In another demonstration of ironic effects, Wegner and Ansfield (1994) asked participants to hold a pendulum and instructed them to make sure that the pendulum did not swing in a particular direction. Half of the participants were given a high mental load task to carry out, that is to count backwards from 1000 in intervals of 7, the rest were given no mental load task. They found that there was much more movement in the forbidden direction for those in the high mental load condition than those in the no mental load condition.
These two studies confirm Wegner's prediction that ironic effects will be more pronounced in situations when we have the distraction of a task with a high conscious mental load. Thus there have to be two general conditions for ironic effects to occur: first we have to initiate some sort of controlling behaviour, whether it be trying to control thoughts or our behaviour; and second we must have reduced conscious processing capacity for the Operating Process to use.
The following experiment is a direct test of the ironic processing theory suggested by Wegner and his colleagues. Much of the previous research on 'ironic processes' has investigated physical effects of such processes, however, the current study aims to investigate cognitive effects of ironic processes. The study will utilize a dual modality task to explore participants susceptibility to ironic effects. The basic idea is that participants are presented with two stimulus sources, one visual and one aural. One half of the participants are explicitly told to ignore one of the stimulus sources, i.e. the aural one, and to concentrate only on the visual source. The remaining participants are simply told that the aural source of information will be there and not told to ignore it. The aural source of information will be a list of words presented using an audiocassette player and a set of headphones. To find out if there is a difference between the two groups in their ability to ignore this information they will be given a memory task. They will be given the list of words, along with some that they haven't heard before and asked if they recognise any of the words as being on the list presented to them using the cassette player.
In such a task, if ironic processing occurs, it would be predicted that those participants who were told explicitly to ignore the taped words will recognise more of these words than those who were not told to ignore them.
Method
Participants
There were 56 people who took part in the study. The mean age of the group was 22.22, range of 18 to 45. There were 45 females and 2 males (gender was not recorded for two of the participants). The participants were all 1st year undergraduate students studying psychology at The University of East London and participated as a part of a Research Methods course.
Materials
A word list was generated which contained 96 words (e.g. sweating, decorate, isolated, calm). This list was then divided into two sub-lists, List A and List B, each one containing 48 words. These two lists were matched as accurately as possible for word length and word frequency using the norms of Carol, Davis & Richman (1971). Each word list (A & B) was then randomised and recorded onto audiotape, one word every five seconds. These lists were to be presented to participants in the encoding stage of the experiment. The lists were presented using a standard tape recorder with headphones. The full list of 96 words was also randomised and printed onto one sheet of A4 paper. These were the words that were used in the recognition task. The participants were simply asked to read through the 96 words and put a circle around all those words that they recognised as having been presented previously using the tape recorder. These instructions were printed on a separate sheet and attached to the word list.
A separate booklet was created which contained 24 pictures for the participants to look at during the encoding stage of the experiment. Eleven of the pictures were cartoons taken from a Far Side cartoon book (Larson, 1992). The remaining 13 pictures were black and white photographs taken from newspapers and magazines (see Appendix A for examples). The 24 pictures were put into a random order and presented in the booklet.
Design
The study was conducted as a between participant, experimental design. The independent variable was whether or not the participants were instructed to ignore the taped words. The dependent variable was an adjusted measure of each participant's recognition ability. Participants were randomly allocated to each of the two conditions and were presented with the words on the tape and in the recognition task in a fixed but random order. The recognition list was divided into two separate but equivalent word lists to be presented to the participants using the tape recorder. Half of the participants in each condition heard word list A and half heard word list B.
Procedure
The room in which participants were tested contained 15 tape-recorders with headphones attached. When the participants entered the room they were asked to sit at a desk which had a tape recorder on it. The participants were told that they were taking part in a study examining memory for pictures. They were then instructed to put the headphones on and listen carefully to the instructions given by the experimenter. The experimenter informed the participants that when the experiment began they would be asked to press the Play button on the tape-recorders. They were told that they would hear someone reading out a list of words and that this was designed to distract them from the other task they would be required to do. They were informed that they should look at the pictures contained in the booklet which was in front of them. They were told that they should look at the pictures carefully and try to memorize them as they would be tested later in the experiment. One group of participants were told that they should also try to ignore the words being presented by the tape recorder. Whereas participants in the other condition were told that they would probably be distracted by the words but that this did not matter. They were not explicitly told to ignore the words. Once everyone had indicated that they had understood the instructions they were told that the task would last for 6 minutes and the experiment was commenced. Upon completion of the encoding task the experimenter asked the participants to remove the headphones and read the instructions on the front cover of the recognition booklet. They were informed that they would have five minutes to go through the booklet and put a circle around all the words that they thought they recognised as having been presented earlier by the tape recorder. Upon completion of this recognition task the participants were requested to assemble in another room where they were debriefed by the experimenter.
