Friday, February 09, 2007

Flying Low with NASA

The bizarre story of NASA astronaut Captain Lisa Nowak has everyone talking. It's raised some predictable and reasonable questions about psychological assessment of astronauts. In a storm of comments at a New York Times blog, The Lede, several commenters correctly observed that this incident is the first incident, and that's a pretty good record. (Of course, that's assuming that it is the first incident. Others may have been covered up.) After an incident like this, however, we should still ask if the assessments are good enough.

I did a search for astronaut selection, and I couldn't find a current list of psychological tests NASA uses. Santy (1994) has an intriguing history (available on Questia), but the book is 12 years old, and given publication lag, the information is even older. In a Google search, I found a recent reference to the Astronaut Personal Characteristics Inventory (ASTROPCI), but little other information. The primary tests they use are predictable, including IQ tests, tests of perceptual-motor functioning, personality inventories, and projectives. It appears that NASA is maintaining an active program of research on personality assessment in astronauts. Nevertheless, several news stories indicate that NASA only assesses astronaut candidates once and never repeats the assessment.

However, here is an intriguing quote from Santy:

The Working Group's position was that personality assessment is underutilized as a resource in astronaut selection, but the empirical record in aviation psychological research of using personality traits as predictors of performance is appalling. This dismal record extends back to World War I and the selection of opponents for the Red Baron. (p. 108-109)

In other parts of the book, Santy correctly points out that psychological assessment is often directed toward identifying psychopathology, but psychopathology isn't a good predictor of success on a job. (Yes, I know there are a host of Dilbert-type jokes here.) This is especially true where the occupation is one in which there are only a small number of people who are employed in the field.

Assessment of personality traits is another way to predict success on various jobs. For example, according to Santy, the 16PF, a personality inventory measuring 16 different personality traits, has been used by NASA in astronaut selection. Logically, different jobs require different traits, so assessing for the right combination of traits would make for a good astronaut right?

No. Personality traits do correlate with behavior, but the correlations tend to be somewhat low. The 16pf is a useful instrument--I've used it myself--but by itself, it's inadequate. No personality inventory is adequate by itself. Generally, we compensate for this weakness by using multiple tests and multiple types of assessment.

That's what NASA does. Each test has a certain likelihood of miscategorization. By using multiple tests, the likelihood of miscategorization declines. NASA takes it farther by also using different types of instruments: Psychiatric interviews are included, as are samples of behavior. Behavioral sampling is done by putting candidates into trainers and assessing their performance under roleplay conditions.

Although NASA doesn't do formal, repeat assessments, there is certainly ongoing monitoring of astronauts. It's the same monitoring that goes on at every job. Both peers and superiors are looking at each astronaut's functioning in training and in everyday performance. I can't prove it, but I suspect that such performance evaluations have washed out people who psychological testing has missed. This brings us back to Captain Nowak.

Why didn't she wash out? From all the reports I've seen, Captain Nowak was a competent astronaut. How could she melt down over a marital separation and a perceived love triangle outside her marriage?

The answer lies in an old argument in psychology: Is behavior controlled by person or situation variables? On the one hand, personality theorists argue that internal variables, such as personality traits, conflicts, and dynamics, control behavior. Behaviorists, on the other hand, argued that the external variables, such as the environment and behavioral consequences, control behavior.

For example, what controls the tendency to cheat on tests? Is it a person's honesty (a trait), or is it the opportunity to cheat (the environment)? The resolution to the argument was predictable. Both person and situation variables are needed together to predict behavior. In many cases, the situation exerts more control than the person. To return to the above example, as the risk of getting caught for cheating drops, the number of students who cheat rises. But cheating never reaches 100%, because honesty is important, too.

The Milgram obedience experiment is another example of the power of the situation, as is the Zimbardo prison experiment. The situation has powerful control over the individual. Change the situation and you change individual behavior.

So, Captain Nowak was able to function in the setting of the astronaut corps. She could work in an environment where death was always looming. She just couldn't deal with the environment where rejection had occurred. You're not going to get rejection on the space shuttle. For reasons that I am not privy to, abandonment and rejection were more threatening than death to her. I don't think any amount of psychological testing would have prevented this awful situation.


Santy, P. A. (1994). Choosing the Right Stuff: The Psychological Selection of Astronauts and Cosmonauts. Praeger.

No comments: