We've seen how their debate with Mischel led trait psychologists to appreciate that behavior is an outcome of the interaction between personality traits and situations. Another important lesson learned by trait psychologists is the value of aggregation when it comes to measuring personality traits. Aggregation is the process of adding up, or averaging, several single observations, resulting in a better (i.e., more reliable) measure of a personality trait than a single observation of behavior . This approach usually provides psychologists with a better measure of a personality trait than does using a single observation. Consider the concept of batting average, which is seen as a measure of a baseball player' s batting ability (a trait). It turns out that batting average is not a very good predictor of whether or not a player will get a hit during any single time at bat. In fact, psychologist Abelson (1985) analyzed single batting occasions over the whole season. He found that batting average accounted for only .3 percent of the variance in getting a hit. This is a remarkably poor relationship, so why do people pay such close attention to batting average, and why do players with a good batting average earn so much more money? Because what matters is how a player performs over the long run, over an entire season. This is the principle of aggregation in action.
To draw an analogy between batting average and personality, let's say you decide to marry someone, in part, because of that person' s cheerful disposition. Clearly, there will be days when your spouse is not going to be cheerful. However , what matters to you is your spouse's behavior over the long term (i.e., how cheerful your spouse will be in general) and not his or her mood on any given day or occasion.
Imagine taking an intelligence test that has only one item. Do you think that this one-item test would be a good measure of your overall intelligence? You would be right if you concluded that a single question was probably not a very accurate or fair measure of overall intelligence. A related example might be if the instructor in your personality course were to decide that your entire grade for a course would be determined by asking you only one question on the final exam. Surely one questio could not possibly measure your knowledge of the course material. Single questions or single observations are rarely good measures of anything.
Recall the Hartshorne and May (1928) study in which the researchers measured honesty by assessing whether or not a child cheated during a game on one occasion during summer camp. Do you think that this one-item measure of honesty was an accurate reflection of the participants true levels of honesty? It probably was not. This is one reason that Hartshorne and May found such small correlations between their various measures of honesty (that is, because they were all single-item measures).
Personality psychologist Seymour Epstein published several papers (1979, 1980, 1983) showing that aggregating several questions or observations results in better trait measures. Longer tests are more reliable than shorter ones (reliability was introduced in Chapter 2) and hence are better measures of traits. If we want to know how conscientious a person is, we should observe many conscientious-related behaviors (e.g., how neat he or she is or how punctual) on many occasions and aggregate, or average, the responses. Any single behavior on any single occasion may be influenced b all sorts of extenuating circumstances unrelated to personality .
Imagine that a trait psychologist is developing a questionnaire to measure how helpful, caring, and conscientious respondents are. She includes the following item on the questionnaire: "How often in the past few years have you stopped to help a person whose car was stuck in the snow?" Imagine further that you live in a place where it rarely snows. You answer "never," even though you are a generally helpful person. Now imagine being asked a whole set of questions, such as how often you donate money to charity , participate in blood donation programs, and do volunteer work in your community. Your answers to that whole series of questions provides a better indicator of your true level of helpfulness than does your answer to any single question.
Psychologists "rediscovered" aggregation in the 1980s. Charles Spearman published a paper back in 1910, explaining that tests with more items are generally more reliable than tests with fewer items. Spearman provided a formula—now called the Spearman-Brown prophesy formula—for determining precisely how much a test' s reliability will increase as it is made longer . Although this formula appears in all the major textbooks on measurement and statistics, personality psychologists seemed to have forgotten about the principle of aggregation until Epstein (1980, 1983) published his reminders in the early 1980s. Since then, other researchers have provided ample demonstrations of how the principle of aggregation works to increase the strength of correlations between measures of personality and measures of behavior . For example, according to a study by Diener and Larsen (1984), measures of activity level on one day correlated with activity level on another day at a correlation of only .08. However, when activity level was averaged over a three-week period and then correlated with activity level averaged over another three-week period, that correlation went up to .66. Clearly, aggregation provides a more stable and reliable measure of a person' s average standing on a trait than any single observation can.
Aggregation is a technique designed to improve trait measures by adding items to a questionnaire or adding observations to obtain an overall score. Aggregation implies that traits are only one influence on behavio . That is, at any given time, for any given behavior , many factors influence why a person does one thing and no another. Aggregation also implies that traits refer to a person's average level. Traits are similar to the set-point concept in weight; a person' s weight will fluctuate from day t day, but there is a set point, or average level, to which they typically return. An otherwise cheerful woman, for example, might be irritable on one occasion because she has a stomachache. If you were to observe this person on many occasions, however , you would be apt to conclude that, on average, she is generally cheerful.
This example illustrates that personality traits are average tendencies to behave in certain ways. Personality psychologists will never be very good at predicting single acts on single occasions. We may know , for example, that there is a strong negative correlation between conscientiousness and an aggregate measure of being late for class, yet, even if we know everyone's conscientiousness score in your class, are we able to predict on which particular day a specific person will be late? That's not likely . We can, for example, predict who is likely to be late over the whole semester , but we are not able to predict, from that person' s personality scores alone, which specific day he or she will be late. Situational forces (e.g., a failed alarm clock or a flat tire) may deter mine why a person is late on any specific da . But personality may play a role in determining why a person is frequently late (e.g., low on conscientiousness).
Was this article helpful?