Attribution Theory
of Fritz Heider
(modified From the Second Edition of A First Look at Communication Theory by Em Griffin, © 1994, 1991, McGraw-Hill, Inc. This text-only version of the article appears on the World Wide Web site www.afirstlook.com. A facsimile of the original article is also available in PDF format.)
My wife, Jean, served on a jury in a federal case involving conspiracy, racketeering, drug dealing, armed robbery, and extortion. The seven defendants were accused of being the lieutenants in the "Little Mafia" gang which terrorized a Chicago neighborhood. The gang leader had escaped from police custody and was on the FBI’s most-wanted list.
The key government witness was an ex-gang member named Larry. Larry was called "the Canary" by the defendants because he turned informer. For two months Jean listened to the testimony and tried to figure out whether Larry’s story was credible. Was his behavior on the witness stand that of a pathological liar, a rejected pal seeking revenge, a petty crook who would say anything to save his own skin, or an honest witness dedicated to the truth? Fritz Heider, the Austrian-born father of attribution theory, said that we all face the same task Jean confronted-trying to figure out personality of people from their behavior behavior.
Heider, who became a psychologist and taught at the University of Kansas, said that attribution is the process of drawing inferences. We see a person act and immediately reach conclusions about him or her that go beyond mere the sensory information regarding what we saw or heard. Suppose Larry yawns while on the stand. Is he bored, afraid, tired, or indifferent? Jean will search for an explanation that makes sense to her. Heider would have seen her as a naive psychologist bringing her common sense to bear on an interpersonal judgment. If he were developing this theory today, he might well describe Jean and all of us as Judge Judy stand-ins, rendering decisions in a people’s court of everyday life.
We’re constantly told we shouldn’t judge others. Attribution theory says we can’t help it. Like my wife, who had to listen to Larry’s testimony for a week, we’re inundated with sensory data, some of it contradictory. Faced with this information overload, we make personality judgments about people in order to explain otherwise confusing behavior. For example, although Jean had earlier thought Larry was a credible witness, she wondered why Larry yawned when describing how a gang member struck a victim on the head with a baseball bat. She made a snap judgment that he was callously indifferent to human suffering. She might have drawn other conclusions (e.g., maybe he couldn’t sleep the night before having to testify in class, or maybe hours of testifying in court was wearing him out). But no, like all of us, she quickly arrives at a decision that reflects her belief about the cause of his yawn.
In addition to our need for clarity, there’s another reason for making causal inferences from behavior. We want to know what to expect in the future. Prediction is a survival skill. During the third week of the trial, Jean came face-to-face with one of the defendants outside a train station. Mildly anxious, she quickly turned aside. Accurate attributions can help us know which people might do us harm and which people might be beneficial to know and be around. In my wife’s case, she made an attribution that this person potentially was harmful to her, and she moved away from him.
ATTRIBUTION: A THREE-STEP PROCESS
Attribution is a three-step process through which we perceive others as causal agents. The steps are perception of an action, a judgment about a person’s intentions behind the action, and a judgment about the dispositions of the person. Suppose you are stopped at a red light, and the driver in the car ahead flips an empty soda can into the gutter. Before the light turns green, you mutter the three thoughts that cross your mind:
I saw that! (Perception of the action)
You meant to do that! (Judgment of intention)
You’re a slob! (Attribution of disposition)
I visited the courtroom the day ex-gang member Larry described the baseball bat attack, so I’ll use my reactions to illustrate Heider’s chain of causal inference. As I’ve already told you, Jean experienced her own attributional sequence as she heard the testimony that day. Since her private world differs from mine, however, I can write with certainty only about one person’s experience-my own. This is a very important point about attribution. Every person’s experiences is different from everyone else's. The attributions that we make are based on our experiences and perceptions, and therefore our attributions about the causes of events will differ from each other.
Step 1: Perception of the Action
The courtroom was on the twenty-first floor of the Federal Building. A well-dressed, handsome man carrying a bag of donuts dashed toward the elevator just as the doors were closing. I slapped my hand against the breaker bar to hold the car, receiving enthusiastic words of thanks and the offer of a donut. We traded a few friendly comments until we got off at the same floor. It turned out we were both headed for the south courtroom. A few minutes later I learned that his name was Brian, and I also heard testimony of how he used a Louisville Slugger to beat up a cocaine addict who hadn’t paid for his drug.
