Does this program want you to contact the investigator?
NIH SIP explicitly encourages contact with relevant investigators after completing the application. NSF REU students apply through individual sites or the specified supplement route. Start with the program's instructions: direct outreach can be useful, but it does not replace an application or override a request not to contact investigators.
For a home-campus lab without a formal recruitment process, look for its undergraduate research instructions and the department's research directory. Ask the contact named there about participation. A faculty member's research page is a starting point for learning about the work, not evidence of a funded vacancy.
Use a small shortlist you can explain
For each lab, write one sentence about its question, one sentence about your relevant preparation and one question you would ask. Read a public project description and then enough of a related paper to understand the method and its limits. Be precise about what you actually read.
A practical starting shortlist might contain three groups: one using a method you already know, one using that method on a different problem and one whose method you want to learn. That is a search exercise, not an application quota. Drop a group when you cannot find a genuine connection or when its participation requirements do not fit.
The point is not to know the whole field before contacting anyone. You should be able to explain why you chose this group without saying only that its university is prestigious.
Send a short message with a real research connection
Invented example for a lab whose public summary discusses missing environmental observations. The sender has read that summary, completed a statistics project and submitted the required program application. Replace those facts with your own.
Subject: Summer research inquiry — missing environmental observations
Hello Professor [Surname], I am a second-year student interested in how missing observations affect environmental analyses. I read your group's public summary on that problem. In a statistics course, I compared two ways of handling missing values and documented where they changed the result. I have submitted my application to [program]. Are you considering undergraduates through that program, and would basic Python analysis be useful preparation for your current work? I have attached my résumé. Thank you, [Name].
The message says what was read, what the student has done and what answer would help. It does not pretend that reading a summary equals studying the full paper. When there is no central program, replace the application sentence with your actual availability and ask about the lab's undergraduate route.
Offer one small piece of work you can defend
A useful sample could be a short reproducible analysis, a careful literature comparison or a documented class project. Include the question, your contribution, the method and one limitation. A reviewer should not have to search a folder of unrelated files to discover why you sent it.
For a team project, label what you did. For a new method, say what you have learned so far and what would require training. Do not attach private lab records or restricted datasets; use work you are permitted to share. The aim is to give the prospective mentor an accurate starting point for training you.
Use the conversation to test the mentoring arrangement
Ask: 'What would I work on first? Who would review it? How often would we meet? What hours and dates are required? Is funding available for this particular position?' The answers tell you whether an opportunity exists and whether the supervision fits a beginner.
Compare two hypothetical replies. 'Read whatever interests you and check in at the end of summer' leaves scope and feedback unclear. 'Begin by reproducing this result, meet with a graduate student twice a week, then choose a small extension' gives you a defined first task and a route for getting help. A famous supervisor's name does not answer those practical questions.
After a meeting, send a short note recording any agreed next step and its date. Without a reply, a single brief follow-up after roughly a week can be a reasonable personal rule unless the program gives different instructions. Then continue your search rather than repeatedly emailing the same person. Before changing classes, leaving a job or arranging housing, obtain the actual participation and funding terms.