Generalization statements may represent either a logical conclusion or the intrusion of a person's individual beliefs. In either case, generalizations do not necessarily represent the truth. Instead, they are a way of estimating the truth, or guessing at the truth. When making a generalization, you must carefully understand your evidence for making the generalization and how it may be beneficial or hurtful to the group being described.
A generalization may represent a logical conclusion. For example, if you encounter a truth about a group or class of items again and again, you may assume that this truth applies to all things in that group. For example, if you have sampled many different types of peppers and all of them are hot, you may generalize that peppers are spicy. The fallacy is that not all peppers are spicy. Your generalization is flawed, but in your experience it is mostly true.
Judgment and Fact
When generalizing, you must be careful to stick to facts rather than passing judgment. This often comes down to semantics, or word choice, so phrase your generalizations carefully. For example, you may believe that the weather in Houston is nicer than the weather in Toronto. To generalize, though, you cannot say, "The weather in Houston is nicer than in Toronto," because this is not empirical fact. You may, however, generalize that "The weather in Houston is warmer than in Toronto," because this is often the case.
Generalizations may be hurtful to the subject, and are not always accurate. Sometimes these generalizations are called stereotypes. A stereotype is a generalization about a culture or group of people, and they are often inaccurate, offensive or both. For example, a common stereotype may assert that people of a certain race or nationality are lazy, violent or arrogant. Expressing belief in stereotypes like these may brand you as insensitive or ignorant.
Generalizations should typically be supported by evidence -- if not your own firsthand experience, then other evidence. For example, you may generalize that a group of political protestors grew restless. The reader, then, is left curious as to how you reached this generalization; to support it, you must present evidence. Using this same example, you could also write that the protestors began throwing bricks and shouting obscenities; this corroborates your statement that they became restless.