Skip to main content

Verified by Psychology Today


Are 'I' Statements Better Than 'You' Statements?

Is "I feel bad when you do that" really different from "You make me feel bad"?

Imagine that you have missed a deadline for providing a report to a coworker. You run into her in the hallway and she asks, "Hey, where's that report you were supposed to submit last Friday? You are holding up the whole project!"

How would you feel? What would you say or do?

Now, imagine the same situation, except this time when you run into your coworker she says, "Hey, I am getting backed up and feeling a little stressed because I don't have that report yet."

How would you feel in this case? Would you react differently?

According to the accepted lore of interpersonal communication, someone who hears the first coworker statement is more likely to feel defensive and resentful. He or she is likely to feel unfairly blamed and therefore is unlikely to cooperate with the coworker. In contrast, someone who hears the second coworker statement is more likely to feel some embarrassment but is more likely to apologize and make amends by promising to get the report to the coworker as soon as possible. Or at least so says the Conflict Research Consortium at the University of Colorado.

So, what is the difference? In both cases, the coworker points out that the report is late and that this is causing problems. The difference is that in the first scenario the coworker is using what communication experts call "You-Statements" or "You-Messages." You-Statements are phrases that begin with the pronoun "you" and imply that the listener is personally responsible for something: ". . . [the] report you were supposed to submit ..."; "You are holding up the whole project."

In the second scenario, the coworker uses what are called "I-Statements" or "I-Messages," which, though the use of the pronoun "I," attribute responsibility to the speaker: "I am getting backed up and feeling a little stressed ... I don't have that report yet."

Did the choice of pronouns make a difference in your own reactions? Does the use of You-Statements or I-Statements make that much of a difference in how people react to bad news?

Personally, I understand the logic behind the claim that the use of You-Statements might communicate an accusatory tone and I-Statements might demonstrate a willingness to take responsibility for one's own perceptions and feelings. Also, I have read about the different consequences of using You-Statements and I-Statements so many times in my career that I recognize it as a verity in the field of interpersonal communication, and I teach my positive psychology students about You- and I-Statements every semester.

Still, I wonder. In real life, my own emotional reactions to the two kinds of statements are not that different. When my wife (a psychologist well-versed in these principles) tells me, "I felt [insert a negative feeling] when [yours truly did something she didn't like]," I feel largely responsible for her bad feelings, no matter how many "I"s she uses and whether she avoids the pronoun "you" altogether. Even though my trained psychological mind tells me, "She is using I-statements and taking responsibility for the way she feels—she is not blaming me," I still feel I am to blame. The way I look at it, if I hadn't done whatever it was she did not like, she wouldn't be feeling bad.

But, hey, maybe that's just neurotic me, being too sensitive, taking things personally. I should know better.

Or ...

Maybe I am not that unusual. Perhaps the consequences of You-Statements and I-Statements are not that different. I decided to do a little research, searching for evidence that You- and I-Statements do create different emotional and behavioral reactions in the listener. I was able to trace the history of these statements back to Carl Rogers, who is famous for advocating non-directive therapy. Instead of attempting to influence a client, a non-directive therapist aims to create a safe space in counseling sessions for clients to express themselves, to increase their self-awareness, and to choose their own course of personal growth. This is accomplished by using reflective listening, modeling self-disclosure, and demonstrating empathy and unconditional positive regard for the client.

One of Carl Rogers' students, Thomas Gordon, was intrigued by the idea that an authority figure (the therapist) could achieve positive results by giving up power and being non-directive. He had a hunch that other authority figures such as leaders and parents might also achieve positive outcomes by using non-directive methods instead of coercive power. This hunch led to the development of his famous Leadership Effectiveness Training, Parent Effectiveness Training, and Teacher Effectiveness Training programs. Judicious use of I-Statements (Gordon actually coined the term "I-Message" to refer to this concept) is an integral part of these training programs.

As I studied what Gordon and his colleagues wrote about I-Statements, I noticed an interesting paradox. The Gordon model claims that effective I-Statements contain three essential components:

1. A brief, non-blameful description of the behavior you find unacceptable.

2. Your feelings.

3. The tangible and concrete effect of the behavior on you.

Here is one of the examples used to demonstrate the three essential components: 'I feel very upset [feelings] when you’re not here at 8:30 A.M. to answer the phone [behavior] because that means I have to leave my work to cover for you [effect]."

