What jesnas subdivided into was more like ways of interacting or media. Interviews really subdivide into unstructured, semi structured and structured. You can use all of the media for each of these.
Unstructured is when you have no plan just an idea or an intersting person etc --chat and find out where it goes--great for new topics etc. Semi structured has a spine that you want to follow e.g. A to B via Z and you can stray on and off as things develop but you get back on track eventally, Structured is much like a set questionnaire.
jesnas highlighted some of the strengths and weaknesses but omitted a couple of biggys. Potential to lead the respondent is one. By skillful prompt you can lead the respondend where you want and get the answer you want. Selection bias is another, picking peole who may give the responses you want. There is also (mentioned by jesnas in part) social desirability bias, people may want to look good in the interviewer's eyes, or may not give their real thoughts because it is again what decent people think.
The big plus side is no other method can gather such rich data as an interview done well.
Interview method subdivided into four - the face to face interview, telephonic interview, Messenger (chat ) interviews, e-mail interviews. But when we talk about interview method we give more importance to face to face interview method.
The mertis of FtF method - the interviewer can establish rapport with the people being interviewd as they get direct attention to the respondents. More probing can be done as per requirement. Hence uselful for untangling complex questions. The response rate would be higher. moreover this is the easisiet way to geresponse if the respondent doesn't posses reading skill to read and understand the questionnaire.
The demerits of FtF method - the flip side of interview mehtod is it advantage itself. The presence of the interviewer may reslut in biased response of the respondent as the respondents may tell the interviewer what they think the interviewer wants to hear. It is also expensive , time consuming, and cannot be used on a large population. Finally there is a problem in supervising the interviewers as they may decide to fake data to save themselves a great deal of travelling and interview time.