One down...
If anyone understands what I just wrote, let me know...I certainly don't.
Two important phenomena occur when we are faced with the task of determining whether two visual stimuli differ from one another. First, when given the same number of features that may or may not differ, as the number of mismatching features - either area, shape or direction - of a stimulus increase, the reaction times for judging two different stimuli decrease. This results because a greater number of mismatching features, the sooner the first of them will be identified. Secondly, reaction times for different stimuli with a fixed number of different features yield longer reaction times when the number of relevant features increases. Sequential tests allow us to test why these phenomena occur because they are self-terminated as soon as a mismatch has been found and after following the search for certain features in a certain order, which is referred to as the “search path”. The mismatched features occur in random positions along this search path and have equal probabilities of occurring in a certain pattern and are therefore unpredictable by the subject. When presented with more relevant features to test, the mean number of tests required to find a mismatch increases. For example, when color is added to the set of relevant attributes of a visual stimuli, not only must mismatches in the area, shape and direction be checked, but it must also established that they are not mismatched, thus increasing the amount of time necessary until we get to the attribute of color to terminate the search process. Therefore the mean number of tests is dependent on the both the number of mismatching features and the number of relevant features and can be operationally demonstrated with a specified formula.
Another important consideration is the linear relationship between number of tests and reaction time. It is assumed that a feature testing process occurs in between the presentation of two visual stimuli and the response as well as residual operations including encoding the stimulus, deciding whether it is the same or different from the first, and organizing and executing the response depending on the prior decision. Hypothetically, the sum of these processes is equal to the response time. Another assumption underlying the relationship is that these processes occur in stages - that is, they occur successively. Furthermore, the idea that only one of the four stages in feature testing duration is influenced by a change in the level of the factors illustrates yet another assumption of selective influence. Lastly, we assume the idea of pure insertion - that no matter what the context - that neither where a particular test is inserted nor the number of tests, should change the tests. Sternberg notes that the relationship between the duration of feature testing and number of tests carries constraints when fitting the model to Bamber’s data; however, relaxing three of these constraints maintain the properties of the proposed model, yet relaxing the fourth constraint (that is, allowing unequal mean test - durations for different attributes) requires either assuming that search paths are equally probable or changing the constraints on test durations in order to fit the model to data. Thus far, the information presented in Sternberg advocates with caution the use of the sequential-test model to explain how we discriminate between visual stimuli, arguing that well-defined features of certain visual stimuli are analyzed to significantly aid in their comparison.