It depends on the program. If you are in a lab in a team doing funded research, you might get very little say in the choice of topic.
Otherwise, in many programs, learning how to come up with a suitable topic is one of the skills you have to demonstrate. Your supervisor might be a good sounding board and help, but choice of topic is your responsibility.
As to the question of how, start by asking people and talking about their ideas. Other than that, here’s the short answer:
Choose your main area of interest in which you have some expertise.
Choose five or six key words that represent it.
Go to Google Scholar and type in those key words. Hit return.
Look though the topics and find titles of articles that interest you.
Click on the articles that you like best.
The abstract is a summary placed at the beginning of each article. It will tell you what the paper is about.
If you're interested in the topic, go to the end of the article. That's where many papers have suggestions for further research. These suggestions are usually the results of the research, because research often opens up more new questions.
Then narrow your focus to those that most interest you. If the article is recent, it is not very likely that anyone else has done research on those suggested topics. They help in other ways tool; the author has found some of the most relevant literature and given you a background statement to the problem.
You might like some of those ideas just as they are, or they might inspire you to create your own new topic. Look for gaps in the literature.
Another thing, if it's your first time, don't expect to make a ground-breaking discovery. You'll be successful if you understand the topic well, learn the research process without making mistakes, and come up with a finding that adds to current research.
First, if you don't choose a good topic, your chances of graduating with a PhD are slim. Here's what could happen:
Your advisor probably won't let you advance to a full proposal, and you won't even get to start on it.
If you get to proposal stage with a weak topic, your proposal will probably be rejected.
And if your proposal somehow gets through, you might get bored or fed up with a weak topic and withdraw.
If you get further and realize how weak your topic is, you might want to change it, but that can be difficult.
You will fail the mid-way review. Your supervisors might decide that you don't have the potential to do a PhD and either downgrade you to a Master degree or discontinue you as a student.
If you get to the end, your supervisor might not recommend you to go to the defence.
And if you get to the defence, it is difficult to pass. (In the British system, you'd face a high risk of being downgraded from a PhD to a Masters degree.)
And if you pass, you might be asked to make major changes, just to get over the line.
And if you actually pass the PhD, you will still have a boring dissertation. Consider this:
About this time last year I was doing a literature review and came across a PhD dissertation on my topic. I didn't need to read much; besides noticing that the topic was very weak, my main impression was “Why did you even bother?”
If you need permits, funding, or ethical clearances, you need a proposal before you start research.
A research proposal specifies the purpose of your research. You can express your research topic as a problem to be solved, as a question to be answered, or as a hypotheis to be tested. Your conclusion will be the solution to the problem or the answer to the question.
A focused question helps you to plan your research, do a focussed literature review, know what data to collect, ignore irrelevant information, and come up with a result. Otherwise you will review lots of literature and collect lots of data, but then have to throw most of it out because it doesn't help you to achieve your particular purpose.
It's a chicken and egg situation. You need to read to find a topic, but you need a topic to select items to read.
Talk to others in the same field and talk about their ideas.
Start reading in your area of interest to find a good topic. The best places to look are in recent journal articles on topics that interest you most and where you have some expertise. Near the end of many articles, you will often find a section containing suggestions for further research. Keep notes of your reading because you might need to refer back to them later on. You might even notice a pattern.
Look for gaps in the literature. It is easier to do if you have already read recent journals.
The research needs a specific purpose, and the need for the research is expressed as a problem that needs solving. Without a clear and specific purpose, you would not be able to plan a solution nor know if you have achieved it.
A specific research question will also save you time and effort from reading many things that won't be much help. It might also save you from throwing out many pages of writing that took a lot of time and effort to write, but that are not helpful in answering your narrowed-down research question.
Do a thorough literature review that includes current research. No short cuts there. It will give you a good idea of what has and has not been done. (Having said that, it's always possible that someone somewhere has done something and nobody has noticed.)
Next, when you discuss your topic with faculty members, they will advise you and won't accept a proposal if they think the topic is not unique.
You can't. In fact, it is quite likely if the topic is important and others are aware of it. Things to do about it … Work faster because a slow research can be overtaken. Update your literature review. In some fields you might even have to adjust your topic.
Various. The following come to mind.
