So I love the idea of what they're saying: that there's an evidence hierarchy. Good stuff. But I'd structure mine a little differently.

1. Synthetic Research. This refers to research that synthesizes other research. It includes the following:
   1. Meta-analysis. A meta-analysis is a study of studies. The key part that makes it a meta-analysis is the published effect size. Rather than do an original study, a meta-analysis author will compile a set of other studies and measure their average effect.
   2. Systematic reviews. This is a narrative explanation of major research findings on a specific question but without an effect size. If the word "systematic" is used, that means the author has looked up every single paper on the specific topic. If the word "literature review" is used, that just means that they've looked through some of the research. Both are valuable, but systematic reviews are more deliberate.
2. Large experiments. Good experiments must be controlled (that is, must have a control and treatment group for comparison) and representative of the population in question. The easiest way to ensure representativeness is random selection. This is why we talk so often about randomized controlled trials, or RCTs, which are often viewed as the gold standard in academic research. That said, other methodologies are not RCTs that are still valuable--we will only focus on RCTs in this course.
3. Small experiments. A smaller RCT is still worthwhile, but not of the same statistical power as a large study.
4. Correlational or observational studies. This is when there is no control or treatment, but there is still data worth looking at. Often you'll hear things like "We found an association between these two variables." We can imply correlation, but not causation.
5. Expert opinion. Experts are smart. But just because it's an expert opinion doesn't mean it's right.
6. Lay opinion. Non-experts can be smart too--but they're subject to all the biases you've learned about. (So are experts' opinions!)