Several types of cognitive bias occur due to an attentional bias. One example is when a person does not examine all possible outcomes when making a judgment about a correlation or association. They may focus on one or two possibilities, while ignoring the rest.
The most commonly studied type of decision for attentional bias, is one in which there are two conditions (A and B), which can be present (P) or not present (N). This leaves four possible combination outcomes: both are present (AP/BP), both are not present (AN/BN), only A is present (AP/BN), only B is present (AN/BP). This can be better shown in table form:
| A Present | A Not Present | |
| B Present | AP/BP | AN/BP |
| B Not Present | AP/BN | AN/BN |
In everyday life, people are often subject to this type of attentional bias when asking themselves, "Does God answer prayers?"[1] Many would say "Yes" and justify it with "many times I've asked God for something, and He's given it to me." These people would be accepting and overemphasizing the data from the present/present (top-left) cell, because an unbiased person would counter this logic and consider data from the present/absent cell. "Has God ever given me something that I didn't ask for?" Or "Have I asked God for something and didn't receive it?" This experiment too supports Smedslund's general conclusion that subjects tend to ignore part of the table.
Attentional biases can also influence what information people are likely to focus upon. For instance, patients with anxiety disorders[2] and chronic pain[3] show increased attention to information representing their concerns (i.e., angry and painful facial expressions respectively) in studies using the dot-probe paradigm. It is important to note that two different forms of attentional bias may be measured. A 'within-subjects' bias occurs when an individual displays greater bias towards one type of information (e.g., painful faces) when compared to different types of information (e.g., neutral faces). A 'between-subjects' bias, alternatively, occurs when one group of participants displays greater bias than another group of participants (e.g., chronic pain patients shown greater bias towards painful expressions than healthy control participants). These two types of bias therefore arise due to different mechanisms, and both are not always present in the same sample of participants. Another commonly used paradigm to measure attentional biases is the Stroop paradigm.