What tips or tricks do you use when researching a topic to find actually useful information?

This falls generally under "research methods".

In theory, the web is a wealth of information. In practice it's increasingly a cesspit of disinformation. This doesn't make it useless, though it's necessary to tread with caution.

The same issue applies, of course, to other media. The Web is not unique, though its scale and response time are.

Methods differ both by topical recency and domain.

Methods by item recency

• For immediate breaking issues: within the past minute to hour, Twitter or its front-ends (I prefer Nitter) are highly responsive. It's best to use these to establish that something has happened rather than trying to establish just what has occurred: first reports are often inaccurate, speculate (usually inaccurately) as to cause or responsibility, and lack context. Emergency alert services may provide reports, but are often disappointingly ineffective. For natural disasters, weather, earthquakes, etc., online notification dashboards provide near-realtime updates on initial events & magnitudes. Sometimes local news services.

• Hour to day: News media, supplemented by other sources. This is the news's job, and shortcomings notwithstanding, they're reasonably good at it. Hitting a few sources (and checking to confirm they're not merely parroting some common source) tends to give a reasonable picture. WNYC's On the Media produces a number of Breaking News Consumer Handbooks which are excellent. Mainstream sources (Reuters, AP, AFP, national dailies, national broadcasters (PBS, NPR, BBC, CBC, ABC (Australia), Deutsche Welle) are good starting points / checks. Financial news outlets (Financial Times, Wall Street Journal (despite Murdoch), Bloomberg, The Economist), whilst conservative, are pretty good with getting facts and maximising information-to-hype ratios. Be aware of reputation and bias. And check against debunking sites.

• Day to month: Wikipedia. For stories which are complex and develop over time, it's extremely hard to beat. I first noticed this following the 2004 Boxing Day Quake and Tsunami. Where traditional media were dribbling out decontextualised facts or retreading background, Wikipedia assembled these into a coherent whole. Some long-form journalism pieces can approach this, but they're much harder to find. The 2011 Tohoku quake & tsunami, Russian interference in the 2016 United States elections, the Oroville Dam Crisis, and ongoing COVID-19 crisis are other excellent Wikipedia topics. Again, not perfect or immune from distortion or influence, but remarkably resilient. Follow up on sources. An expert and highly reputable blog or social media source may be useful. Also various reporting dashboards (e.g., ongoing Covid statistics).

• Month to year: Also a Wikipedia strength (most of the examples I listed above span this range), but add in detailed survey coverage, long-form magazine treatments, official reports and investigations, journal (academic rather than commercial press) sources. In some cases, books. Lectures, podcasts, online video, speeches, and presentations are often useful in this timespan (and longer). For technical (computing) topics, blogs, good online discussions, conference proceedings, etc.

• Year to decade: if possible start looking for published sources in books or academic journals. This will vary by topic, but the process of compiling, structuring, researching, editing, and publishing such works tends to shake off a lot of cruft. You're also starting to reach the point where tracking references and citations starts to pay off. If a document (blog, journal, article) is mentioned, or an authority is namedropped, track it down. One of my more useful research digressions (a half dozen books) grew from a five second reference within a conference seminar I'd found on YouTube. Good books and articles have bibliographies which are often more valuable than the texts themselves. Conversely, a book lacking an index or bibliography loses half or more its value compared to one that does.

• Decade to millennia: Again, books, but strongly emphasizing both primary sources and compendia, surveys, and digests. Topics here tend to be science, literature, culture, philosophy, and religion. Primary sources give the unvarnished original (all the more so if you can read the original language), surveys reveal the larger landscape and highlights. At this level, if you've not already begun one, a knowledge-capture tool becomes essential. I use a mix of index cards, online blogs and comments (backed up!), notebooks, and unpublished drafts and notes.

Methods by topic

For basic facts and figures: some authority tasked with providing them. Often government agencies which can be excellent, if difficult to navigate.

For technical topics, especially geared at accomplishing some task; domain experts, or the communities in which they hang out. Beware imposters. Despite the appearance of constant change and churn, underlying concepts often have lasting value, and evolve slowly.

For news and political analysis: sources not directly engaged in the matter, at least for general background. Commentary of direct participants may have points for immediacy, but is virtually always biased or distorted. A disinterested foreign source (and both qualifications matter) is often best. Know the biases of various sources.

For "low-tech" topics, such as your olive oil example, you'll have to hunt. Here, searching based on author (etymological root, authority), is often best. Ask people who know their shit who they'd turn to, and if the same names keep coming up, prioritise those.

Hi-tech or low, you'll have to practice the methods directly. That, and explaining yourself to others, is the best learning method.

For conceptual and academic topics: books, texts, journal articles, bibliographic search. I'll contact authors directly in cases, with specific questions I can't otherwise resolve.

Developing a framework of a domain, whether from existing models or your own personal organisation, off which to hang newvconcepts and methods, is also useful. New facts don't simply pile up but interrelate.

On assessing expertise

It can be hard to tell true experts from charlatans, and many methods aren't entirely reliable. An expert may not have a deep humility, though many do, and most (at least the ones you'd want to have teach) are genuinely appreciative of an interested student. Cross this with a disdain for flakes and casuals. The huckster is virtually always trying to sell you something, usually directly, sometimes indirectly. The best experts not only explain simply and clearly, but those explanations produce results.

Another fairly reliable guide: average (or worse) instructors will show what or how, often providing a simple cookbook recipe. True masters will explain why, show variants, risks, pitfalls, rescue or troubleshooting procedures, and side effects --- useful or otherwise. This may strike many people as TMI, and there's a knack to delivering this information without overwhelming students. I (looks guiltily at length of this respose...) ... generally appreciate this.

TL;DR:

For online search, increasingly I bypass general Web search for:

• Wikipedia.
• Specific domain-based searches.
• Google Scholar / Google Books.

As noted, bibliographies and references tracking is powerful, as is some simple, maintainable, and useful knowledge capture system.