Screener questions don't support skip or branching logic. Every applicant answers every question, in the order you set, and you see all applicants with all of their answers.
This is deliberate, for a few reasons:
You keep the full picture. Screeners rank applicants rather than blocking them, so you see everyone who applied, including near-misses. An applicant who isn't a perfect match on paper can still be insightful to speak with, and the spread of answers gives you a rough read on the market.
It protects answer quality. Branching paths make it easier for applicants to work out which answers a study is "looking for". A flat list of non-leading questions makes gaming the screener much harder.
Scoring stays simple. Eligibility is scored from the multiple-choice options you mark approve or reject. Open-text answers carry no weight in the score. See how custom screeners work.
Many times, a participant who may not be eligible according to your screeners can be surprisingly insightful to consider speaking with.
How to handle questions that don't apply to everyone
Add a "Not applicable" option to the question so every applicant has something truthful to select. Mark it approve or reject depending on whether those people fit your study.
For example, if question 2 only makes sense for people who answered a certain way on question 1, include "Not applicable, I don't use this product" as an option on question 2 and mark it reject.
Having a holistic view of both eligible & ineligible participants in your list is also added data to show a rough representation of the market or population!
Looking for logic inside your study instead?
Unmoderated studies do support conditional logic between blocks, so you can route participants to different tasks and questions based on their responses once they're in the study. See conditional logic. That logic applies to study blocks only, not to screener questions.