Organic consumers not very concerned about GE

posted in: Commentary, Science | 34

(Hat tip to Elton Robinson at South East Farm Press)

The Consumers Union wanted to know what consumers felt about genetically engineered crops cross-pollinating with organic crops. So in early February, they conducted a poll. They called a thousand random people over the phone and asked them just two questions:

1. Do you buy organic food, such as produce, meat or dairy products? (Yes/No)

2. Please rate your concern with organic food crops that are contaminated by genetic engineering. Are you…

  • Extremely concerned
  • Very concerned
  • Somewhat concerned
  • Not concerned at all

Sounds like a pretty simple exercise. However, I question the use of the term “contaminated.” This is a loaded term, and assumes one of the things that they want people to believe – that organic agriculture should not include genetic engineering. This introduces a bias into the poll. For instance, if you asked these two similar and benign questions, you would get two different results:

A. Please rate your concern with organic food crops that cross-pollinate with hybrid crops.

B. Please rate your concern with organic food crops that are contaminated by pollen from hybrid crops.

Of course, hybrids are allowed in organic agriculture, but I’ll bet my backyard garden harvest for this entire year that if you ask these two questions the word ‘contaminated’ will have a measurable effect and cause people to answer that they are more concerned than they would be otherwise.

Ok, that issue aside, it is good that the Consumers Union did a poll such as this, because there hasn’t been very much research investigating what people really think about genetic engineering and organic agriculture, and there’s been some talk about it in numerous channels. So how concerned are consumers about this “contamination?”

The Consumers Union announced its results on the 2nd of March:

Consumers Union Poll: Two-Thirds of Organic Foods Consumers Concerned with Genetically Engineered Contamination

Results Contradict USDA’s Position That Consumers Don’t Care

Yonkers, NY—Consumers Union, the nonprofit publisher of Consumer Reports, today released new poll data showing that two-thirds of organic food consumers are concerned about genetically engineered (GE) ingredients contaminating organic food. Given the popularity of alfalfa sprouts among health-oriented eaters, Consumers Union urges the U.S. Department of Agriculture (USDA) to consider the overwhelming consumer concern before deciding to allow GE alfalfa on the market. USDA has until Wednesday, March 3 to receive public comment on its draft Environmental Impact Statement (EIS) on approval of GE alfalfa. The poll results can be found online at Poll_Public Release_Feb 2010.pdf.

What did they find? Although they linked to the paper itself, those less inclined to delve into those details can easily just read a few more lines down to find out what people thought.

A majority of respondents expressed some level of concern with genetic engineering contamination of organic food crops. Overall, 58% said they were extremely concerned, very concerned or somewhat concerned with this contamination.

Two-thirds (66%) of consumers who purchase organic food indicated being concerned versus half (50%) of those who don’t make organic food purchases.

This announcement came one day before the close of the comment period for the Environmental Impact Statement for GE alfalfa. It seems that this poll was conducted specifically to address the question of organic concern over GE crops, which has been amplified recently by calls from the Center for Food Safety that has been challenging the biotech alfalfa and recently, sugar beets in court. 58% of consumers concerned about pollen drift from GE crops to organic crops sounds like an impressive opposition! However, if you read the poll results themselves, you will find that they found the opposite of what was claimed in the press release.

Luckily, they published the poll results so that inquiring minds could figure out what was being obscured from casual readers. I have reproduced their main data table, showing the breakdown of consumer opinions:

Buy OrganicDon’t Buy Organic
Buy OrganicMenWomen18-3435-5455+Don’t BuyMenWomen18-3435-5455+
Respondent Base998528236292183193151469247222122187159
Extremely concerned1112121312151094151187
Very concerned131811241817197510968
Somewhat concerned34353635323639333136463028
Not concerned at all41334028373230495939345455
Don’t know1111102111011

As you can see, when you look at the individual percentages for each response, it comes out completely different from how they describe. The category with the highest number of responses is “Not concerned at all” at 41 percent followed by “Somewhat concerned” at 34 percent.  “Very concerned” and “Extremely Concerned” only make up 13 and 11 percent each. While their categorical description that consumers expressing any level of concern at all add up to 58%, this hides the distribution of the level of concern.

