What are the differences between papers that are chosen for presentation vs those which are accepted for poster presentation? Is one less reputable than the other one? How does a committee select received manuscripts for these two categories?
Disclaimer: This is mostly from personal experience on the sending side of the process, and a little bit of talking to professors that have actually been on selection committees.
As far as I can tell, the two main points for selection of poster vs. presentation are completeness and impact.
By completeness I mean whether the work is actually finished or close to finished. If the selection committee cannot tell whether work is finished, or knows it isn't, this is grounds to select the work for a poster instead of presentation. The main reason is that a presentation about work that isn't finished will most likely be somewhat boring, because it lacks strong conclusions, whereas at a poster a discussion about the work that isn't finished might actually be much nicer than work that is completely done.
Impact is a bit of a vague notion, but in this context it consists of relevance to the audience of the conference and the level of innovation in the work. If work seems to be only relevant for a few people attending the conference the work will most likely be selected for a poster, because the people that are interested can then look up the poster. In case of a presentation the room would be mostly empty, because most people are not interested which is undesired. If the work is highly innovative, instead of a small step forward in a big process, this will most likely interest a lot more people, thus making the work suitable for a presentation.
Finished work with high relevance to the conference audience and preferably large steps forward in the field will be selected for presentation, the opposite case will be selected for a poster or even rejected.
In general you could say that a presentation is more prestigious, because it is sort of a quality stamp. However, the boundary between presentation and poster can shift substantially based on the type of conference, the number of submissions to it and the level of submissions and (I know, not fair, but they're only human) the personal preferences of the selection committee.
It might be useful to look at it from the conference organizers' perspective:
- Accepted full papers usually require the author to give an oral presentation at the conference. But the number of time slots for those presentations is usually limited by the size of the venue (number of conference rooms), and extending it either costs the organizers additional money or is not possible at all. So there is a strong motivation to limit the number of accepted full papers, and to choose only those that represent the largest impact/progress to the research field.
- Accepted posters on the other hand are usually presented at a short (~1-2h) poster session, where many posters are presented at the same time in a rather small area. Each conference room booked for the oral presentations can probably be used to present 50-100 posters in a single session, and scheduling a second poster session can cheaply double that. So there is no real motivation to limit the number of accepted posters, beyond making sure that they make sense and represent at least a minor step forward.
These perspectives also match my experience on acceptance and gained reputation in the field of computer science: You can get pretty much anything accepted as a conference poster, while getting a paper accepted is much harder. For that reason, accepted posters are largely irrelevant scientifically, and - in my experience - are usually not even actually published in the conference proceedings (sometimes the proceedings at least contain the extended abstract that was submitted for poster acceptance, sometimes not even that).
In conferences where there is a poster session (or poster sessions) usually there is the option for authors to submit a poster and not a full paper. If this is the case, then there is a page limit for poster and full papers, e.g. poster 4 to 6 pages, full paper 6 to 8 pages.
If there is not such a discrimination but the committee has the option to choose papers for poster session then the selection is based a) either on reviewers comments or b) from the judgment of the organising committee.
In a), there should be a field on the reviewer 's form asking something like: "This submission is for presentation or for poster?" along with the other reviewer form 's fields.
In b), the committee takes into account the subject of conference, its scope, areas of interest and the contribution of the submission. There are two cases: a) The paper is accepted but not, qualitatively, pass the standards of the committee for presentation and is registered for poster, b) There are many good papers but not all of them fit in the time slots of sessions and thus some have to go for poster presentation and the choice is made according to the same criteria for case a).
As for the "reputable" part of the question, if the paper is peer reviewed then is the same either it is presented as poster of presentation. In ones CV the section of conferences' publications (usually) is divided in peer reviewed and non-peer reviewed/abstract reviewed papers.
Regarding "Is one less reputable than the other one? " , in the University I studied they assign points to every contribution (and then sort us for PostDoc positions etc.) A paper chosen for presentation can give you up to 0.75 points whereas a paper chosen for poster can give you up to 0.4 points.
Of course the trick here is the "up to" part, but normally you can count on receiving more reputation from a paper chosen for presentation.