Participant observation is when the researcher is part of the group that s/he is studying. So, if I am involved in a classroom discussion but at the same time carrying out an observation, this is a participant observation. Festinger's classic study of the Doomsday cult was a participant observation. Members of Festinger's team joined the cult in order to study it.

Advantages include that the researcher is less detached from the people being observed. The context in which the behaviours are demonstrated can be better understood by the researcher. This allows the researcher to understand the perspective of the participants. The researcher can also interact and ask questions of the people being observed to clarify behaviours.

The disadvantages include that the researcher may unintentionally change behaviours by interacting with the people being observed. S/he may also lose objectivity by developing a close relationship with the people, or actually finding their behaviour problematic. It is also difficult to collect data when the researcher is part of the group and engaged in the same activity.

One participant effect is reactivity. Reactivity is when a person changes his/her behaviour, simply because they know that they are being observed. This is often seen in clinical psychology where someone who is to be diagnosed acts differently simply because they know that a diagnosis is expected. Reactivity is very similar to "expectancy effect", but it also is at an unconscious level where body language, as well as spoken language patterns, may change. This may also happen when a teacher is being observed by his boss in the classroom.

Another example is the social desirability effect. This is a type of reactivity where the person being observed does things that are socially appropriate and avoids behaviour that may look him/her look bad in front of the researcher.

A naturalistic observation takes place in a natural environment - that is, it is not carried out in a laboratory. One strength is that is high in ecological validity. The environment is not artificial so reactivity should be lower and the participant should act as "he usually does." The greatest limitation of a naturalistic setting is that there is no control over other variables. This means that extraneous variables could have a significant effect on the behaviour of the participant. In addition, naturalistic observations are difficult to replicate. In planning a naturalistic observation in a public space, it is also difficult to manage the sample. For example, if I want to do an observation on how tourists cope with the language barrier in Czech restaurants, I may find that on several of my visits to a popular restaurant, there are no tourists.

There are several advantages of a non-participant observation. First, it is easier to take notes. Since the researcher is not part of the activity being observed, note-taking is relatively easy. It is also non-invasive - that is, it does not interfere with the behaviour of the participants or influence their behaviour in any way. It is also easier to have several researchers watching the same behaviours to establish inter-rater reliability.

In a covert observation, the participants are not aware that they are being observed. This technique is used when the researcher feels that if the participants knew that they were being observed, their behaviours would change. In some rather classic observations, such as Festinger's cult study, covert observations were used to study a very closed group that would not have given consent. This has significant ethical problems.

There are several limitations to this technique. Ethics is perhaps the most obvious. In addition, if it is also a participant observer, the researcher has to be careful not to be "caught." This may lead to unnatural behaviour on the behalf of the researcher which could in turn influence the behaviour of the participants. Also, notes have to be taken when participants are not there. This often leads to the problem of memory distortion where information is written down several hours after the behaviour is observed. Finally, in observational research the debriefings are often important for establishing credibility - in other words, the psychologist asks the participants if they agree with his interpretation of their behaviour. In the case of covert observations, this may not be possible or the reaction upon finding out that they were being watched may lead to a withdrawal of the data or a negative response to the interpretations.

There are several things that a researcher must do to prepare for an observation. This includes what type of observation will I carry out? Ethical considerations. How will the sample be obtained? How will data be recorded? How will variables be defined? Will the observation be inductive or deductive in nature? What type of data will be obtained and how will it be analysed?

Both of these terms refer to the amount of agreement among a team of researchers. In deductive research, researchers use a priori coding. This means that they decide as a team before they carry out the observation what they will look for. They then usually collect quantitative data regarding the frequency or duration of the behaviour. For example, they have a checklist of aggressive behaviours and they watch to see how much aggression is demonstrated on the playground. The frequency with which this was observed is compared between the researchers to see if they have similar findings.

Inductive approaches to observations use emergent coding - that is, they look for themes in behaviour after they have collected their data. The researchers tend not to have a hypothesis, but they want to see "what happens during recess." So, after watching recess on the playground, researchers will then read through their notes and look for trends in the data. From the trends, they will then write an interpretation. When looking to see if there is a similarity in the findings this is called inter-coder reliability. Both terms define the level of similarity, but the philosophy upon which the conclusions are reached is different.

Event sampling is when a researcher makes a note every time a specific behaviour is observed. An example would be to take notes every time someone answers a question in the classroom. Time sampling is when notes are regularly made based on a specific amount of time. For example, the researcher makes notes every three minutes about how many students are on-task in the classroom. Finally, point sampling observers the individual behaviours of participants one after the other. If Jack, Mary and Harry are in the room, point sampling would note what Jack was doing, then what Mary was doing, then what Harry was doing and then begin again.

This is very similar to question 7. In inductive content analysis, the researcher would take her field notes and first read through them to see what "emerges" from the text. She would take out what she thought were significant observations and begin to organize the raw data into themes. From the themes/trends that emerge from the notes, an interpretation would be written. In order to increase the credibility of the findings, either researcher triangulation would be used and the interpretations from the same field notes would be compared. In addition, the interpretations could be given back to the people that were observed and ask if they can confirm the interpretations. Finally, other research methods could be used to see if they result in similar findings.

The limitation of inductive content analysis is that it is highly subjective. This means that it all depends on what "strikes" the researcher has important. Therefore, there is a risk of researcher bias in the interpretation of the field notes. This is why establishing credibility is important. It is also important to employ reflexivity where the researcher reflects on how their own background and behaviours may have affected the interpretation of the observation.