Form a testable question
State what quantity you expect to change, what you will measure, and what would count as surprising evidence.
Research skills
Make claims that your data can answer.
Research training starts before advanced mathematics. It starts with a question that can be tested and a record someone else could follow.
Control variables
Change one parameter at a time before trying combinations. Otherwise, you cannot tell which cause produced the effect.
Read graphs as evidence
Record trends, scales and transitions—not merely screenshots. Ask whether the graph supports every part of your claim.
Know a model's limits
List its assumptions, then deliberately seek a condition where the conclusion may stop applying.
Keep a reproducible log
For each trial, record parameter values, initial conditions, an observation and an interpretation separately.
Communicate uncertainty
Distinguish what the data shows, what the model predicts, and what remains an open question.