Usually when you design a poll (assuming you don’t use loaded terms!) you try to include several responses that can inform you about the distribution of respondents’ opinions. This is often on a 1 to 5 scale, where 3 is the neutral or undecided category. This poll had five categories, with “Extremely concerned” at one end and “Not at all concerned” at the other end. The next two categories inward from the extreme positions are the moderate opinions, “Very concerned” and “Somewhat concerned”. Somewhat concerned is the moderate partner of Not at all concerned, and if you were to properly categorize is on the scale of concern, it goes with “Not at all concerned.”

To put it another way, say you designed a poll that asked people if they had any politically conservative opinions. The responses could be, “Extremely conservative,” “Very conservative,” [Neutral],  “Somewhat conservative,” and “Not at all conservative.” You would find moderate liberals who have some conservative opinions falling under the “somewhat conservative” category. But would it make any sense to say that people in this category count overall as conservatives? That’s what the Consumers Union did.

You could say that they could have designed better responses. This would go hand-in-hand with a more neutrally-worded question.

“Somewhat concerned” = “Not very concerned” (but still a little concerned…)

The vast majority of consumers polled by the Consumers Union are not very concerned or not at all concerned about organic “contamination” by GE crops, at 75% of the sample. Those who are very or extremely concerned only make up 24%. Even amongst the people who say that they buy organic food, it is still 68% not to 30% yes. Those who don’t buy organic are 82% to 16%.

Mind you, half of the people who answered the phone at dinnertime said that they buy organic food, and since we all know that organic makes up only about 2% of the food out there, they’re not all buying a lot of it. I imagine that the hard-core organic folks are far more in the “very” and “extremely” concerned categories. Perhaps they should have asked how much organic food respondents purchased, perhaps there would have been an interesting trend from low concern to high concern as you go from infrequent to frequent organic food consumption?

In case there was any doubt in your mind about the shape of public opinion, here’s a handy graph of the answer to their poll question:

So I think it is pretty clear that the actual distribution of the results differ from the way readers were led to believe. And these results are right in line with the USDA’s analysis of general consumer opinions on GE crops. And, might I add, maybe Michael Pollan’s as well. The fact is, most consumers do not really care very much if at all – it’s not on their radar screen. The Consumers Union, although accurately quoting the 58% figure, is still misleading people when they say that their poll data disagrees with the USDA’s conclusion.

But it doesn’t stop there, look at what Michael Hansen, their senior scientist said:

“The EIS states that consumers and organic farmers don’t care if their organic food is GE contaminated,” said Michael Hansen, PhD, senior scientist with Consumers Union. “Consumers Union’s poll states the exact opposite: consumers care greatly.”

But as we have seen from looking at the actual results from their own poll, this is entirely NOT the case. I don’t know very much about the Consumers Union, and I know even less about Michael Hansen, but he is being flagrantly dishonest about their results. Since he is the senior scientist for the Consumers Union and this is an official press release, this reflects poorly on the organization itself as well. There are all kinds of cases where wishful thinking plays a role in people’s opinions of polls and trends like this, but you cannot conclude what Hansen has from this data. It is rare that I find cases where I can say for sure that falsehoods are knowingly being propagated, and this is one of them.

If this is what they do with reasonable and understandable data, I daresay we should be wary if the Consumers Union does any more polls on genetic engineering in the future.

Finally, here’s the big irony – let’s come back to what I said about using the term “contaminated.” Even though the question was loaded, they still couldn’t produce the result they were looking for!

  • Ewan R

    And Inorganic consumers (yay robots!) are concerned even less!

    Clearly though when 65%+ of people couldn’t give a hoot a technology categorically should be slated for purely populist reasons, right? Especially when the sector which is/isn’t impacted by the cause for concern is equivalent to less than 5% of agriculture.

    I find it a tad worrying however that there aren’t more ‘dont knows’ – my own personal experience would suggest this should be far higher than 1% of people (as I have personal recollection of more than 2 people who were perfectly willing to admit having no idea about transgene contamination, and I’m pretty sure I’ve not discussed it with more than 200 people, unless I have some serious memory blackouts going on)

    Another loaded aspect of the question is that it appears to paint the picture such that it is a given that organic foods are contaminated, rather than asking how concerned people are about the potential for contamination (which I’m assuming is what they’re pretending to try to ask?)

