A Screening Protocol Before You Search

Deciding what counts as evidence first, so the search cannot be steered by what it finds.

elicit research systematic-review screening

What it is for

Decide what counts as evidence BEFORE searching. Otherwise the criteria quietly bend towards whatever the search returned, and the review concludes what the first page of results already said.

The skill

Write the protocol before running the search. A protocol written afterwards is
a description of what you found.

State, in writing:
1. **The question**, in PICO form if it is clinical, or as a single sentence
   with the population and the outcome named if it is not.
2. **Include if:** study design, population, date range, language.
3. **Exclude if:** the specific things you will not count — no control group,
   n below a threshold, secondary analyses, preprints.
4. **The extraction columns** you will fill for every paper. Decide these now;
   adding a column later means re-reading everything.

Then in Elicit:
- Let it extract the columns, and then **open the PDF for any row you intend to
  cite**. The extraction is a good index and a bad source of truth; a number
  pulled from an abstract is often the wrong arm of the study.
- Record the count at each stage — found, screened, included — as you go. It is
  the only part that cannot be reconstructed afterwards.

If a paper changes your question, that is a finding. Write it down as one rather
than quietly editing the protocol.

How to install it

1. Open the tool and start a new conversation, project or template.
2. Paste the text below wherever the tool keeps standing instructions
   — usually called system prompt, instructions, or persona.
3. Save it so it applies to every run rather than a single message.

Did it work for you?

Where it came from

Skills app Original