Was Brian observed? Yes, Larry was an eyewitness and gave an account of the act. My perception was secondhand through Larry’s description, but Jean had told me the night before that she thought Larry was a credible witness. My vicarious observation of the scene triggered the start of the attribution chain. If I hadn’t heard the story, the process would never have started. Perception (of Larry) can begin second-handedly -- through haring about an event from someone else, rather that seeing it yourself.
I’m sure that my perception of the action was subject to all the biases listed in the introduction to this section. It was initially difficult to picture such a brutal act performed by Brian, because that information contradicted my initial, personal image of a generous, joking Brian. This first impression, which was formed on the elevator, was one of warmth, a trait that casts a halo over all other qualities.
I like to think of myself as a good judge of character, so my continuing desire to see Brian in a favorable light could well have compromised the integrity of my courtroom listening. That’s why jurors are excused "for cause" when personal motivation prevents their being impartial. Despite these biases, the new information crashed into my world and I moved to the second stage of causal attribution.
Step 2: Judgment of Intention
Heider argues that we need to find two factors involved in a person's action before we are likely to hold the individual personally responsible for it. We need to find intention (motivation) and ability in the person. Intention refers to two things: simply said, they are want and try. If we find that a person wanted to cause an outcome and tried to cause it, we generally conclude that he intended to cause it. For example, if we know that Brian was angry at the dead man and swung a bat at him, we are likely to conclude that there was intention to hurt the dead man in Brian's mind. If we conclude the person had the ability to cause an outcome, we are saying that he was present at the scene, had the necessary physical (strength) or material (baseball bat) resources to do it, etc. Since Larry’s words convinced me that Brian was present when the man was struck with a bat and had the bat in his hands, intention seems to be present in his actions.
So, is Brian guilty of murder? Not necessarily. We still do not know the degree of his intention to harm the victim. The next question is, “To what extent had Brian wanted it done?” That may sound like a strange way to ask the question, but Heider didn’t consider intention as an either-or matter. He identified five levels of personal causation. We can see these gradations reflected in the American legal code. Suppose for a moment that the victim had died. Our judgment of intention could place the killing blow into five different categories before the law.1
1. Association Level. At the association level a person is judged to be the cause of an event that he, in fact, did not cause. Often the person is around the event, but was not actually involved in its causation. It’s like your parents accusing you of something that your sister had done. Or, it’s like someone being convicted of a crime that he did not do.
In our case, despite the testimony, maybe Brian didn’t direct the attack, swing the bat, or even know the man who did. Maybe Larry was not telling the truth. He was merely in the vicinity at the time. Chance proximity is no reason to assign causality. Arguing that Brian was guilty of murder in a situation like this would be assigning him responsibility at the association level.
2. Casual Level. At the casual level a person is judged to be the cause of an event that he did, in fact, cause. However, it was an outcome that he had no reasonable way of foreseeing would occur. As an example, suppose that you are walking into your bank. The bank has solid wooden doors that you can’t see through. You push on the door to open it. Just as you are doing this, a bank robber is backing out of the building from the inside. When you push the door, he is hit in the back and falls down. The security officer in the bank then apprehends him. You are held as a hero for causing the capture of the bank robber. Is there any reasonable way you’re your could have known that your opening the door would cause the capture of this criminal? No. But, did you cause the capture? Yes. Calling you responsible for his capture in this circumstance would be holding you responsible at the casual level.
In the case of our trial, perhaps the event took place at a sandlot baseball game. Brian took a mighty swing at a pitch that fooled him, and the bat slipped out of his hands, striking the unlucky fellow standing forty feet away in foul territory. It’s true that Brian’s hand was the ultimate cause of death, but he had no motive or desire to do harm, and he could not reasonably have foreseen that the bat would fly so far and so accurately toward the person’s head. Here, a coroner’s jury probably would rule that the death was accidental. However, if they tried to prosecute Brian for it, it would be holding him responsible for an action at the casual level.
4. Foreseeability level. At the foreseeable level a person is judged to be the cause of an event that he did, in fact cause. He also should or should have been aware that he might cause the outcome, although it was not something he really wanted to occur or was actively pursuing. Having an accident while driving drunk is a perfect example of this level. The drunk driver did not want the accident to occur, but should have known that his impaired driving could produce the outcome.