  • Ewan R

    Fast work on the edit… now change my math so it is right again! =p (200 people…)

  • Mary

    Have to agree with Ewan on the inorganic consumers. My roomba occasionally eats the cat food if it gets into the kitchen when I’m not paying attention. It doesn’t seem to matter what’s in the cat chow.

  • Karl Haro von Mogel

    Bah! That’s it, I’m installing a self-edit plugin for those of us *cough* Ewan *cough* that have a hard time with numbers! :)

    It took me a moment to get the joke about the roomba.

    I did find the low number of ‘I don’t know’ responses to be odd. That’s probably the smallest undecided category I’ve seen for any poll on GE

    • Ewan R

      Meh, I don’t have a hard time with numbers… it’s just my attention span in reading the table is lacking… by the time I got to the don’t know section I was back to assuming they were talking in terms of %ages rather than absolute numbers, hence 0.1% and 2000 people etc – which will only make sense to Karl as he edited my (correct – only the underlying assumptions were ass backwards) figures to ones which were more reality based

    • Benjamin Edge

      Thanks for providing this data Karl.

  • pdiff

    You’re right on target, Karl. The survey is both biased and unbalanced. Kind of like looking at a physiology study with 6 rats …

    But, since you had data …. I just couldn’t resist a look see :-) I stuck to the original categories, rather than the lumping they did. My quick analysis shows:

    People who don’t buy organic don’t give a frap.

    Males have higher concerns about this.

    The 18-34 yr olds had no preference, but older people (especially middle aged) did.

    So, the anti GMO target from this data is the middle aged, male organic buyer :-)

    Also of interest, the overall odds of being concerned for the individual categories of Extreme, Very and Somewhat are not significant. It is only when you put them together that they become so.


  • Ewan R

    A physiology study with 6 rats, or a safety study with 400 rats (I went and looked up the Study (or at least a study… I may have missed the pertinent one) and the only reference to ‘6’ I could see, at least in the abstract, was 6 extra groups (of 40 rats per grp))

    I’m assuming this is the study Seralini reanalyzed as it’s on MON863

    • pdiff

      Sorry Ewan. Admittedly I was being a bit snarky there, but I had in mind the study by Stanley W. B. Ewan and Arpad Pusztai (no relation, I assume). In that study “Six rats were randomly allocated to each group…” I am suspicious of this study as they also state “Data are the means of six animals calculated from five observations for each.” This smells of pseudo replication where they are trying to inflate the sample size to 30 per group. The paper is not clear, but does give means and standard deviations. I will elaborate on that on the other thread in the forums, but bottom line: the statistical power assuming n = 6 for the detection of a “GM” effect is very low for all responses.


  • Ewan R

    I was thinking that on the way to work this morning actually (that I may have reacted to a point you didnt make…) – getting forum discussions confused with this here… I’m in perfect agreement that 6 reps is low (I was actually surprised that the monsanto study had as many as 40 rats per group when I looked into it though, definitely a number worth remembering next time it is brought up that the safety studies are woefully inadequate)

  • Karl Haro von Mogel

    Pdiff, I think you got some of the number backwards – the men seem to be less concerned about cross-pollination than women, and although the young age bracket is less concerned in the organic group, they are more concerned in the non-organic group. Older organic buyers are more concerned, and older non-organic buyers are less concerned. But in all categories, the distribution of concern is on the low end.
    I was talking with a professor of mine this morning, and one thing we rtalked about is how the ”not concerned’ category might include the undecided people, and that this poll does not separate out whether people are concerned that ‘contamination’ could happen, or that it does happen, or if they are concerned about when it will happen.
    Also, I want to point out that when you see these gender or age-specific differences discussed by some of the people who are anti-GE, it is often believed that this means they should focus on those groups. For example, I know Jeffrey Smith has said that since women are less inclined to buy GE foods, that they plan to market more to women. However that could just be preaching to the choir, and that they could gain more traction by trying to market to men. For instance, why convince people who buy nothing but organic food not to buy GE food – they already don’t! Conversely, it would be good for pro-GE folks to appeal to women since the most ground could be gained in that group.
    HOWEVER – Gender, age, race, socioeconomic status, political affiliation – most of these categories give mixed results from published study to published study. Gender as a factor in GE acceptance is not very reproducible.

