I have proposed a 2-process explanation of the task interference or cost that occurs when a prospective memory task is embedded in an ongoing activity (Guynn, 2003, 2008). According to that explanation, the cost arises because individuals adopt a retrieval mode (a sustained readiness to encounter the targets that indicate that it is appropriate to perform the intended action) and engage in target checking (a transient process of evaluating whether an environmental stimulus is a prospective memory target). This cost is evidenced by worse (slower) performance on the ongoing task when there is, versus when there is not, an embedded prospective memory task.
My current research involves evaluating the involvement of these processes using a 4-quadrant paradigm that I developed for this purpose (Guynn & Underwood, 2014). Using lexical decision as the ongoing activity, letter strings are presented one at a time in a different random quadrant or corner of the computer screen. In blocks of control trials, lexical decision is the only task to be performed; there is no embedded prospective memory task. In blocks of experimental trials, a prospective memory task (e.g., press the Enter key if you ever see an animal word) is embedded in the ongoing activity. Moreover, experimental participants are told that if an animal word occurs, it will appear only in one particular quadrant (e.g., the top right). Thus, there are two different types of experimental trials – non-target trials, which are those for which a letter string appears in one of the three non-target quadrants, and target trials, which are those for which a letter string appears in the target quadrant. Note that any letter string (not necessarily an animal word) may appear on a target trial; it is perhaps best thought of as a “possible target” trial.
Research in our laboratory currently involves testing this 2-process explanation using the 4-quadrant paradigm and exploring the factors that impact the overall cost on the ongoing activity as well as the estimates of retrieval mode and target checking.