In the case of our trial, picture Brian trying to hit fly balls to a group of friends in a park that is crowded with people. This is a dangerous game from the start, as there are many people in the near area of the game. Brian becomes angry at his inability to hit the ball in the air He impulsively flings the bat aside, blindsiding and killing a man playing with his children nearby. This is foreseeable commission, as Brian should know that throwing a bat in crowded conditions might cause someone to be hit and hurt. Brian might honestly claim that he never meant to hurt anybody, but he should have known that his reckless behavior might do so. In our legal system, the law would regard him as responsible for the outcome of his careless act. He probably would be convicted of manslaughter.
5. Intentionality level. This level represents most of our normal behavior. This level concerns events that you want to occur and work consciously toward causing. If you want an “A” on an exam, you might study heard for it intentionally. If you want a job, you might intentionally go out and apply for a job at several places. If you are thirsty, you might intentionally go to the refrigerator to get a soda. Intentional behaviors reflect your own personal motives and goals.
In the case of our trial, Larry’s account of Brian’s behavior reflects intentionality. Larry described Brian as the sole cause of the attack and was convinced he meant to destroy. The police would label this as "premeditated homicide." We’d call it murder.
3. Justifiability level. At the justifiability level of commission, a person has again performed a behavior intentionally. The difference here is that the intentional behavior reflects very little about the personal goals of the individual. Rather, the intentional behavior is something that almost all other people would have done had they been in the same situation. As an example, suppose that you were driving on a highway when a dog ran onto the road in front of you. You could hit your brakes to miss the dog, but a car is close behind you and hitting the brakes would cause a bad accident. You could try to swerve to miss the car, but there is a steep bank on your right and an oncoming car on the left. This action also would cause you to have a bad accident. You end up driving on, hitting and killing the dog. Your intent was to drive on; you knew that this would cause the dog to die. However, in this case we are not going to hold you completely responsible for the death and call you a “dog killer” because we recognize that almost anyone else in your situation would have done the same thing.
In the case of our trial, suppose the event took place in Brian’s apartment. Returning from work, Brian surprised an intruder who came at him with a knife. Brian grabbed the bat, which was propped in the corner, and swung it to protect himself. Some might wonder about excessive force, but most people would see it as self-defense. Almost anyone else in this situation would have done something similar to defend himself. In the legal system, we probably would call this justifiable homicide.
Common Biases in Judging Intention
Because I’m illustrating attribution theory in a courtroom setting, I’ve cast judgment of intention in legal terms. But Heider emphasized that the issue transcends accountability before the law. What we’re really dealing with is moral culpability - perceived responsibility in the court of public opinion. When we judge another’s motives, we move past Sergeant Joe Friday’s dispassionate "Just the facts, ma/am," and enter the realm of values, "shoulds," and "oughts." It’s easy for bias to shade our judgments about other peoples’ behavior. Three volumes of attribution research edited by Harvey, Ickes, and Kidd verify the human tendencies described in Heider’s original work.2 They find three general biased tendencies in our attributions for the causes of other peoples’ behavior:
1. We tend to hold others more responsible for negative results than for positive outcomes. If the first-year student who sits next to us in class flunks a test, we are likely to think that he’s stupid (personal cause). If he aces it, we’re more likely to think he’s lucky (situational cause).
2. We tend to hold others more responsible for not trying (poor effort) than for incompetence (poor ability). It’s worse to be lazy than to lack ability.
3. We tend to hold others more responsible when they aim to improve their position (positive goals) rather than avoid loss (negative goals). For example, we tend to judge more harshly a hungry person who steals food than we do a well-fed person who won’t willingly share it. (If you don’t think that this is the case, then consider how we treat these people in the law. The hungry person who steals is jailed; noting is done to the wealthy person who does not share.)
4. We tend to hold others more responsible for their outcomes when we fear the same thing may happen to us. A veteran skydiver haughtily claimed that anyone who "bounced’ got what he or she deserved. The skydiver used defensive attribution as reassurance that death by sudden impact always happens to someone else. The other person did something stupid that he won’t do himself. Someone who does not skydive might not be so harsh on the unfortunate skydiver.
5. We tend to hold others more responsible for their outcomes than we hold ourselves for our own outcomes. Apparently, we use a double standard as we decide who should be held accountable for mistakes and errors. When things turn out badly for others, we assume it’s their fault; but for our own failures, we tend to blame circumstances or other people. We see others as causal agents, but we give ourselves an excuse. If someone else has a car accident, I think that he or she is a bad driver who should have his or her license taken away. If I have the same accident, something in my environment was the cause (slick road, glare from sunlight, etc.). I’m still a good driver.
