Exercise 1: Why Published Criteria Resist Cherry-Picking — Possible Solution
===============================================================================

Both review types could, in principle, be built from the exact same
pool of available studies - the real difference is not what's
available, but how visibly and accountably the selection happens.

In a narrative review, the author decides which studies to include as
they write, with no requirement to state the selection rule in
advance. If a study contradicts the author's own expected conclusion,
it can simply be left out with no one else able to detect that
omission - the selection process is invisible to the reader.

In a systematic review, the inclusion and exclusion criteria (date
range, study design, population, outcome measures) are defined and
published before the search results are even reviewed for content.
Once published, a chosen study either does or doesn't meet the stated
criteria - there's no room to quietly drop a study that meets the
criteria but produced an inconvenient result, because the criteria
themselves, not the author's own preference, determine inclusion. A
reader can also independently check whether the review actually
followed its own stated criteria, which a narrative review's own
undocumented process doesn't allow.

The key mechanism: pre-registering the selection rule removes the
author's own opportunity to selectively apply it after already seeing
which results support or contradict their expectations.

ANSWER: A systematic review's inclusion/exclusion criteria are set and
published before the studies are evaluated for their results, so a
study can't be quietly dropped just because its finding is
inconvenient - the criteria, not the author's own preference, decide
inclusion. A narrative review sets no such rule in advance, so the
selection process is invisible and unaccountable, even if it covers
the same underlying pool of studies.

WHY THIS WORKS AS AN ANSWER
------------------------------
This explains the actual mechanism (pre-registration removing
after-the-fact discretion) rather than just restating that one process
is "more structured," directly answering what the exercise asked.