New study shows “BPA-free” labels may increase risky behavior

In All, General by Andrew Maynard

Products with the label “BPA-free” have become ubiquitous on store shelves in recent years.  It’s a trend that has been driven by consumer concerns that the chemical bisphenol-A, or BPA, may be harmful at low doses.  Yet a recent study suggests that the label may mislead consumers into thinking that “free” means “safer” — even when there’s a chance that the substances used to substitute for BPA may also be harmful.  The study is one of the first to explore how consumer responses to uncertainty and ambiguity in risk information may lead to “regrettable substitutions” — the replacement of one material with another that is potentially less safe.


BPA — or Bisphenol-A — is a chemical that’s used in products ranging from plastic bottles and canned food liners to cash register receipts.  Studies have shown that it behaves like a hormone in the human body, and at high exposures can potentially lead to or exacerbate a range of health impacts, including damage to the liver and kidneys, and possible impacts to the reproductive, nervous, immune, metabolic and cardiovascular systems.  When low, long-term exposures occur though, the science is less clear.  Many experts interpret available data as showing that there are no significant health risks from current consumer exposure levels. Regulators around the world agree with this assessment and have established what they consider to be acceptably safe levels.  Nevertheless, there are are some scientists who argue that people are overlooking subtle but important health impacts that are potentially associated with exposures below levels considered to be safe.

Regrettable Substitution

Even though the balance of professional opinion is currently that BPA doesn’t present an appreciable health risk when used in most consumer products, there has been a widespread move — driven largely by public opinion — to remove BPA from consumer products.  The move quickly spawned products with “BPA-free” labels – especially water bottles, where there were initial concerns that BPA could leach out of BPA-based polycarbonate and into the liquid.  In most of these products though, BPA was used for specific purposes, meaning that removing it required that another substance — a substitute — was used in its place.

But how do consumers decide which is the better (or worse) deal when such substitutions occur — the original BPA, where there is some information on likely risks, or the substitute substance, where in most cases there is very little known about possible risks to health?

This was a question that Laura Scherer and a number of colleagues (myself included) addressed in the recent paper “The psychology of ‘regrettable substitution’: examining consumer judgements of Bisphenol A and its alternatives“, recently published in the journal Health, Risk and Society.

Consumer Responses to Ambiguous Risk

In the study, we were particularly interested in two questions:  First, if consumers are presented with two near-identical products (we chose canned tomatoes), one of which contains BPA, and the other a BPA-substitute (we used polyethylene terephthalate or PET), do they have clear preferences? Second, if consumers discount available risk-information on BPA because of the public controversies surrounding it, what then influences their their decision?

Common to both of these questions is how people handle ambiguity in decision-making.  In this situation, consumers must choose between a reasonably well-characterized risk which some experts are still debating (BPA), and a totally unknown risk (PET). To study this,

The order in which information is received matters

The first part of the study evaluated participants’ perceived risk from BPA and PET.  We asked a representative sample of the US population to read two newspaper-style articles about the used of BPA or PET in canned tomatoes.  From the publication:

For Bisphenol A, participants were told that there was some evidence that Bisphenol A caused some specific harm, but that this evidence was controversial for a variety of reasons. Participants were additionally told that some information was known with relative certainty, including the fact that ‘BPA [Bisphenol A] clearly does not have as significant an impact on health as smoking, eating an unhealthy diet, or living a sedentary lifestyle’. In the case of the alternative chemical … participants were told that there was no information about its safety or lack thereof, because it had never been scientifically scrutinised.


Participants were randomly assigned whether they read about about BPA or PET first.

The results of the survey were not entirely expected — they showed that after reading the two articles about BPA and PET, people ranked the risk of the second substance they read about as higher than the first substance.  The differences weren’t great, but they were statistically significant. The order in which the information was presented changed their ranking of perceived risk, even though participants were told that there was evidence for the risks and safety of BPA, and none for PET.

This order effect seemed to suggest that participants were interpreting uncertainty over the safety of BPA in the same way that they interpreted a lack of information on safety or risk.  Interestingly, the uncertainty we presented them with was journalistic uncertainty rather than scientific uncertainty: the article they read focused on expert debates around what is not known about the substance rather than what is known.  While this requires further investigating, there are indications that how ambiguities around risk are reported in the media may have a direct impact on risk decisions made by consumers.

BPA-free Label  on water bottle

How does “BPA-Free” labelling influence risk perception?

The findings of this first phase of the study were consistent with what might be predicted from other research into how people make decisions based on ambiguous information.  But they didn’t seem to reflect the seeming widespread rejection of BPA-containing consumer products that has occurred over the past few years.

One explanation for this is that many products using BPA substitutes are labeled “BPA-free”, with no mention of what the BPA has been replaced with.  We were interested in how this labeling affects perceptions of risk and decisions, and so conducted a follow-up study.

In this study, we presented another set of survey participants with the newspaper-style article on BPA used in the first part of the study, followed by the article on PET.  We then asked them about their perceptions around the risks.  But we included an added dimension: For half of the participants, the articles were modified to explicitly refer to canned tomatoes using PET as being “BPA-free”. With these participants, we asked them about their product purchasing preferences after reading the information provided.

We found that when participants were not exposed to the “BPA-free” language, their perceived risk of the BPA-containing cans was lower than that for the PET cans.  This confirmed the order-effect findings of the first part of the study.   However, when the “BPA-free” language was added, participants ranked their perceived risks of the two chemicals as being similar.  In other words, labeling PET-containing cans as “BPA-free” countered the order-effect on risk perception seen in the absence of the label.

The participants who read about the “BPA-free” alternative were also asked how risky it would feel to eat tomatoes from a BPA-free can. They indicated that they perceived ‘little risk’ here. When asked whether they would like the option to purchase “BPA-free” cans, they were strongly in favor of this.  They also indicated that they would be willing to pay 28 cents more for a product labeled as “BPA-free”.

In other words, adding the label “BPA-free” influenced people’s perceptions of risk and safety, even though the lack of knowledge about substitute chemicals was clearly stated and the BPA-free labeled cans could possibly present a greater risk to health.

Substance-free labels can mislead consumers

These two studies suggest that participants did not have a strong or a stable preference for well-studied chemicals with possible but low likelihood health risks (BPA) versus a chemical that is virtually unknown with regard to its health risks (PET).  They also indicate that controversial evidence of harm — or at least widely reported controversy — may be viewed as being no better or worse than having no information at all.

When it comes to labeling, the studies indicate that a product labeled as “BPA-free” leads to greater acceptance of that product, even when consumers know that BPA has been substituted with another substance that may potentially be more harmful. In other words, the findings indicate that labels like this are misleading, and are likely to cause some people to accept substitute chemicals that they might otherwise reject.

Irrespective of the current state of science on BPA and health risks, this study sheds new light on how people respond to information where they have a choice between well-studied but still controversial substances, and poorly studied substitutes.  It indicates that labeling a product as being free of a particular chemical leads to a lack of consideration over the risks potentially presented by substitute materials.  And it suggests that care needs to be taken in how evidence and risk are communicated if substitutions aren’t to become regrettable substitutions — whether making decisions on BPA, or any other substance where there is some element or doubt over risk and safety.

Paper: The psychology of ‘regrettable substitutions’: examining consumer judgements of Bisphenol A and its alternatives.  Laura Scherer, Andrew Maynard, Dana Dolinoy, Angela Fagerlin and Brian Zikmund-Fisher.  Heath, Risk & Society, 16:7-8, 649-666 (2014).  DOI: 10.1080/13698575.2014.969687