All our biased judgments involve a decision between personal and environmental control. This tension is a crucial ingredient in the third step of attribution. Having taken notice of Brian’s aggressive action and believing that it was an unprovoked and wanton attack with malicious intent, I’m now in a position to make a dispositional attribution about him.
Step 3: Attribution of Disposition
Heider considered attribution as an effort to "predict and control the world by assigning transient behavior to relatively unchanging dispositions."3 You can see that process in my train of thought as I heard the testimony about the assault:
That was a brutal attack. Brian’s a brute! I can’t imagine bashing someone over the head when they haven’t done anything to hurt me. I wonder if he’s insane. No, there was nothing crazy about how he acted on the elevator. And there wasn’t any pressure from the gang leader to do the job. Brian offered to work the guy over for $500, yet he didn’t even need the money. No doubt about it; he’s a brute and he’ll probably be violent again.
Note how quickly I have jumped from judgments about the cause of Brian’s behavior to assigning dispositions to him to explain why he causes such an event. I assumed that people who do things like that have dispositions in them that cause them to do such things. As naïve psychologists, we constantly assess how much an action is due to personality as opposed to environmental pressure. When judging others, our tendency is to discount external factors and put our thumb on the character side of the scale. Our tendency is to assign responsibility to people for their actions (intentionality) and to assign traits and dispositions to them to explain their behavior. So, Brian killed the guy because he is a brute.
ATTRIBUTIONS DEPEND ON PERCEIVED FREEDOM OF CHOICE
The key issue is choice. If we see others as compelled to act as a result of circumstances beyond their control, we won’t assign their behavior to enduring traits of character. I tried to consider the possibility that Brian was driven by madness, coerced by his boss, or forced to the act by financial need. These mitigating circumstances would short-circuit the attributional chain. But I couldn’t find any outside explanation which would account for the severity of the action.
Heider stated that we judge an actor’s freedom as proportional to the difficulty of performing the act. It’s not easy to crack someone’s skull. It takes tremendous desire and exertion to lay waste the human head. Being face-to-face with the victim makes it even harder. (Dante condemned the designer of the catapult to an inner circle of Hell. By giving a warrior a means to achieve death without having a visual link to the victim, the inventor made killing easier.4) Since Brian performed this difficult task with apparent ease, I considered him as having true choice, his act free of constraint. It became easy for me to say that the behavior was intentional and for me to attribute to Brian a cluster of personality traits and attitudes consistent with this voluntary violence. My conclusion: Brian is a hardened criminal.
Don’t be surprised that I ended up explaining Brian’s behavior by the type of person he is rather than by the circumstances surrounding the attack. None of us is immune from the bias that Stanford psychologist Lee Ross calls the "fundamental attribution error."5 This is the tendency for observers to underestimate situational influences and overestimate dispositional influences upon behavior. Whether it’s a police officer’s callous reaction that a rape victim was "asking for it," the football coach’s analysis that a player missed a tackle because he didn’t try hard enough, or parents’ assumption that the crumpled fender on the family car is due to their son’s carelessness, we usually assume that people are responsible for the things that happen to them.
This extended account of judgment may give you the impression that attribution is a prolonged, conscious deliberation which takes place only in formal settings of guilt or innocence. Not so. Causal inferences are usually snap judgments made whenever we see others in action. Our judgments deal with praise (e.g., praising me for opening the door that hit the bank robber) as well as blame. Heider’s theory has generated thousands of studies that blanket the map of interpersonal relations. One of the fascinating extensions of the theory is the work of Cornell University psychologist Daryl Bern, who is interested in the way we look at our own actions.
SELF-PERCEPTION: A SPECIAL CASE OF ATTRIBUTION
Bern is concerned with the dispositional labels we assign to ourselves. He claims we don’t have favored status when it comes to figuring out our own prevailing moods. Our weak internal signals may give clues to our attitudes or emotions, but behavior is the acid test that confirms or rejects our intuition. So we watch ourselves act and then draw conclusions about our own inner disposition just like outside observers do.
Conventional wisdom suggests that behavior follows attitude: "I play tennis because I like it." Bern’s radical behaviorism says it works the other way around: "I like tennis because I play it." Later in the course we’ll see that cognitive dissonance theory also predicts that actions precede attitude, but Bern explains the sequence on the basis of self-perception. We see ourselves put a dollar in a beggar’s cup and decide that we are compassionate.
