background

Clean meat (also called cultured meat, in vitro meat, etc.) is grown from animal cells in a clean facility without the need to slaughter an animal. It is not yet commercially available, but several companies are working toward having products in stores in the next few years.

Meanwhile, consumer researchers are trying to gauge the likely reaction to clean meat when it comes to market. So far, reactions have been quite mixed. Some people are very excited about it, but others are extremely reluctant. Previously observed rates of willingness to try clean meat have ranged from 16% to 65%.

As we described in a previous post, one key concern is the “ick factor”—the idea that clean meat is unnatural and therefore bad to eat. This concern is an important one to address because similar consumer concerns likely contributed to policies restricting the cultivation of genetically modified (GM) foods in Western Europe.

The goal of this study was to find ways of describing clean meat that could address naturalness concerns and increase acceptance of this new product. Participants read one of three messages intended to address those concerns or a control message similar to those currently in use. They then answered questions about their acceptance of clean meat: willingness to try it, beliefs about it, emotional reaction to it, willingness to pay for it, and more.

We looked at whether different messages produced more or less acceptance of clean meat, and at overall rates of acceptance in the study relative to previous studies. Successful aspects of these messages can be used by advocates, lobbyists, businesses, and others to promote clean meat. The ultimate goal is to reduce reliance on animal farming by encouraging as many people as possible to switch to clean meat once it becomes available.

Key Findings

1. Acceptance of clean meat was high (and higher than has been observed in many previous studies): Specifically, in this study, 66% of people were willing to try clean meat, 46% were willing to buy clean meat regularly, and 53% were willing to eat clean meat as a replacement for conventional meat. We think these rates likely stem from the positive descriptions of clean meat we gave participants, as well as the term “clean meat” itself. However, further research will be needed to determine which aspects of the messaging are effective, as this study did not directly compare them.
2. Telling potential consumers about the unnatural side of conventional meat was effective: People who read about the unnatural conditions in which farmed animals are raised came to believe that conventional meat is unnatural.
3. Describing conventional meat as unnatural produced the most acceptance of clean meat: Potential consumers who read this message were willing to pay more for clean meat than those who didn’t. People who read this message also tended to be the most positive about clean meat in a variety of other ways: attitudes, feelings, and beliefs. 
4. Trying to directly reduce naturalness concerns was ineffective: The other two messages tested in this study—which described the natural side of clean meat and questioned the idea that naturalness is important, respectively—were not convincing to participants. Given that these messages were developed by subject matter experts with multiple rounds of feedback, we conclude that these arguments may be difficult or impossible to use effectively.

Study Design

The design and experimental procedure for this study were pre-registered with the Open Science Framework.

A census-balanced, representative sample of 1,185 U.S. adults was recruited through the research firm Ipsos. This number exceeded the 1,100 that our power analysis deemed necessary.

At the beginning of the study, participants answered questions about their familiarity with clean meat and read a brief passage describing it. Most participants were not at all (64%) or a little bit familiar (21%) with clean meat before the study.

Each person was then randomly assigned to read one of four messages about clean meat. All four began with the same introductory paragraph, followed by one of the four messages about naturalness:

• an argument that clean meat is natural,
• an argument that conventional meat is unnatural so clean meat is preferable,
• an argument challenging the appeal to nature, or
• a control message about the benefits of clean meat for health, the environment, and animals. This was meant to be similar to current messaging about clean meat.

After reading one of these messages, participants answered questions designed to measure their interest in clean meat, as described below.

Perceptions of Naturalness

The main purpose of this study was to try and influence perceptions of naturalness. The results showed that of the three experimental messages, one was persuasive but the other two were not.

Most people don’t tend to think that conventional meat is unnatural: only 20% thought it was unnatural, while 49% disagreed. However, people who read a message arguing that conventional meat is unnatural were significantly more likely to believe it: 28% agreed and only 39% disagreed.

