Why Aren't We Saving the Planet?: A Psychologist's Perspective - Geoffrey Beattie (2010)

Part I. Notes on attitude

Chapter 6. Uncovering implicit attitudes to carbon footprints

We created our own version of the IAT which compared the categories of high- and low-carbon-footprint products and the attributes ‘good’ and ‘bad’ and tested it on a random sample; a mixture of students and ordinary working people. To get an idea of the IAT procedure for yourself, try the following example. All you have to do is put the pictures and words that appear down the middle of Tables 6.1 and 6.2 into the categories that appear on the left hand side (‘Low Carbon Footprint or Good’) or the right hand side of the page (‘High Carbon Footprint or Bad’) as quickly as you can. In the normal IAT, items are assigned to categories on the left by pressing a key to the left of the keyboard (e.g. ‘Z’) or to categories on the right by pressing a key to the right of the keyboard (e.g. ‘M’). However, for the examples included here you can just tap either the left-hand side of the page or the right-hand side of the page. As in the experiment itself, you must try to do this as quickly as you can.

For example, the first item in Table 6.1, ‘Awful’, fits into the category ‘High Carbon Footprint or Bad’ because ‘Awful’ is clearly ‘Bad’, so tap the right-hand side of the page. (The underlying psychological reasoning here is that if you unconsciously think that high-carbon-footprint products are bad then this assignment should be relatively easy.) The second item, a Waitrose plastic bag, is clearly ‘High Carbon Footprint’ so here you should again tap the right-hand side of the page (‘High Carbon Footprint or Bad’). Again, the reasoning is that if you unconsciously think that high-carbon-footprint products are bad then this assignment should be

Table 6.1 Sample IAT procedure 1

Table 6.2 Sample IAT procedure 2

relatively easy. ‘Wonderful’ is ‘Good’ so now you should tap the left-hand side of the page, assigning ‘Wonderful’ to the ‘Low Carbon Footprint or Good’ category, and so on.

In Table 6.2, the category pairs have been reversed and we have ‘Low Carbon Footprint or Bad’ on the left-hand side of the table versus ‘High Carbon Footprint or Good’ on the right-hand side. The first item is a small eco-friendly car, so you should tap the left-hand side of the page assigning it to the ‘Low Carbon Footprint or Bad’ category because this small car is clearly ‘Low Carbon Footprint’. The psychological reasoning here is that if your underlying unconscious attitude is very pro-low-carbon-footprint products then this will be (relatively speaking) harder to do than the previous task when the categories were paired differently, because the generic category you are assigning it to (‘Low Carbon Footprint or Bad’) also covers words or concepts that are ‘bad’. The second item is ‘Superb’ so you should tap the right-hand side of the page (‘High Carbon Footprint or Good’) because ‘Superb’ clearly is ‘Good’. Again this should be more difficult than before, if you unconsciously do not think that high carbon products are ‘good’. The third item is a low-energy light bulb so you should tap the left-hand side of the page (‘Low Carbon Footprint or Bad’), which is relatively speaking (and again we are talking about milliseconds here) more difficult for most people than the earlier task when the categories were paired in the reverse manner. So go ahead and try it for yourself to work out your own implicit biases.

If you do (in terms of your unconscious attitude) associate low carbon with ‘good’ and high carbon with ‘bad’ then you should have found the first table easier to do than the second. On the second table you may well have noticed a slowing in your reaction time. If, on the other hand, you associate low carbon with ‘bad’ and high carbon with ‘good’ (perhaps because you feel that people are trying to force green issues down your throat) then you should have been faster when categorising items in the second table, and your responses should have been slower when you were categorising items in the first table (‘I was only getting started,’ I hear you say; the mind is after all great at rationalising!).

In Table 6.3, the blocks of trials used in the IAT (with the critical comparison trials outlined in bold) are shown:

The computerised versions of the seven trials are shown in Figures 6.1-6.7. This is what the participants actually saw on the computer screen in our IAT.

The D score (or difference score) is the critical measure used in the IAT. This is a statistical measure that calculates both the difference in the latency, or time taken to respond, in the critical trials and the error rate. The main point to remember here is: the more positive the D score, the more positive the implicit attitude to low-carbon-footprint products. The actual D scores we found are given in Table 6.4.

The results of our carbon footprint IAT revealed that 59% of our participants showed a strong implicit preference for products with a low carbon footprint. An additional 24% showed an implicit bias towards products with a low carbon

Figure 6.1 First trial: ‘Low Carbon Footprint’ vs ‘High Carbon Footprint’.