Emotions work the same way. You might think it’s safe to assume that a fellow knows when he is sexually aroused. Not necessarily, according to Bern. He cites a study by Stuart Valins, a research psychologist at the State University of New York (SUNY) at Stony Brook. College men were wired with fake electrodes that supposedly picked up their heartbeat and amplified it through a speaker for them to hear. The experimenter occasionally varied the bogus biofeedback as the males looked at pictures of nude women. The men reported being most "turned on" by the photos that were associated with a change in heartbeat.
Some cynic has suggested that love is a feeling you feel you’re feeling when you feel you’re feeling a feeling. The statement is consistent with Bern’s description of self-attribution. We aren’t sure what we feel, so we look to behavioral clues to fill in the gaps. "It must be love cuz my heart skipped a beat."
CRITIQUE: RENDERING A VERDICT
Heider suggested that people systematically evaluate causes of behavior in a commonsense search to understand why things happen. If Heider was right, however, some of the jurors Jean was with failed to get the message. After ten days of bickering, nine were convinced of the defendants’ guilt; three saw them as innocent. One jury member chose to ignore all the evidence, relying instead on a gut feeling that the men were innocent. In reaction to this member’s stubborn refusal to discuss wiretap recordings, another juror lay on the floor for days and wouldn’t discuss anything.
The jury’s bizarre behavior illustrates a serious weakness in Heider’s case. Attribution theory stresses human rationality and ignores the role of emotion. Heider described the process as one of making "causal inferences," but are we the detached observers of the human scene that the phrase suggests? It may be more accurate to describe the attribution process as "jumping to conclusions." And Heider’s naive psychologists might be more correctly labeled "self-serving perverters of the truth."
Just as Jean and the other jurors had to decide whether to believe the testimony of Larry the Canary, so you and I must render a verdict on the validity and usefulness of attribution theory. A vote to reject means we still need to find a way to explain our quick judgments of personality based on behavior. But if we find Heider’s attribution principles to be true beyond a reasonable doubt, sheer human decency requires us to resist some of the built-in biases that tilt our perceptions.
The chief culprit is our consistent tendency to assume that other people are the sole cause of their actions, that they are free to move in any direction they want. In the absence of overwhelming evidence to the contrary, we’d be much closer to the truth if we viewed others as enclosed in a maze of environmental constraints. Most people on welfare aren’t poor because they’re too lazy to work. Pilot error is only part of the story in airplane accidents. There are times when others really can’t help being late. In short, other people’s lives are just as complicated as ours.
In spite of its questionable ability to deal with the passionate side of relationships, attribution theory provides a helpful analysis of the way we parcel out praise or blame. It has stimulated thousands of research studies that investigate the way people interpret the behavior they see. Heider’s ideas may not be perfect, but as the theory itself suggests, few objects of our judgment are as good or as bad as we want to give them credit for being.
QUESTIONS TO SHARPEN YOUR FOCUS
1. What happens when we observe an action but decide that the other person had no intention to do it?
2. Tracy’s coffee ends up all over Lacy’s new shirt. Before you jump to the conclusion that Lacy is a slob, can you construct a scenario for the five categories of causation-association, causality, justifiability, foreseeability, intentionality?
3. What steps could you take to make certain that you don’t commit the fundamental attribution error?
4. According to attribution theory, what is it that we attribute to other people?
NOTES
1 M. E. Shaw and J. L. Sulzer; "An Empirical Test of Heider’s Levels in Attribution of Responsibility," Journal of Abnormal and Social Psychology, Vol. 69, 1964, pp. 39-46.
2 J. H. Harvey, W. Ickes, and R. F. Kidd (eds.), New Directions in Attribution Research, Vols. 1-3, Lawrence Erlbaum Associates, Hillsdale, N.J., 1976, 1978, 1981.
3 Fritz Heider, The Psychology of Interpersonal Relations, John Wiley & Sons, New York, 1958, p. 79.
4 Dionysiue, the Elder, 330-367 B.C. He perfected the catapult in sieges of Italian cities (Ralph Payne-Gallwey, The Projectile-Throwing Engines of the Ancients, Rowman and Littlefield, Totowa, N.J., 1973, p. 5). Dante places the "tyrant of Syracuse" in the seventh circle of Hell, submerged in a stream of boiling blood (Dante Alighieri, The Divine Comedy: Infkmo, Canto x11:107).
5 Lee Ross, "The Intuitive Psychologist and His Short-comings: Distortions in the Attribution Process," in Advances in Experimental Social Psychology, Vol. 10, Leonard Berkowitz (ed.), Academic Press, New York, 1977, p. 184.