    • pdiff

      Short on time right now, but I’ll check again when I have a chance. I will post the data and codes as well, if anyone wants to check. It was quick and dirty on my part, so I may have a mistake (don’t tell anyone I said that :-) ). My recollection is that a logistic regression found the odds of concern higher for organic buying males as well as those 34 and up. Of course, odds and actual numbers can seem conflicting too. Will get back to you on that …

    • pdiff

      I understand what you are saying and, in terms of the numbers, it is correct. Overall, the males are 50% unconcerned, while the females are around 33%. The logistic regression analysis I looked at, however, is aimed at examining the differential between buy and don’t buy categories.

      Before I get to that, let me explain the data I used, as rounding errors may cause difference between these results and the survey. I took the data you provided (actually from the article itself) and multiplied the percentages given by the totals reported to get back to “number of individuals”. I also rounded these off to whole numbers. These are small errors and will not affect the outcome. Additionally, I have collapsed the last category, “Don’t Know” into the “Not Concerned” category. It was very sparse (suspiciously so, as others have noted) and will have limited influence through collapsing (keeping it on its own causes estimation problems). So, I have 4 categories “Extremely Concerned”, “Very Concerned”, “Somewhat Concerned”, and “Not Concerned”. Unfortunately, because of the way the data is reported, we can not look at cross classifications with all the demographics, e.g. race, income, education, etc. The data is limited to the overall “Buy, Don’t Buy” category and its cross classification with “Gender” or “Age”, therefore, my analyses are limited to those three scenarios. Lastly, I chose to use logistic regression (generalized logit) as the analysis tool. This will look at the odds and odds ratios of various classifications based on the probablility of being concerned relative to the probability of not concerned. Such analyses are standard protocol for these types of surveys. All data, programs and results are available at: http:\\\GMO\GMO.NRC.html .

      Meanwhile , ….

      I don’t think I can reproduce the tables here in this reply, so I’ll refer you to the PDF results file on the web site above. The Gender tables are on Page 5. The odds of Female buyers being extremely concerned relative to not concerned would be 12.67/28.42 = .44, while that of non buying females is 14.86/39.64 = .37. The odds ratio comparing buy vs not buy would be .44/.37 = 1.18 or roughly equal. The males, on the other hand, show quite a difference: Odds for male buyers = 12.02/41.20 = .29 ; Odds for male non buyers = 3.69/60.25 = .06. The male odds ratio for buy vs not buy is .29/.06 = 4.7. All the relavent odds ratios and their confidence intervals are computed and reported on Page 8. Note that the Female ratio is non-significant (confidence interval covers 1.0 => Equal Odds), while that for males is significantly greater than 1.0 indicating male buyers are more likely to be extremely concerned (about 4 times more than non buyers).

      Similar analyses are given after that for the Age groupings. The 18-34 group shows equal odds for all concerned categories while older groups show more concern.

      I realize that all this may seem confusing given what you noted earlier regarding the overall preference for not concerned. That’s the oddity of these types logistic regression analyses. They look at relative differentials. You were correct to note the overall proportions. This type of discrepency is why I despise virtually every reported clinical trial and survey result. All they ever report are the odds ratios, but these, as you can see, do not tell the whole picture and can be decieving. They should always be reported with the underlying numbers and probabilities. Ok, ok … soapbox mode off :-)

      Hope that didn’t confuse things too much!


      • pdiff

        Must Have Edit Mode ….

        Also, I noticed that, in my exceedingly long replies, the input box does not allow me to see everything I type (I have to type elsewhere and paste in => problem above). Then again, maybe you want shorter replies :-) Will work on that …

        • Karl Haro von Mogel

          Yes I need to fix the edit box. It is weird that it does not allow you to scroll down if the comment is long. Every theme has its issues that need to be fixed… let me see if I can do something about that right now. Self-edit plugin will be on its way too!

