For those wanting a little more detail, this page outlines some of the more important elements of my PhD Dissertation research design.
The key fact that could not be discussed while the experiment was running was the primary dependent measure, which was verbal memory. It could not be discussed because in keeping with the naturalistic design of the experiments, memory was tested incidentally, meaning that the participants could not be aware their memory would be tested until just prior to the test. In addition to functioning more like everyday memory, this procedure also controls for potentially different degrees of rehearsal across participants.
| 
Integration illustrated. 1=Relation between Person and Environment, achieved through environmental exploration. 2=Relation between Environment and Task, achieved with the word hiding task. 3=Relation between Person and Task, achieved via word choice. |
Therefore, the task that was or was not experimentally tied to the environment and/or the participant, in an attempt to create a more or less contextually integrated situation, was a verbal learning task that was disguised to varying degrees as a randomly occurring verbal task meant to intrude upon the ostensible main task of exploring and rating the environment. In other words, all participants were instructed to explore and rate the garden environment, a task that was interrupted for the performance of the verbal task. The nature of that verbal task, and its relation to the exploration task, were the principle sources of variation.
Thus, a task requiring the creation of a random story using the provided list of 5 words was considered a less contextually integrated task than the imaginary hiding of each word in places in the immediate surroundings, with a verbalized rationalization for the hiding location (thereby creating an association between the word and the environment). Further integration (between the person and the verbal task) was manipulated by having some participants choose their words from a larger set, while others received the same "yoked" choices in a manner that made them appear random. The details on how that was accomplished are covered on the Methods page.
The simulated environments component of this project, mentioned in other portions of this web-site, was designed to explore the possibility of inducing the same type of person/environment/task integration with the use of a mere visual representation of a place. If comparable results can be obtained in a simulation, then there would be a possibility for the creation of more elaborately integrated situations. Preliminary data suggest that similar ratings were given in response to the simulated versions of the same garden environment used in the rest of the project, but there were not enough memory data collected to address the integration issue.
The results of this research cannot be properly discussed without more in-depth information than is appropriate here. So as not to completely disappoint the curious reader, the following is the briefest possible summary: There is some indication that the task used to create high "contextual integration" contributed to better incidental memory performance, but the engaging nature of the hiding task cannot be eliminated as a simpler (though not competing) interpretation. Less equivocal was a result showing that participants pre-testing as comparatively high in the personality construct "field dependence" had better memory for words if they were presented in the context of the highly structured hiding task as opposed to the storytelling task, which required more self-direction and involved a fair amount of ambiguity.