I have no freaking idea what this means:

Researchers identified 72 female students who said they favored voluntary euthanasia. Researchers then gave orange juice to these subjects, but half of them got juice spiked with caffeine. The students then read a series of arguments against voluntary euthanasia. An after study showed that the subjects receiving the caffeinated juice remembered more of the arguments AND were more likely to shift towards anti-voluntary euthanasia views. Similar results obtained in a study of 76 males.

["Coffee for Persuasion, "The Chronicle of Higher Education, July 7, 2006, A17; thanks Timothy Murphy UIC]

More seriously, I guess it’s not surprising that short-term memory retention would be affected by neuroactive drugs – though it’s a bit worrisome that such an ubiquitous one would have such a notable effect. Taking a total wild-ass guess, I would assume that the change in position is a function of the greater retention of the material – that is, that ingesting caffeine doesn’t inherently make you anti-euthanasia, but rather that a differential retention of arguments specifically against that position, caused by the drug, would then tend to skew respondents’ answers in that direction simply because they then had more such arguments in their heads.

Looking further, the original report is here. The situation is more complicated than explained in the blurb above: The students were selected for having opinions favorable to voluntary euthanasia (so the researchers could test the affect of the reading on changing their opinions). They were not just told to read the articles about euthanasia, but were divided into groups and given one of two tasks: either a mechanical editing chore or a specific instruction to read the articles carefully and consciously try to remember the arguments they used; they were then tested on retention and the affect of the articles on influencing their opinions. They were then given counter-messages (articles in favor of voluntary euthanasia) and re-tested on the degree to which receiving the counter-messages undid the change in opinion they had undergone from reading the original arguments.

The results indicated that reading the first (anti-euthanasia) arguments had no effect on opinion or retention, with or without caffeine, for the students who were given the simple editing task without being told to concentrate carefully. However, reading the first argument with careful concentration did improve both retention of the arguments and a change in the students’ opinions; the effect was present in both the caffeinated and the no-caff groups, but it was greater with caffeine. Then, after reading the counter-arguments, students who did not take caffeine reverted their opinions back to their original opinions, but students who did take caffeine were not affected by the counter-messages and retained the new opinions they had adopted after reading the first arguments. (A summary of the report appears below.)

The researchers attribute this to the differential affect of initial arguments and counter-messages: apparently, there is a theory in psychology that when people are exposed to new information, they tend to favor the first message they hear, which sets up a defense in their minds against a counter-message that they hear afterwards. Interestingly, in the above experiment this effect was not observed for the non-caffeine group, but the caffeinated group did show a defensive effect against the counter-message. This seems to me just as important a result from this experiment as the basic effect of the caffeine itself.

The results seem to suggest that caffeine not only aids retention of information (that one is consciously processing already), but somehow fixes it more firmly in the mind or increases one’s susceptibility to being swayed by it. The first part doesn’t seem so startling, but the latter is, to my inexpert perspective at least. It would be interesting to repeat the experiment with, say, bioethicists or others well-versed in the issue, to say whether their opinions would be more vulnerable under caffeine than those of students presumably reading about the issue seriously for the first time. I would predict the professors would not change their existing opinions even with the caffeine boost; if they did, that would suggest that the caffeine not only increases receptivity to new messages but somehow overrides existing strongly held opinions (a result that, frankly, I hope is not the case).

Report Summary

Each participant consumed 330 mls of a commercially available sweetened orange juice.

In a randomised, double-blind procedure, half of the drinks contained anhydrous caffeine at a concentration of 3.5 mgs/kg of body weight and half had nothing added (placebo condition). Participants were asked to drink the juice as quickly as possible. To allow for the maximum absorption of caffeine, participants waited for 40 minutes before continuing with the study. This absorption time interval is optimal for caffeine to peak in the blood (Maisto, Galizio, & Connors, 1991; Reynolds, 1996). Third, participants were informed that there was a debate regarding voluntary euthanasia in various communication mediums (such as newspaper, radio, television). As a cover story, they were told prior research had shown that students at their university were divided in their attitudes towards voluntary euthanasia—some were in favour of voluntary euthanasia, whilst others were against. Before participants read the arguments the extent of messagedirected processing was manipulated. In the low message processing condition, participants were asked to go through each word in the message and cross out every letter ‘o’. The aim of this task was to encourage only a superficial consideration of the content of the arguments contained in the message. In the high message processing condition, participants were asked to read the arguments and try to remember them because they would be required to recall them at the end of the study. This latter instruction was designed to encourage systematic processing of the arguments. As a persuasive message, six convincing narrative arguments against voluntary euthanasia were then presented. Such a message has been successfully employed in other studies of attitude change (see Martin & Marrington, 2005; Martin & Martin, in press; R. Martin, Gardikiotis, & Hewstone, 2002; R. Martin et al., 2003). Fourth, after reading the message, participants completed a thought-listing task and the same 9-point scale measuring their attitude towards voluntary euthanasia as in the first session (post-test I: initial message). Participants were then asked to recall as many of the arguments in the message as they could. Finally, participants were given six strong arguments that argued the opposite position to the initial message, i.e. pro-voluntary euthanasia. The initial and counter-messages both contained the same number of arguments and were of similar word lengths. The participants then completed the same attitude scale as that following the initial message (post-test II: counter-message). At the end of the study participants were asked to read the thoughts they had listed following the initial message and to indicate whether each thought was in favour, against or neutral towards voluntary euthanasia. . . .

Analysis of simple main effects showed that participants did not change their attitude towards the messages when message processing was low in either the placebo (Ms=2.47 vs. 3.00 vs. 3.00), (2, 112)=1.17, n.s., or caffeine (Ms=2.67 vs. 3.33 vs. 3.53), F (1,56)=2.53, n.s., conditions.

However, participants did change their attitude towards the messages when message processing was high in both the placebo, F (2,112)=5.93, p< 0.004, and caffeine F(2,112)=31.75, p< 0.001, conditions.

Focusing first on attitude change to the initial message (pre- vs. post-test I), there was a reliable change in attitude towards the initial message when message processing was high in both the placebo (Ms=2.40 vs. 3.73), F(1,56)=13.18, p< 0.001 and caffeine (Ms=2.40 vs. 5.33), F (1,56)=63.77, p<0.001, conditions. However, the amount of attitude change in the caffeine condition (Mdifference=2.93) was reliably larger than that in the placebo condition (Mdifference=1.33), F(1,56)=19.04, p< 0.001.Focusing on attitude change to the counter-message (post-test I vs. post-test II), in the placebo condition the reliable change in attitude to the initial message (Ms=2.40 vs. 3.73), yielded to the second counter-message (Ms=3.73 vs. 2.73), t(56)=5.19, p< 0.03. Indeed, there was no reliable difference between pre-test and post-test II showing attitudes had returned to their original level (Ms=2.40 vs. 2.73) t(56)=0.69, n.s. However, in the caffeine condition the reliable change in attitude to the initial message (Ms=2.40 vs. 5.33), resisted the second counter-message (Ms=5.33 vs. 5.00), t(56)=0.58, n.s. The difference between the pre- and post-test II attitude scores (Ms=2.40 vs. 5.00) was reliable, t(56)=42.25, p< 0.001.>