          Update: comment box embiggened and enscrollbarred. All too easy…

      • Karl Haro von Mogel

        I understand now what you meant. You were talking about the differences between organic buyers and non-buyers with regard to gender and age. What you said now makes perfect sense, both intuitively and from the numbers. Men who buy organic foods are much more likely to be concerned about GE ‘contamination’ than men who do not, whereas women are little different whether they buy organic foods or not.
        Like I mentioned before, I think the anti-GE people really do not understand that they are preaching to the choir in their strategies. The focus is on convincing organic buyers, convincing women, etc, that GE crops are bad. Indeed, the marketing that happened during the Sonoma County, CA ballot initiative, Measure M, focused on marketing to women-as-moms. This was recognized by a sympathetic geography grad student to be reinforcing sexist gender stereotypes for one thing, but also missing out on ground that could be gained amongst men. (Maybe I shouldn’t say this too loudly, otherwise they might catch on!)
        So that is interesting that the younger crowd is not very different between organic and non-organic buyers. I wonder what that lack of a difference reflects? No difference in openness between young people in either group? This might make sense since the older group (baby boomers) was around when organic got started, and some of them may carry the philosophical baggage that predisposes them to be more concerned about this admixture of agricultural technologies.
        Glad to have you with us!

        • Ewan R

          Preaching to the choir may pay off in that if you don’t perpetually reinforce a bad idea it is more easily displaced – it’d be a major blow to the anti-GM ideologues if their grass roots got turned off to their message – I’m not convinced either that targetting those who do buy organic is preaching to the choir – particularly in light of this survey (when looked at in a more binary fashion, lumping exteremely and very) more than (close call!) 60% of organic consumers essentially don’t care – which is a sizable majority in a nation normally split 50.1:40.9 on contentious issues (at least that’s how it seems….) – basically you split the population into two subsets – organic and robot killing machines (or whatever nomenclature is current for those who don’t buy organic…) and pick which subset is more likely to support your cause (which undoubtably is the organics, as robot killing machines have very little concern for the environment above and beyond how awesome it looks when exploding) or consider each population as a bell curve of support -your better off focusing on the one which if shifted is more likely to bring you more support (shifting the organic curve slightly to the against side would garner more support than a similar shift in the robot curve)

          My take on the younger crowd being more homogenous in terms of response is that as far as I can tell environmental concern, and the erroneous linking of GM to environmental damage, was far more prevalant as an educational tool during that cohort’s formative years – at least in my limited experience

          • Ewan R

            49.9. Edit function please =p (and that’s typing, not math!)

        • pdiff

          Thanks for the fixes and edits above.

          “So that is interesting that the younger crowd is not very different between organic and non-organic buyers. I wonder what that lack of a difference reflects? No difference in openness between young people in either group? This might make sense since the older group (baby boomers) was around when organic got started, and some of them may carry the philosophical baggage that predisposes them to be more concerned about this admixture of agricultural technologies.”

          Well, seeing things in the data are one thing. Assigning causes, especially in a biased survey like this, is more difficult :-) . It could be the younger crowd is just more concerned about other things or perhaps they are more accepting of the newer technologies than their elders. If those elders are holding on to “philosophical baggage”, it will be interesting to see how they face up to the revised views of their mentors like “The God-Father of Green”, Stewart Brand.

          I was disappointed to see no detail on the income bracket. I would like to see how that breaks out versus buy-don’t buy as well as the concern question. I would suspect that people with higher incomes could better afford to be concerned about such things.

  • Ewan R

    It would also have been somewhat more interesting if the split was between people who buy only organic, mostly organic, some organic, occasional organic, and no organic – I think the split would be a little more telling there, with a guess that only organic would sort out the hardcore who are extremely concerned, mostly organic the rest of the extreme concern and the bulk of the moderate concern, leaving those of us who buy occasionally (guilty as charged, but predominantly because organic food stores locally tend to have far better fish departments – I won’t buy fish from anywhere that smells fishy) to not give one hoot, or at least only be a little concerned.

  • Karl Haro von Mogel

    This also ties into the question of ‘who exactly is the organic community?’ Margaret Mellon said that the use of genetically engineered crops in organic agriculture should be decided by the organic community. Would that be the farmers? Advocacy groups? Consumers? If we included everyone that bought organic at some point, that would be half the country!
    And it would include you and me.

