Hey everyone, I've been diving into Elicit over the past few weeks to help with some market research for a new campaign segmentation project, and I have to say, I'm really impressed so far. But there's one column in the results that I kept staring at without fully grasping: **"Strength of evidence."**
I understand it's a score (from 0 to 1), and higher is better, but *what exactly* is it weighing? Is it a measure of the journal's prestige? The study's sample size? The statistical significance (p-values) reported in the paper itself? Or something about the study design, like RCT vs. observational?
For context, I'm coming at this from a marketing ops background where we often score leads on "confidence" — but that's based on concrete data points like email engagement, firmographic fit, and activity thresholds. I'm trying to translate that kind of practical scoring logic to understanding academic papers. When Elicit surfaces 50 papers on a query like "effects of personalized email on conversion rates," I want to know which ones I should prioritize reading based on this score.
Could someone walk me through what goes into that metric? A real example would be super helpful. Like, if I have two papers:
- Paper A: Strength of evidence 0.95
- Paper B: Strength of evidence 0.72
What is Elicit *likely* telling me is the key difference between them? Is it that Paper A has a larger *n*? Or that it's a meta-analysis? Or that it's cited more?
I want to make sure I'm using this feature correctly and not accidentally overlooking good studies because I'm misinterpreting the score. Any insights from your own workflows would be amazing!
~Sarah
Data is the new oil