First, it must be within the abilities of the institution to supervise and assess, given that it may co-opt personnel from other institutions. This usually rules out many topics that students might like.
Second, it must be feasible. If it’s not actually possible, then students won’t be allowed to try.
Third, it needs to have an intellectual challenge, which normally means a contribution to the theory of the field. This rules out many “how to” practical topics, but the border is sometimes not very sharp.
Fourth, it needs to be a new contribution to knowledge. In some institutions, this includes critical reviews as long as the critique produces something new and leads to a particular conclusion. Other institutions don’t allow critical reviews at all.
Some topics clearly have no potential for a passable dissertation. The reasons? Try these:
The student's main premise is already known to be wrong and the student has not suggested a way to realistically challenge it.
It would be too easy to solve the problem. (I've had students suggest topics that are not in the research literature but I could resolve in ten minutes by phoning a practicing professional.)
The problem is only practical and has minimal theoretical component. (Cf. “too easy.”)
The proposed topic is in essence two or more separate topics.
The student wants to collect information and believes that a big pile of data is enough. (That is, he/she doesn't understand the nature of research.)
Student believes the main challenge is to explain something that is already known. (That is, he/she doesn't understand the nature of research.)
The plan has poor alignment, that is, the problem, purpose, and methodology are inconsistent.
The student wants to use the dissertation to learn a new field where he/she has very little academic or professional background.
The topic is too ambitious for the scope of the PhD. The student proposes to “solve a grand challenge using an approach that would take a large research group a decade to complete.” (With thanks to Βαrbαrα Rοbsοn).
“Complex” is often used to avoid explaining something that one doesn't understand or is too lazy to explain.
If something is “complex” it usually refers to a phenomenon affected by lots of different factors (e.g. cause, effects, assumptions). The research would presumably comprise description and analysis. That is, define (or at least describe) those factors and say how they relate to each other. The phenomenon might vary from case to case. (See “grounded theory,” which is a methodology designed for this kind of problem.)
Another factor is eidetic. That is, how do you know that it is only one phenomenon to be analysed, not multiple phenomena that look similar?
The list of relevant factors could be very long but let's make a start. Consider these:
Can you get all required permits in the location?
Will the location meet the requirements of the proposed research purpose?
Will travel be okay or will it be too difficult, expensive, and/or time-consuming?
Will I need accommodation?
What about my family? Should I bring them too? Will we need accommodation?
Can you provide adequate supervision for work in that location?
Even if you have several good options, which one is best?
There is no specific formula. The way to do it is to anticipate what it will involve and how complex it will be. This requires some experience; one cannot anticipate what will be involved in a particular topic if one has never taken a research project through the whole cycle. It is also a judgement call.
Having said that, many PhD students seem to look at dissertations of other students. The limitation is that they then tend to choose similar research problems and use very standardised methodologies.
These things are helpful in estimating the length of a dissertation:
Does it use standardized chapters (e.g., Introduction, Literature Review, Methdology, Analysis, Conclusion)?
Does it use standardized chapter contents? For example, an introduction might comprise:
Introductory pargraph(s)
Background of the problem
Purpose
Significance
Research questions
Definition of terms
Delimitations
Limitations [Potential weaknesses of the study]
Assumptions
Closing paragraphs
Does it use standardized methodology? Students can see what is involved by looking at dissertations of other students. (The limitation is that they then tend to choose similar research problems and methodologies.)
Let's call the two ideas A and B.
First, do you have a good reason to suggest that A is related to B? If so, you can ask the question, “What is the relationship between A and B? Do they correlate, or does one cause the other?” This could be a good research topic.
However, if you have no good reason to relate them, you might be taking yourself into a trap. One of the requirements for a topic is that you have one topic, so you cannot write a paper on it. You might do research on A or on B, but not one research on two unrelated topics.
This question has come up in another form: “What is relevance in an academic project?”
“Relevance” is an interesting word. It has often been used in a sense that was very unclear. For the word to be clear, I suggest that it needs an object; in other words, “relevant to what?” These have similar meanings but seem to me to be clearer:
“What is the relationship between A and B?”
“What are the implications of A for B?”
“How does A affect B?”
Yes, but it must be different. If it is in the same specialised field, it will most likely be an extension or development of the Master's thesis.