How should we study the mind?
In Precursors to Cognitive Psychology we saw that psychologists adopted a number of methodologies throughout the late nineteenth century and the twentieth century, including focusing only on observable behaviour. One early methodology was introspection, which was used to study the elements of perception, although this ran into difficulties. However, for higher-level cognitive processes such as decision making and voluntary behaviour, many social scientists continue to ask people about the reasons for their own behaviour. In a key paper published in 1977, and titled Telling more than we can know: Verbal reports on mental processes, Richard Nisbett and Timothy Wilson reviewed several studies that appeared to show that people are sometimes unaware of the reasons underlying their own behaviour.
In one study, Nisbett and Wilson observed consumers choosing between identical packs of nylon stockings arranged in a row. They found that people were biased towards choosing from the right side of the array but, when asked about their choice, people did not refer to the position of the stockings in the array. When specifically asked whether they might have been affected by the position of the item, almost everybody denied this. Based on this and the results of many other studies, Nisbett and Wilson concluded that people's explanations of their own behaviours were not based on having access to the underlying cognitive processes, but rather were theories about the causes of those behaviours. In other words, people may be aware of the products of their cognitive processes but are not aware of those processes themselves.
Introspections may not always be wrong, but Nisbett and Wilson's analysis indicates that they are not a reliable guide to the causes of behaviour. They may be useful in generating hypotheses, but these should then be subject to more rigorous testing. The nature of empirical investigation is the subject of the rest of this section.
The scientific method
The scientific process can begin with a simple observation or it may result from a deliberately-designed study in order to test a specific hypothesis. A set of observations, once analysed, may lead a researcher to develop a theory, which is an organised set of principles that together explain the data. From a theory, a researcher may then derive specific hypotheses, which are statements about what might be observed under particular circumstances. To test these hypotheses the researcher then designs research studies that will set up those circumstances. The resulting data are collected and analysed, and conclusions are reached about whether the data support or do not support the theory. It is important to note that the scientific notion of a "theory" is often different from the way the term is used in everyday speech. People sometimes use the word "theory" to refer to a vague idea or a guess about something, but this is not the same as the scientific conception given above - that a theory is an organised set of explanatory principles.
Another important aspect of the scientific concept of a theory is that it should meet Karl Popper's criterion of falsifiability. This states that any theory, to be considered scientific, must be described in such a way that there is a potential to show it to be false. Sigmund Freud's proposal of denial as a psychological defence mechanism is widely considered to be unfalsifiable. If a patient makes statements that appear not to be consistent with the analyst's interpretation of their condition, then these statements can be explained away in terms of the patient's denial to face up to their condition. Thus, the analyst's interpretation is put out of reach of empirical testing. Although philosophers continue to debate exactly how science does and should proceed, the basic notion of falisifiability continues to be very powerful.
Research designs in cognitive psychology
The following sections give a brief sketch of some different methodologies. Further information can be found - among other places - at Jeremy Miles's wiki for Research Methods in Psychology.
In a true experiment, a researcher manipulates a variable in order to see what effect it has on another variable. For example, suppose I want to know whether background noise affects performances on maths problems. One way of studying this would be to take a group of people and randomly assign them to two different groups, a no-noise group and a white-noise group. The first group is asked to solve maths problems in a quiet environment and the second group tries to solve the problems whilst being exposed to white noise. In this case, the presence/absence of white noise is referred to as the independent variable. Our outcome measure is referred to as the dependent variable. In this case, we might decide that the dependent variable is the number of correct answers on the test. Alternatively, we could measure how long it takes to produce answers. It is quite likely that we might actually measure both. Yet again, because there might be a trade-off between time to complete and accuracy of answers, we might set a time limit for the test and only measure accuracy.
The random assignment of participants, and the ability to include variables of interest whilst exluding many unwanted factors, mean that the true experiment is a particularly powerful kind of design. In the example given above, if we found that people performed considerably more poorly in the white-noise condition then we could have quite a high degree of confidence that this was indeed due to the presence of white noise.
Not all experiments involve the comparison of different groups. For instance, in the example above we could have used a single group of people, but asked them all to take part in the two conditions of the study. The two types of design are referred to as between-subjects and within-subjects, respectively.
The following video does not show a true experiment, as such, but does illustrate the importance of a researcher maintaining control over a situation:
Question: What do you think might be a potential drawback(s) of the true experiment?