The paradox lies in the assertion that the person using the I-Statement is allegedly not blaming the other person's behavior for his or her unhappiness, but at the same time is saying that the behavior is causing an undesirable, unacceptable effect on the speaker. As I wrote in another article on blame, blaming is the act of claiming that someone's behavior is the cause of my unhappiness. So, despite taking some of the focus off of the other person by saying, "I feel very upset" instead of, "Your tardiness is upsetting me," in the end, the speaker is still blaming the receptionist's tardiness for his or her upset feelings. So, it seems to me that I-Statements with all three essential components cannot be non-blameful.

If an I-Statement were truly non-blameful, it would look more like this: "When I am the only one here and I have to cover the phone, I get really upset. But, hey, that's my problem. I am telling myself that it is a horrible thing to be answering the phone instead of doing other work, but this is just irrational, limiting self-talk. I am "awfulizing"—magnifying the problem beyond all proportion. I need to take responsibility for my own feelings by monitoring and adjusting my self-talk."

If I am sharing my feelings while taking full responsibility for them, I am not blaming the other person for my feelings and not expecting him or her to change in order to make me feel better. But that is not the case with I-Statements; the whole purpose of using I-Statements instead of You-Statements is the premise that they are more likely to get the other person to change his or her behavior! (See Tanya Glaser's comments on "The Power of Vulnerability," especially "[I-statements] can cause the other party to change their behavior by their own choice.") An I-Statement is still meant to manipulate or control the other person's behavior, and the hope that your odds of successfully manipulating the person will be better than if you had used a You-Statement! So all this talk about being self-disclosing and non-directive with I-Statements strikes me as a little disingenuous.

By the way, in Gordon's history of his model, he actually admits that I-Messages can make the other person feel blamed and defensive. At that point, the speaker is supposed to "shift gears" and become more sympathetic to the listener's hurt feelings. And guess what? "It didn’t take long to learn that I-Messages sometimes fail to influence a person to change behavior, even after you have Shifted Gears to Active Listening." Gordon then recommends the six steps used by John Dewey for creative problem-solving. So, not only does the I-Statement fail to prevent the listener from feeling blamed, it also quite often does not achieve its goal of getting the listener to change his or her behavior.

Thomas Gordon was not the only psychologist to attempt to apply Carl Rogers' ideas beyond communication between a therapist and client. Another psychologist, Bernard Guerney, extended Rogers' thinking to couples counseling, where he encouraged couples to choose empathy over blame and to express themselves genuinely and take responsibility for their own feelings through the use of I-Statements toward each other.

Although Bernard Guerney's extension of counselor empathy and genuineness to empathy and genuineness between couples sounds great on the face of it, and despite Guerney's claim that his relationship enhancement methods are backed up by "[a]ward-winning research," some relationship experts have questioned whether the use of I-Statements over You-Statements really makes any difference in marital therapy. In his book, The Marriage Clinic: A Scientifically Based Marital Therapy, John M. Gottman writes:

"Rogers' approach inspired Bernard Guerney ... to develop empathy training for couples. Eventually, all the other marital therapies followed his lead. In fact, they expanded this suggestion in creating 'communication skill training components.' But what was the scientific justification for teaching these communication skills? How did they decide that 'I-statements' are better tha[n] 'you-statements'? How was any of it decided?"

Although Gottman himself had followed the overwhelming consensus among psychologists that couples are better off when they speak to each other using I-Statements rather than You-Statements, he began to have some doubts. A major marital study reported in Hahlweg and Jacobson (1984) found that couples who received intensive training in active listening (which includes the use of I-Statements) showed a decrease in negative interaction but no increase in positive interaction. In contrast, couples who received a behavioral treatment showed both a decrease in negative interaction and an increase in positive interaction. Furthermore, couples trained in active listening relapsed back to pre-treatment levels unhappiness eight months later, while couples receiving the behavioral treatment show continued improvement over eight months.

The fundamental problem, as Gottman sees it, is that in therapy the counselor can easily empathize with a client who is complaining about someone else, a third person. It would be a different story if the client complained about the counselor, who might consider this to be resistance and would no longer empathize with the client. In a marriage, says Gottman, the proper place for empathy is "colluding to trash a third party, for which there appears to be no more satisfying way of engendering solidarity and we-ness." But when partners feel hatred toward each other, they find it almost impossible to generate I-Statements. When they are stung by a remark from the other person, they want to sting back. Gottman found that the difference between partners in happy, stable marriages versus unhappy, unstable marriages is that when partners in happy marriages are attacked they respond with an attack of equal magnitude. Partners in unhappy marriages respond by escalating the attack.

Even critics such as Gottman do not recommend against using I-Statements and active listening; they only wish to warn us not to expect profound results. If you want to try to change someone's behavior by blaming the person for your negative feelings, go ahead and use an I-Statement. Just keep in mind that your mileage may vary.

More from John A. Johnson Ph.D.
More from Psychology Today