Figure 6.2 Second trial: ‘Good’ vs ‘Bad’.

footprint. 10% of people were neutral, showing little or no preference for either high- or low-carbon-footprint products, and 7% showed an implicit preference for products with a high carbon footprint, as shown in Figure 6.8.

Put simply, implicit attitudes would seem to be even

Figures 6.3 and 6.4 Third and fourth trials: ‘Good or High Carbon Footprint’ vs ‘Bad or Low Carbon Footprint’.

Figure 6.5 Fifth trial: ‘High Carbon Footprint’ vs ‘Low Carbon Footprint’.

Figures 6.6 and 6.7 Sixth and seventh trial: ‘Good or Low Carbon Footprint’ vs ‘Bad or High Carbon Footprint’.

Table 6.4 D scores from the carbon footprint IAT

Figure 6.8 D score percentages.

more biased towards low-carbon-footprint products than explicit attitudes. Overall 83% of participants showed some preference for low-carbon-footprint products compared to 70% or 67% on the earlier explicit measures. While our explicit measures suggested that 26% of our participants held neutral attitudes, the IAT measure suggested that only 10% of participants held a neutral view towards high- and low-carbon-footprint products.

What could possibly account for the decrease in the proportion of people holding neutral attitudes? I have one possible ‘methodological’ explanation for what is going on here. It could be due to participants showing positive implicit bias towards specific products and images whereas, of course, on the explicit measures, participants are forced to imagine an undefined set of products with either high or low carbon footprints and therefore participants may have sat on the fence more with these abstract concepts. It is a possible explanation.

Of course the IAT measures implicit attitudes to high-and low-carbon-footprint products generally, but one can also focus on the response times to the different types of product used in the experiment and this does highlight some interesting differences between the products. By breaking down the overall response times into the individual mean response times for each picture item used in the IAT, we can gain an insight into how quickly our experimental participants categorised each individual item. This new focus reveals that certain food items, particularly fruit, are readily classified as having a high or low carbon footprint, suggesting that these items are easily automatically recognised as having either good or bad environmental characteristics. Pineapples (exotic, have to be transported great distances, therefore high carbon footprint, and therefore bad) were categorised most quickly of all, followed by English apples (low carbon footprint) and locally grown blackberries (low carbon footprint). The same is true for items representing certain energy sources (like wind) and modes of transportation (like cars). Oddly and somewhat counter-intuitively, the bicycle (the counterpoint to the car) took a lot of time to categorise.

It is only really within the past few years that packaging and the bags we use have been highlighted as a major environmental issue and an area where consumers can make a real difference, and the results showed that reusable carrier bags were the slowest to be categorised in this experiment. This result might surprise many people since the reusable bag has become such a symbol for whole sections of society who wish to flaunt their social identity as primarily concerned with green issues. It may act as a potent (and conscious and deliberate) social signal but it does not seem to have the same automatic, unconscious impact that some people might imagine. The results would seem to indicate that people need a little bit more time to process these iconic representations and assign them to one of two categories.

Similarly, our participants were quick to recognise that certain kinds of light bulb were good, but they took much longer to recognise that the alternatives (normal light bulbs) were bad. They also needed quite a lot of time to think about beef and chicken (the beef shown was high carbon footprint because cows generally have a higher carbon footprint than chicken and much beef comes from overseas). Figure 6.9 shows the response times for each IAT item.

So what are the implications of all this? The implications would seem to be that people have the right attitude, both implicit and explicit (both conscious and unconscious) to low-carbon-footprint products. Since such attitudes, and their combination, set up a predisposition to act, then one might expect people to be doing much more for the environment in terms of their everyday supermarket shopping than they actually are. Green choices are becoming more popular, but not as quickly as some might have imagined. So there is clearly something else going on here, but what sorts of additional processes are critical here?

And something else is evident in these data, the first clear hint that people might say one thing but believe another. The research had shown that people, generally speaking, were very pro-low carbon in both their explicit and implicit attitudes, but a significant proportion of people showed a marked discrepancy between these two measures. They were

Figure 6.9 Response times to high- and low-carbon-footprint IAT items.

much more pro-low carbon in their explicit measure than in their implicit measure. This was the first hint in my data that some people might like to exaggerate their green credentials. They would report, when asked, that they were pro-low carbon products, and that, of course, they cared about the environment. The IAT, however, revealed something different. I was surprised to find (for all kinds of reasons) that I was actually one of them.