On the other hand, participants who read a message arguing that clean meat is natural or that meat naturalness doesn’t matter did not appear to believe those statements. Overall, about a third of people (34%) agreed with a statement that “clean meat is unnatural.” Another 34% disagreed and the rest didn’t take sides.

Almost two thirds (66%) agreed with a statement that “it is important for meat to be natural,” while only 9% disagreed. The remainder did not choose a side. These findings suggest that different people have a different idea of how natural clean meat is, but most people think that it’s important for meat to be natural, and we weren’t able to convince them otherwise.

Behavioral Intentions

In this study, 66.4% of people were (probably or definitely) willing to try clean meat, 45.9% were willing to buy clean meat regularly, and 52.8% were willing to eat clean meat as a replacement for conventional meat. More details are provided in the full report.

It is a little surprising that more people said they were willing to eat clean meat as a replacement for conventional meat than were willing to buy it regularly—the former seems like a larger ask. We believe that this “reversal” may be driven by the wording of the replacement question, which does not specify exclusively eating clean meat instead of conventional meat.

In contrast, in a study conducted by Wilks and Phillips (2017) that examined rates of acceptance without positive messaging, about two thirds (65.3%) of people were willing to try in-vitro meat (their term), but substantially fewer were willing to eat it regularly (32.6%) or eat it as a replacement for conventional meat (31.5%). Other studies that did not use positive messaging have also found low rates of willingness to eat clean meat, between 16% and 40%.

Willingness to Pay (WTP) for Clean Meat

Participants were asked to imagine that they were at the grocery store and saw clean and conventional meat products available. They were shown a price for a conventional meat product and asked how much they would be willing to pay for the clean meat version of the same product. Each person made this decision three times: for chicken nuggets, beef burgers, and fish sticks. They could also indicate that they wouldn’t buy the clean meat product or that they wouldn’t buy either product.

Figure 1 shows the results, which were very similar for the three products.

When comparing the messages, it is important to understand how confident we can be in the results of the comparisons. Statistical tests give us a good idea of how likely it is that we would consistently see similar results on this survey if we administered it to other groups of people. Generally, we consider a test result significant (i.e., sufficiently trustworthy) if we are confident that 95% of the time we would get the same result. If we expect to get the same result 90 to 95% of the time, this is sometimes considered “marginally” trustworthy.

With that in mind, people who read an argument that conventional meat is unnatural were willing to pay significantly more for clean fish (with 97% confidence). They were also willing to pay marginally more for clean chicken (with 92% confidence).

Finally, people who read that conventional meat is unnatural were willing to pay a bit more for clean beef, but this result didn’t reach conventional standards of significance (confidence was 87%).

We know that advocates and manufacturers of clean meat would like a better idea of the actual amounts people will be willing to pay. Because of this keen interest and the lack of available data, we will provide rough estimates in a follow-up blog post. Although we hope this analysis will be useful, it is also quite speculative, with several important limitations to bear in mind.

Other Measures of Clean Meat Acceptance

The remainder of the study included more typical self-report questions measuring clean meat acceptance. They included the following:

• Beliefs about clean meat (safety, health benefits, etc.)
• General attitude about clean meat
• Affect (mood)

The overall picture of the results for the self-report measures (including these and the behavioral intentions measures above) was very clear: Reading an argument that conventional meat is unnatural makes people more accepting of clean meat than the argument that the naturalness of meat is unimportant (p = .008). Thus, it is clear that of these two messages, arguing for the unnaturalness of conventional meat is the better choice.

However, none of the three naturalness messages was significantly better or worse than the control condition (all ps > .12). Therefore, although the ‘conventional meat is unnatural’ condition did the best of the four, the only strong evidence for using it rather than current messaging comes from the hypothetical grocery store choice above. The current messaging performed as well as the naturalness messages in most cases.

Implications

In a real-world context, consumers will not answer questions about their willingness to eat clean meat, they will be faced with a choice between it and the more familiar, conventionally-produced meat. These results suggest that, in that choice context, focusing on the unnatural aspects of conventional meat may be the most effective way of increasing interest in clean meat. In short, it appears to make consumers more aware of the positive contrast between them.

