19 Experiment Stimuli
19.1 Generate engaging stimuli for your experiments.
ChatGPT’s potential for generating stimuli and materials for experiments is as considerable as its capacity for creating rich content for coursework. Currently, we are harnessing this AI tool to create intriguing insight puzzles and vignettes for fictional brands, thereby paving the way for more engaging and efficient experimentation.
To illustrate, one of my PhD students spent a considerable amount of time developing very specific anagrams. These were designed to hit just the right level of difficulty, comprised of 8-13 letters, and associated with a two different semantic categories. Now, with the assistance of ChatGPT, we can generate similar materials within seconds.
In another experiment, another student used written vignettes focusing on four pro-environmental concepts: electricity conservation, paperless options, recycling and reusing plastics, and vegetarianism. Each concept was depicted through two sets of vignettes. One set presented a wide range of unique and variable arguments, and the other concentrated on consistent, pro-environmental arguments. The creation of these vignettes took several weeks, as we meticulously ensured they covered aspects like environmental benefits, personal health benefits, and economic benefits. All the while, we had to minimise confounds and noise related to differences in content length, framing, argument structure, language simplicity, and the presence of numerical estimates. The aim of the experiment was to evaluate the effect of the breadth versus depth of information on participants’ trust and intentions towards pro-environmental behaviours. If we had employed ChatGPT during this experiment, the generation of these vignettes would have taken mere minutes instead of weeks.
The generalisability crisis, as outlined by Tal Yarkoni (2020), brings into focus the mismatch between verbal hypotheses and their corresponding statistical expressions in psychology. This crisis casts doubt on the validity of generalisations made by researchers. In this light, ChatGPT’s transformative capacity for generating experimental materials becomes even more crucial. It enables the creation of unique stimuli for each participant, even on each trial, thereby directly addressing the generalisability crisis. With ChatGPT expediting the generation of experimental materials and augmenting the ability to produce rich materials, researchers can redirect their focus towards designing experiments that more closely align verbal hypotheses with their statistical expressions. In doing so, they can address the generalisability crisis more effectively and fortify the intellectual rigour and reliability of psychological research.
