This is the stuff that provides the evidence for your work. Often, but not always, it means material created during the time period you are studying, or by people involved in the subject of your research. Usually, works by historians or other scholars are not primary sources.
These are the scholarly interpretations of the subject written by people who have studied the evidence and are familiar with existing scholarship. Through engagement with primary and secondary materials, these experts come up with new theses and arguments about a subject. Their findings are typically published as monographs (books) and articles. Monographs tend to be broader in scope. They provide more context and draw more sweeping conclusions. Articles (except for literature reviews) published in scholarly journals report on the results of highly focused and specific studies. All of these (but especially articles) assume you are already familiar with the field and the current debate about the topic of research.
These are scholarly works that generally do not engage directly with primary sources. Instead, they are overviews or summaries of what the academic community knows about a subject in contrast to the new knowledge presented in secondary work. Often called encyclopedias, companions or handbooks, these reference works are superb entry points to those unfamiliar with a field since they summarize and contextualize the topics included in that field. They also often have short bibliographies to point you toward the next step. Usually, they do not count as sources for your research because of their distance from the evidence, but if you aren’t settled on a topic, browsing through an encyclopedia can help you to develop one.
If you imagine a body of secondary literature as a conversation among experts about a topic, peer review is the way scholars decide what gets included in the conversation. Journals that follow this editorial process are called “academic,” or “scholarly” or simply “peer-reviewed” and they typically have words like journal or quarterly in their titles.
When a scholar submits a book or journal article for publication, an editor sends the work out to experts in the field. Essentially, the experts are looking for three things:
1: Accuracy: Is the work well-researched? Does the evidence it provides support its claims?
2: Originality: does it add anything new to the conversation?
3: Significance: Is what it is arguing important?
If it does all these things, the reviewers will recommend publication (often with revisions) and then everyone working in that field will know that they need to pay attention to this research.
To ensure that the work is judged on its merits and not the reputation of its author, peer review is typically double-blind. That is, authors and reviewers do not know who each other are.
This process evolved in the decades after World War II when the pace and scope of research meant that individual editors could no longer rely on their own knowledge and judgement when deciding to accept or reject a work.
Importantly, the peer review process assumes the integrity and intellectual honesty of authors; it is not designed to catch fraud, although it sometimes does so.