Treisman[ref.4] produced a modified model that allowed for several channels to be analysed; information that was not immediately relevant was not lost but attenuated. A semantic "dictionary" then decoded meanings according to importance and pertinence. It was implicit to this theory that all input was semantically analysed in order that its relevance could be evaluated.
These models were problematic however, particularly in explaining how a physical filter can act upon meaning. This was addressed by Kahneman's capacity model[ref.5]. Kahneman redefined attention as "mental effort", limited resources being allocated according to momentary intentions (tasks related to current goals) and enduring dispositions (responses to other important stimuli). Furthermore, Kahneman believed that some tasks required little processing as they were overlearned, automatic skills. Thus, several activities might share limited cognitive resources.
In a series of studies, Shiffrin and Schneider[ref.6] identified certain basic properties of these automatic processes: they are relatively free of capacity limitations, operate in parallel, require considerable training, are difficult to "unlearn", and are unconscious actions. Reading is such a task; it takes a significant amount of practice but eventually becomes automatic. As an indicator, Healy[ref.7] found that subjects had difficulty in counting the frequency of letters in phrases containing common words such as "the". This recognition of whole words, rather than individual letters, is a result of an overlearned, automatic process.
J. Ridley Stroop[ref.8] had described the difficulty in naming ink colours of words that were themselves the names of colours. It appeared that the meanings of the colour words were interfering with the perception of ink colour. Shiffrin and Schneider explained that reading is automatic and not always under voluntary control and so intrudes on an attended task.
The rationale for this experiment is to retest the Stroop effect, and to quantify the intrusion of automatic, unconscious semantic processing. The experimental hypothesis is that the time taken to identify the ink colours of a list of colour names will be longer than that of a control list of neutral words. The null hypothesis is that there will be no difference in the processing time for the two lists.
Method.
Design.
This was a one-tailed, related study. The independent variable was represented by two conditions; two word lists printed in various coloured inks. One list consisted of colour names and the other of neutral words. The dependent variable was the time taken to name the ink colours. The number of errors in identification was also recorded.
Participants.
Sixteen natural scientists volunteered to take part in the study. The panel consisted of eight females and eight males between the ages of 20 and 50. All were educated to 'A' level standard or higher, and all were fluent speakers of English. To the best of my knowledge, none suffered from any visual impairment or colour interpretation difficulties.
Apparatus and Materials.
Two lists of 24 words were prepared on separate A4 sheets. One list contained the names of four colours; red, blue, green or yellow. The other was a list of four matched neutral words; run, boat, glass and yankee. Each word appeared six times in its list, the order being randomised. Each word was printed in one of four colours. The colours were randomly distributed between the words, but the same ink colour sequence was followed for both lists. All words were printed in 18pt Arial Bold, on white matt paper.
Procedure.
Participants were invited to sit at a desk in a quiet, well illuminated room and were asked to read a set of instructions. They were asked if they fully understood these and a practice list of 24 words was then placed, face down, on the table in front of them. This list was prepared using words from both the test sheets. When ready, the participants turned the paper over and proceeded to name the colours of the ink for each word in order. A timer was used to measure their performance, and a record was kept of the accuracy of their replies. They were given 30 seconds rest before the main test. This consisted of the first list, followed by a 30 second rest, followed by the second list, all in the same manner as before. The order of presentation of the lists was randomised.
Results.
The times taken by each of the 16 subjects to perform the two tasks are shown in appendix C and a summary in table 1.. The experiment was structured as a related study and the data was subject to analysis using the Wilcoxon signed ranks test. The rationale for using this design was that it would detect any significant differences between the involuntary responses of the participants to the two experimental conditions, also considering the limited number of data subjects available which would be insufficient for a conclusive unrelated structure. The analysis showed a very significant increase (p<0.01) in the times to process the colour word list. The number of errors was small and limited to the colour word list.
Table 1. Mean time and errors for the two experimental conditions.
Discussion.