As we have seen, true experiments often involve the comparison of two or more groups, where people have been randomly assigned to those groups. The quasi-experimental design refers to a situation where the researcher is studying naturally occurring groups. For example, if a researcher wishes to compare men's and women's performance on maths problems, then he or she is unable to actually manipulate the sex of the participants. All that can be done is to obtain some men and some women and then compare their maths performance.
In the correlational design the researcher is interested in a possible relationship between two variables, neither of which is under the researcher's control. For example, I might hypothesis that IQ scores and working memory capacity are positively related (as indeed they are). I cannot easily manipulate people's IQs or memory capacities, so the best thing for me to do is to take a sample of people, measure both variables for each person, and see if the two tend to be related to each other. One practical advantage of this type of design is that the researcher does not have to worry about the random assignment of participants.
Question: What do you think is a key disadvantage of the correlational design?
Some researchers investigate the relationship between cognition and the brain's structures and activities. This is psychobiological research. One way of looking at such relationships is to conduct post mortem studies, to compare the brains of normal individuals with those who were known to have some kind of cognitive deficit. For example, Paul Broca's patient Tan was so-called because he had a speech deficit so severe that this was the only syllable he could utter. In 1861, following Tan's death, a post-mortem autopsy of his brain revealed a lesion in the left frontal lobe, now known as Broca's area. Broca's original report can be read here, though it should be noted that the contributions of an earlier researcher, Marc Dax, have been unfairly neglected by history.
It is rather difficult to probe the brains of living people, but an alternative is to use animals. In one piece of Nobel-prize winning research, David Hubel and Torsten Wiesel inserted microelectrodes into the brain of a cat and were able to record the electrical activity of individual cells in the visual cortex. However, whilst this kind of study cannot normally be done on human participants, electrical activity can be studied using electroencephalograms (EEGs). In this method electrodes are placed in various locations across the scalp, and electrical activity is recorded over a period of time.
A number of methods can be used for obtaining static images of the brain's structures, the best-known of these being magnetic resonance imaging. Other imaging methods are based on the increased consumption of oxygen and glucose that occur in active areas of the brain, and include Positron Emission Tomography (PET) and Functional Magnetic Resonance Imaging (fMRI).
Process-tracing studies aim to record the focus of a participant's attention across time, often when working on a particular task. One form of process-tracing is protocol analysis, whereby a participant is asked to continually think out loud as he or she works on some task; these verbalizations are recorded, transcribed, and coded, and then analyzed by the researcher. The transcribed verbalizations represent a record, though not necessarily a complete one, of the participant's focus attention at various times, and may indicate the complexity of thought. Other forms of process tracing include eye movement tracking, computer mouse tracking, and diaries.
Case studies are intensive investigations of individuals, usually people of exceptional ability or people with some sort of deficit. One of the most famous case studies in cognitive psychology is Henry Gustav Molaison (1926-2008), until his death known only as HM. Following an operation on his brain in 1953 to relieve epilepsy, HM was unable to form new memories. For the rest of his life HM took part in research studies designed to help understand the nature of memory. One finding that resulted from the study of HM was that he could learn new motor skills, but every time he did a previously-practiced task he thought he was doing it for the first time. This is evidence for different memory stores. Although much-studied, HM did not feature in the media and it was not until 2007 that his voice was heard by the public (for the NPR radio broadcast click here).
Another individual who is unable to form new memories is Clive Wearing. Unlike HM, Wearing has been the subject of a number of TV and radio programs. There are many clips of Clive Wearing on YouTube, a couple of which are shown below:
Another methodology open to researchers is to observe people in real-life settings, such as at home or at work. Observations may be done with the knowledge and consent of those being watched, or they may be covert, in which case people are not aware that they are being watched. The latter type of observation obviously requires the researcher to give particular thought to ethical considerations.
Question: What do you think is a key advantage and a key disadvantage of naturalistic observation?
Computer simulations and artificial intelligence
As used in cognitive psychology, computer simulations aim to imitate aspects of human functioning. A particular cognitive theory may be implemented in a computer program. If the program runs successfully and produces outputs that resemble human responses, then we might conclude that the theory is coherent and plausible. By contrast, researchers in Artificial Intelligence (AI) aim to develop computers that can behave in an intelligent fashion, but not necessarily by implementing a theory of human cognition. However, the immense challenges faced by AI researchers have often been highly informative for cognitive psychologists and cognitive scientists.