You suspect that the answers from interviews may not be correct.
You have access to the environment or place that you wish to study.
You wish to look for unknown or subconscious behavior or patterns.
You believe that the environment is crucial to your project
You find it difficult to simply interview people in your study

Observation research is essentially about directly observing and recording what happens in contrast to questionnaires or interviews where you are one step removed from what's happening. Observations are usually conducted in the field - in real-life situations, where the researcher finds things out first hand. In contrast to experiments in laboratories observations are conducted in a natural setting and it's important to not disturb this environment more than necessary.

There are basically two different types of ; One basic type of observation research is the classical activity of watching and taking notes in an orderly fashion using a schedule, table or similar. The data produced is usually quantitative and it's important to design the research in a way that's relevant to your project.

Here are some typical things that you may look for during systematic observation research:

Frequency of events: How often does something happen? Usually you have to decide beforehand what you will look for. Such as; how often does the carpenter drop a tool?

Time interval recordings: At a given rate (ie. Every 25 seconds) the events are recorded. Either as a general description or on a list of things to look for such as; position, tool in use, workload, etc

Duration of event: When something happens, it is timed to record how long it takes. This may be done according to a list of things to monitor, or based on events as they unfold.

Individual sample: A particular person is observed for a set period of time. After that, the next person is recorded. This is useful when many people are involved in a setting and/or if the interaction between these people may be important.

Disadvantages:

In any type of observation there is a process of interpretation present. We simply see different tings based on who we are. A few things that may influence the way we see things are; familiarity, we tend to notice things that we are used to see or have seen before. The emotional or physical state that you are in may also influence how you interpret events. If you are stressed or anxious or cold, you may see things in a different way.

Total participation
When the researcher takes part in the environment, but his or her role is kept secret. The researcher is the only one who knows that the situation is being recorded or observed. This may be considered unethical.

Semi-secret participation
In this case, a few participants in the study know about the researcher's role as an observer. This allows the researcher to act in an environment that's difficult to gain access to without permission such as industrial plant or similar.

Participation as observer
This is also called "shadowing", when the people that are a part of the study knows the researchers role. This may disturb the "naturalness" of the situation but also provide a more open insight when the participants are consenting.

Total Observation
This may be the most thought of method, where the researcher doesn't take part at all but simply observes and records. This has the advantage of producing better records, but important elements of the situation that is being observed may be lost due to the distanced approach.

Workshops
When it's unsuitable to be on location it may prove appropriate to recreate the situation in a workshop and observe in a controlled environment.