Georgi N. Nikolaev; Orlin Baev
Department of Psychology and Cognitive Science, New Bulgarian University
In two preliminary experiments the influence of mood and strategic priming on performance on the famous Raven matrices task was measured by using the thinking aloud methodology. Although the results did not reach statistical significance, differential effects on the usage of searching vs. exclusion strategy was observed in Experiment 1 and the extent to which subjects relied on superficial features vs. relations in Experiment 2.
In human thinking research the method of priming is widely used and often generates surprising patterns of performance without subjects being aware of the manipulation and its relevance. In the current experiments we explored the possibility for changing both the strategy employed and focus of processing in the Raven task by varying the preceding task which constituted the priming manipulation.
In Experiment 1 we examined whether solving a classic “odd one out” task could induce an exclusion strategy – first removing the inappropriate options – in the following Raven task. We reasoned that by appropriate priming we can increase the usage of this strategy as compared to a neutral control condition.
A simple between subject design was used with two levels of the independent variable – presence of an “odd one out” task before the main Raven task and a control level without priming.
Subjects were tested individually in sound proof booths by the second experimenter. They were instructed to think aloud and verbalize their thought process as they go along the problems. First all of them were given a practice task that was supposed to accommodate them to the uncommon in everyday life practice of thinking aloud. Following that procedure, half of the subjects were given three pictorial tasks such as: Which is the odd picture from these?
1. saw 2. axe 3. hammer 4. tree
The remaining half proceeded directly to the Raven task, which consisted of three matrices of medium difficulty. Verbal protocols of the procedure were collected.
10 NBU students volunteered to participate in the experiment. They were split equally in the two conditions and were rewarded nothing but unforgettable experience.
Results and discussion
After conducting the experiment, we coded the protocols using three categories – a random strategy, a search strategy and an exclusion, or odd one out strategy. Since the random strategy was rarely used and not really relevant to out study, we conducted T-tests to determine the significance of our results using only the search and odd one out strategies:
Although some specific trends were observed in the usage of the Odd-one strategy, our results remained insignificant. We believe this was caused by the small number of subjects and the usage of only one coder, and that if we increase the number of subjects to 30 and use two independent judges for greater validity, our results could show the predicted pattern.
In experiment 2 we used three other matrices from the Raven stimuli, which were chosen because of being rich in both details and relations.
The same simple one-factor between-subjects design was used. This time we had no control level. The two levels of the IV were positive and negative mood and our DV was the amount of statements focusing on details vs. relations. Our hypothesis was that in a positive mood subjects would tend to ignore the difficulty of the task and focus on the easily detectable superficial features while in negative mood they would rely on a more relationally-oriented strategy.
The principal difference with the procedure of Experiment 1 was the priming task. Subjects we asked to describe one of their best/worst days vividly and describe it to the experimenter, as well as briefly record it in written form. After a very short delay they were given the Raven task.
11 university students participated for no payment. Six were in the positive priming condition and the remaining five in the negative priming condition.
Results and discussion
We used only two pre-determined coding categories – features and relations.
Again as in Experiment 1 the expected patterns of performance was observed without obtaining statistical significance. We tend to attribute that fact to the small number of participants. Even if the coding of the protocols was unreliable because of the insufficient number of coders, we still believe that the predicted pattern is genuine and could be verified by using 30 or more subjects.
In two short thinking aloud experiments we observed the influence of different kinds of priming on subsequent intellectual performance. This is the place to make the remark that none of the total of 21 subjects expressed any suspiciousness during the procedure and nobody reported any conscious impact of the priming task on their target task performance. It would be interesting to merge the two experiments into a more complicated factorial design and observe the possible interactions between the factors. Before that, however, we have ideas for replications of the current designs using more subjects, coders, and importantly – eyetracking technology to aid us in the data acquisition process.