Results
In order to take account of guessing in the recognition task, the number of false alarms was subtracted from the number of correct responses to obtain an adjusted recognition measure for each participant.
The adjusted recognition scores for each group were subjected to exploratory data analyses using box plots.
Figure 1. Box plots of the adjusted recognition scores for the ‘Ignore’ and ‘Not Ignore’
The box plots presented in Figure 1 suggest that the adjusted recognition data for both the Ignore condition and the not ignore condition were approximately normally distributed. The distribution in the not ignore condition is if anything very slightly negatively skewed but not enough to suggest any systematic violation of the assumption of normality of distribution.
The mean and standard deviations for the adjusted recognition scores for the two conditions are presented in Table 1. This shows that the participants in the Ignore condition performed slightly better in the recognition task than those in the Not ignore condition
Table 1. Mean and standard deviations for the adjusted recognition measures for the participants in the 'Ignore' and 'not ignore' conditions
The adjusted recognition scores were analysed using a Wilcoxon Rank Sum test. This revealed that there was no significant difference between the two groups that could not be explained by sampling error (W=727, N= 56, p= .81), thus providing no support for the experimental hypothesis.
Discussion
The analysis of the adjusted recognition scores showed that there was no difference between the ignore and not ignore conditions beyond that attributable to sampling error. This clearly represents a failure to support the experimental hypothesis.
The findings have also failed to support the ironic processing theory proposed by Wegner (1994). Wegner suggested that ironic effects are most likely when participants are given high mental load conditions. This was achieved in the current study by informing participants that they were to memorize the pictures in the booklet because their memory would be tested later in the study. This clearly represents high mental load as defined by Wegner, however, the ironic effects that were predicted in this study failed to materialize. It could be that the experimental design used in the current study has highlighted conditions under which ironic effects do not occur. The current study used two tasks which require conscious cognitive effort. One of the tasks, listening to and remembering the words requires that the words would have to be consciously processed in order for them to be remembered. The distraction of requiring participants to consciously attend to and try to memorize visual stimuli was obviously competing for cognitive resources with the verbal stimuli. The current study thus suggests that ironic effects are not as likely if the task that is trying to be inhibited or avoided also requires high levels of conscious cognitive effort. In contrast, the previous work into ironic effects has not used task which compete so much for cognitive processing resources. For example, Ansfield et al. (1996) looked at sleep instructions whilst playing arousing music. Falling asleep is not so much a conscious cognitive task unless you are deliberately thinking about strategies to help you fall asleep. Therefore, in such a study there would be less interference from the two tasks. A similar conclusion can be applied to the pendulum study of Wegner and Ansfield (1994).
Before accepting the above interpretations however, a number of issues need to addressed. The first problem with the current study is highlighted by the large variability in recognition scores within each of the two conditions (see Table 1 for standard deviations). The large variability along with the relatively low scores within each group suggests that the participants were to a large extent guessing. This in turn suggests that the participants heard few of the words and thus were unlikely to produce ironic effects. If this is the case then the current design represents an inadequate test of Wegner's theory. Perhaps a better test of the theory would have been to give the participants fewer words to remember and have the words presented more than once. The memory test as it stands is probably too difficult.
An interesting possibility for future research might be to assess the effects of anxiety on the likelihood of ironic effects. It has been well documented by Eysenck (1992) that anxiety restricts the amount of conscious cognitive processing capacity available to people. This thus represents a naturally occurring situation of high mental load. In the current study high mental load was induced by requiring participants to memorise pictures. It may perhaps be better to have naturally occurring instances of high mental load. It would thus be of theoretical interest to test participants for ironic effects when they are in an anxious state and see if such effects are more prevalent than at times of low anxiety.
In conclusion, the current study has shown that there was no real difference between the ignore and not ignore groups in terms of their ability to recognise the words presented to them on the tape recorder. This represents a failure to support the experimental hypothesis and the ironic processing theory proposed by Wegner (1994).
References
Ansfield, M.E., Wegner, D.M., & Bowser, R. (1996). Ironic effects of sleep urgency. Behaviour Research and Therapy, 3,7, 523-531.
Carrol, J. B., Davis, P., & Richman, B. (1971). The American heritage word frequency book, Houghton Mifflin, Boston.
Eysenck, M.W. (1992). Anxiety: The Cognitive Perspective, LEA, New York.
Larson, G. (1992). A Far Side Collection: Unnatural Selections, Futura Books, Boston.
Wegner, D.M. (1994). Ironic processes of mental control. Psychological Review, 101, 1, 34-52.
Wegner, D. M., & Ansfield, M. (1994). Magnification of the Chevereul pendulum illusion under mental load. Manuscript in preparation.