Someone misses a task, and you think, they are lazy. Someone gives a short answer, and you think, they are rude. You have moved from an action to a description of the person. That explanation might be possible. The question is whether the information you have supports it. Correspondence bias means giving someone's character too much weight while giving their circumstances too little. Checking for it does not mean excusing every action. It means asking what the person could realistically do in that situation, before deciding what the action says about them. Psychologists Daniel Gilbert and Patrick Malone reviewed several ways this can happen. You may not know about a constraint. You may know about it but underestimate its effect. Or you may form an early character judgment and fail to revise it when more information arrives. Imagine a colleague answers you briefly while facing an urgent deadline. If you do not know about the deadline, you are missing part of the picture. If you know about it but still expect their usual relaxed response, you may not have allowed enough for its effect. The example is fictional; it illustrates a possibility to check. This can affect favourable judgments too. Someone answers every question correctly, but they wrote the questions and chose a subject they know well. Their performance may be impressive. Inferring broad brilliance without considering that advantage would still go beyond what the occasion establishes. A useful first step is to replace the character label with an action you could describe. Instead of they are unreliable, try they missed two agreed deadlines on this project. Now there is something more specific to investigate, and a clearer basis for discussing what needs to change. Ask what is known about the conditions. Was the task clear, feasible and agreed? Did the person have the time, information and resources it required? Include their choices and responsibilities too. A constraint is something to check, not a reason to assume that the person had no influence. Next, compare a relevant occasion. Does the same difficulty occur when the instructions are clear and enough time is available? Do other people really have the same information, skills and choices? Saying anyone else would have managed is a claim that needs evidence as well. A difference between situations can weaken an always-like-this claim. It does not prove which condition caused the difference. And repetition in one setting may reflect a repeated constraint as well as a personal tendency. Keep both the pattern and what remains uncertain in view. Do not replace a harsh judgment with an equally confident invented excuse. A deadline matters if it is known or worth asking about, not simply because it would be a comforting explanation. You can ask for a change, or set a boundary, while the cause remains uncertain. There is another tempting shortcut: the idea that we always explain other people through character and ourselves through circumstances. Bertram Malle's analysis of many studies did not support that as a general rule. Check the particular judgment, including your own, rather than assuming the shortcut explains everyone. Research gives some reason to practise. Training experiments by Carey Morewedge and colleagues included errors involving too little attention to circumstances. People improved on the tested judgment problems, with effects still found at least two months later. That does not establish a permanent change in everyday relationships from reading a page. Our practical suggestion is a short pause before an important character judgment. Describe the action. Check a relevant constraint. Compare an occasion. Then review what happened next. If the behaviour continues despite a workable agreement, address that pattern. Understanding a difficulty does not make its impact disappear. Here is a separate fictional example. Someone misses a task you asked them to do, and your first thought is, they are lazy. Before deciding that, what would you need to know? Choose one missing fact that could make your next conversation more useful. Pause the player if you want longer. [Pause: 9 seconds. Pause the player for longer if helpful.] You might need to know whether they understood the task, agreed to it, or had the resources it needed. Those questions do not establish the answer. They help you choose a clearer request or comparison. You can still discuss the missed task and its impact. The distinction to remember is this: correspondence bias concerns the explanation you give an action. Confirmation bias concerns how you handle evidence for an explanation. You can do either, or both. The written entry has the research links, further comparisons and space to record what you try.