That being said, such an approach would represent a fairly aggressive stance towards conventional meat producers, which may not be an optimal strategy for advancing clean meat. Several conventional meat producers are already backing clean meat technology, so encouraging others to do so as well may be a better strategy than fighting them with legal challenges or marketing. This question warrants further consideration.

Given the care that was taken in developing the experimental messages, and their lack of influence over the outcome variables, we believe it is reasonable to interpret these results as an indication that arguing for clean meat’s naturalness or the unimportance of naturalness are difficult strategies to use effectively.

Limitations

As with all research, this study was subject to several limitations. First, because only U.S. adults were studied, the findings may not be generalizable to other cultures or countries.

In addition, the proportion of would-be participants who were removed for failing attention checks was higher than we would like. Although their removal ensures our respondents were paying attention, it may introduce some selection bias. More generally, it may be indicative of low panel quality.

It is also worth noting several limitations of the WTP measure in particular. First, it is important to bear in mind that this measure directly followed positive messaging about clean meat, potentially producing higher values than would be observed in reality. In addition, because this measure is hypothetical, it is susceptible to the commonly-observed hypothetical bias, in which consumers tend to overestimate how much they are willing to pay for a product (e.g., Loomis, 2011). It is for this reason that we have provided only broad WTP categories above and focused on the comparison between conditions.

Participants’ self-report responses may also be subject to bias. First, forecasting error is probable: Predicting one’s own future attitudes and behaviors towards a product which is not yet available is difficult. Unfortunately, there is little that can be done to avoid it, as clean meat is not yet available. Hypothetical and predictive questions are the only option, though we took care to frame them as realistically as possible.

Finally, participants may have been subject to social desirability bias—answering as they believe others would want them to—for questions about a product with such profound ethical and environmental implications. That said, because even participants who read our control message were exposed to arguments about these implications, we believe that the potential impact of this bias is minimal.

Future Directions

We suggest that future research carefully consider whether trying to directly overcome perceptions of unnaturalness is the most effective option before pursuing it further—a few of this study’s effects suggest there may even be potential for it to backfire. These results suggest that a focus on the unnaturalness of conventionally-produced meat is more likely to be effective, but as noted above, this is not without risk of alienating potential allies.

In addition, the effectiveness of the ‘conventional meat is unnatural’ message in this study was limited, with mixed results across different outcome measures. We recommend that, if this is to be considered as a strategy for advancing clean meat, further testing of similar and stronger messages should be carried out.

The high rates of clean meat acceptance observed in this study at the overall level and in the control condition suggest another potential strategy: that providing potential consumers with positive educational messaging about the benefits and characteristics of clean meat may be a good way to reduce the emphasis on naturalness before it becomes the focus of the conversation. This study does not provide strong evidence about this possibility because we did not include a no-message control group, opting instead for current messaging. Previous research that has directly examined the impact of positive messaging has found that it can be effective (Verbeke et al., 2015; Bekker et al., 2017).

We recommend that future research do more to examine which aspects of educational messages are most effective in increasing acceptance rates: for instance, information about the taste, texture, and nutritional profile, or the health, environmental, or animal welfare benefits. This study included all of these to apparent good effect, but further experimental research will be needed to narrow down the key ingredients so that they can be emphasized.

References: See full report for references.

Research team: Jo Anderson (Faunalytics), Chris Bryant (University of Bath), Kathryn Asher (Animal Charity Evaluators; formerly Faunalytics), Che Green (Faunalytics), Kris Gasteratos (Cellular Agriculture Society), Bruce Friedrich (The Good Food Institute), Jeff Rotman (Deakin University), Jamie Macfarlane (The Good Food Institute).

Funding: We gratefully acknowledge Animal Charity Evaluators and the Animal Advocacy Research Fund for their assistance in funding this project.

The full report for this study is available here. It provides details of all measures and analyses, as well as a higher level of technical detail.