The results showed a significant increase in the time taken to read the colour words over the neutral words. This is in line with previously reported data[ref.8] and supports the experimental hypothesis of this study.
The data used in this study was limited. Stroop used 100 subjects for a comparable experiment and measured a 75% increase in the time to report the ink colour of the colour words. A more limited replication using 40 subjects[ref.9] reported a 70% increase. The reason for the lesser value of 27% in this study could be linked to the only major procedural difference; Stroop and MacLeod used colour squares rather than neutral word as a control condition. From this, it might be deduced that the neutral words used in the current experiment were themselves intruding; a rerun using colour squares as an extra treatment could resolve this dilemma.
The Stroop effect contradicts the "bottleneck" theories such as Broadbent's because, if channels were selectively filtered, information on a second channel would be ignored and there would be no conflict. The results here indicate that involuntary semantic processing is taking place. Intrusion such as this is consistent with the automatic processes of Kahneman's model, and with the definitions of automaticity cited by Shiffrin and Schneider. The act of reading is so well learned that, despite attempting to attend to the colour of the ink, the unconscious process of interpreting the meanings of the words is still going on. When the word is itself a colour, the meaning of the word conflicts with the subject's attention to the ink.
The reason for an increase in reading time is unclear. It was observed that some of the participants seemed consciously to slow down when reading colour words, possibly to provide more attention. Others would read both lists at the same pace but slip up; realising their mistake, they would take up time in correcting it. Debriefing comments by the participants confirmed this view. It is possible that a longer period of practice would have allowed the participants to customise their response times more accurately to the task, thus reducing errors. On a few occasions, uncorrected mistakes were made; all but one subject were unaware of these.
The fact that most participants were oblivious of their mistakes indicates the practical consequences of the failure of selective attention in real life events. An example is the complexity of the tasks performed by air traffic controllers was reported by Roth[ref.10] who pointed out the need for them to monitor spoken and visual messages at the same time; intrusion into a current task by meaningful input on an unattended sensory channel could lead to errors.
Conclusion.
The experiment found that unconscious semantic processing of words on an unattended channel was intruding upon a task of naming ink colours. This was consistent with the Stroop effect. The extent of the effect varied from previous studies, being dependent on the neutrality of the control stimulus.
Raw data.
References.
-
1. Broadbent, D. (1958). Perception and communication. Oxford: Pergamon.
-
2. Cherry, E.C. (1953). 'Some experiments on the recognition of speech, with one and two ears'. Journal of the Acoustical Society of America, vol. 25, pp. 975-9.
-
3. Gray, J.A. and Wedderburn, A.A.I. (1960). 'Grouping strategies with simultaneous stimuli'. Quarterly Journal of Experimental Psychology, vol. 12, pp. 180-4.
-
4. Treisman, A. (1964). 'Monitoring and storage of irrelevant messages in selective attention'. American Journal of Psychology, vol. 77, pp. 206-19.
-
5. Kahneman, D. (1973). Attention and effort. Eaglewood Cliffs, NJ: Prentice-Hall.
-
6. Shiffrin, R.M. and Schneider, W. (1977). 'Controlled and automatic human information processing: II Perceptual learning, automatic attending and a general theory. Psychological Review, vol. 84, pp. 127-90.
-
7. Healy, A.F. (1976). 'Detection errors on the word the: evidence for reading units larger than letters'. Journal of Experimental Psychology: Human Perception and Performance. Vol. 2, pp. 235-42.
-
8. J. Ridley Stroop (1935) Studies of Interference in Serial Verbal Reactions. Journal of Experimental Psychology, vol. 18, 643-662
-
9. MacLeod, C.M. (1986). A replication of the classic Stroop effect: Experiment #2. Unpublished raw data at http://www.cgl.uwaterloo.ca/~bgbauer/chapters/stroop.html
-
10. Roth, I (1990). Introduction to Psychology, vol. 2, p533. Milton Keynes: Psychology Press.
-