  • Ewan R

    Awesome, I vote yes, where do I sign? (although I get the feeling that those who vote yes are no true members of the organic community)

  • Karl Haro von Mogel

    Organic agriculture wants to be more mainstream, but what comes with that is either including mainstream opinion – or dictating what it should include/exclude from the top-down. If organic holds to democratic principles, it might change. But the first step would be allowing that debate to take place within the organic community, and that means inviting people who have compelling arguments for integration to organic conferences to speak… (hint hint)

  • Karl Haro von Mogel

    Great discussion going on here. I just posted some comments and links on other sites that are discussing this poll, so maybe we will have a chance for others to take part in it.

    • pdiff

      Did they kill the forums? :-)

      • Karl Haro von Mogel

        Looks like we had a forum plugin bug that is now fixed. Discuss away!

    • Anastasia Bodnar

      May I ask where you posted?

  • Becca

    I’ve been an organic consumer since the mid-80’s, and have been purchasing more and more certified organic agriculture as it has become available.

    The more I know about how conventional food is grown, the more I will spend every food dollar fighting GMO’s and toxic untested agriculture. I admit that I was slow to act, choosing to gather information first before becoming an activist.

    But, my way of helping the economy is to fight the Obama Administraton like I did with Bush-Cheney before them. The lies and deceit about the so-called safety of our food.

    The only safe food that is nutritionally rich is organically grown in accordance to the original law passed by Congress –not the weakened law as it is today. But, even at that, I’ll still buy from organic farmers over Monsanto, given a choice.

    Of course, U.S. Senate bill 510 eliminates that choice. It allows BIG BROTHER to control all the food and gives the consumers, backyard gardeners, small farmers, no choice about what they grow or buy.

    This report of yours is not factual. You have taken things out of context. Organic consumers trust in organic agriculture and organic farmers they know, which is the reason Monsanto has to eliminate our rights to get their way.

  • Becca

    My opinion contains no links! But this biotechnology fortified website is pure propaganda, lies and deceit. Just like the presidencies since the 1980’s and many of our U.S. Senators and Congress persons, legislators, and colleges and universities which received generous donations and bribes from the biotech-pesticide industry.

    • Karl Haro von Mogel

      It doesn’t appear that you put any links in your comment. Feel free to try again, 2 or more links might have your comment held by the blog for moderation, just to warn you, since at your first comment hiccup you decide to make huge assumptions about us. Biofortified does not accept money from the biotech industry.

      • Ewan R

        Ridiculously long posts also cause comments to be hung up – as I’ve discovered on a couple of occasions.

  • Kevin Folta

    Hi Becca,

    I’m curious, do you think that the science of climate change is all bogus too? Evolution? I’m always curious when people accept real science in some cases, yet deny it in others.

    Honestly, if I ever found evidence of transgenic plants being hazardous I’d publish it on the cover of Science AND Nature. I’d get grants for the rest of my career and start a whole new area of scientific inquiry. The best evidence just does not point that way. Lots of people worldwide are looking for such evidence and the best they can get are irreproducible studies in obscure journals.

    This is a technology that can help feed more people and have less impact on the environment. That’s what I’m in it for. Don’t be too short on your assumptions of my motivations as part of this website.

  • Scott

    Contaminate: to make impure or unsuitable by contact or mixture with something…And considering that all genetic engineered pollen is patent protected, the taint/contamination aka pernicious legal retribution is passed along to the contaminated plant and owner. The owner of said patent can unilaterally defend and sue everyone where their property exists. You’re damn right it’s contamination…of the pure natural order and of our individual constitutional rights.

    • Karl Haro von Mogel

      Scott, the precise language of questions matters a lot when it comes to social science research, including surveys. Using loaded terms such as “contaminate” that carry with themselves a value judgement will skew the results and make them less of an accurate description of what people think. So I completely disagree with your assessment of whether this word was used appropriately. Ironically, however, as I noted in the post above, despite the use of loaded terms, they still couldn’t get the result they were looking for – most people still didn’t care or cared little when the word “contaminate” was used.