Primary sources are original data -- documentation of an event itself, so to speak, rather than interpretation of it. In history, this means letters or diaries and so on. In science, this would be reports of empirical research. In government, the text of laws and statutes are primary sources.
We use these because: they're the originals. Whatever we're looking at, this is straight from the horse's mouth.
Limits for this type: Not placed into a context.
A secondary source is a layer removed from the primary source. The historian reads several diaries and writes an article about an event based on the different points of view. A political scientist or a journalist writes out the implications of a statute.
We use these because: (generally) they contextualize the primary sources, or they provide incredibly current information (i.e. news articles).
Limits for this type: Variety of viewpoints at play in interpreting and applying the primary sources. Authors may be trying to interpret subjects they aren't well-versed in.
Scholarly sources are a very specific type of source, and "scholarly" means more than just good or credible. These are sources that have been written by experts in a field: actual experts, with advanced degrees in the topic they're discussing. They have often (but not always) been through a peer-review process before publication, in which other experts of the topic critique the piece looking for bad information or wrong conclusions or badly-constructed experiments.
We use these because: they present strong analysis, giving meaning and context and showing connections. The authors have the strongest level of knowledge about the topic.
Limits for this type: Slower to publish, so not as current.
Once you've completed some preliminary research, you're ready to start digging in to find more scholarly sources. Depending on exactly what you're focusing on, you could end up exploring databases relating to business, health, education, and so on.
Consider which disciplines apply to your specific topic when choosing your databases.
Google Scholar is sort of like Google's version of our research databases in that it indexes scholarly content, making it easier to pick out than if you did just a normal Google Search. However, you're less likely to have direct access to these otherwise expensive resources.
You can discover articles here (and it'll come up in Google searches, too) but this also streamlines the process for requesting an article we might not have access to -- assuming the author is active on the site.
Scholarly articles end up behind the paywalls of their publishers, but the authors are allowed to share copies out to people who ask for them. When or if the author is paying attention to the request may be a bit of a gamble, however... but it's worth a shot.
Of course, if you really want to be enterprising, you could also email an author directly if you can find their current institution! Same deal -- you can't be sure if they'll see and respond to your request on the timeline you need, but nothing ventured, nothing gained.
Hey, this tab is really sparse! How come?
In your librarian's experience, it's often easier to focus on finding secondary sources first, then "backtrack" to find primary. Sure, you don't only want to look at your primary sources under the influence of what someone else has already said about them... but it's very hard to know what your options are for primary sources for a given topic without any of that secondary insight. Besides, if you're doing it right, you're not saying the exact same argument as your secondary sources are, and thus you'll be scrutinizing your primary sources in a different way, too.
Let's say you're researching government policy supporting entrepreneurs, and you read a news story about business owners fretting over a new tax law. That news story is likely secondary (and is credible but not scholarly). But aha, there's a new tax law: noted!
You're just looking for leads to get you, ultimately, to the law itself, probably by way of a secondary name-drop of its official title. Doesn't matter how you get there, really.
In the library databases (see Scholarly Articles tab), there are reports from studies. If the authors have conducted new info-gathering and new analysis of that data, those articles are most likely primary sources, as well.
There is no filter for these, but you can look for the word "study" in the title or abstract (summary) to help pick these out. (You can also search for "study" as a Title keyword, but understand that this will exclude a lot, as they don't all mention that word in their titles.) The articles themselves may also have specifically-named sections.