When a research organization wants to know what became of the people it trained, the natural first step is to ask them. Send a survey: where are you working now, in what role, and did your time with us help?
It’s a reasonable instinct, and surveys have real strengths. But as evidence of what happened to your alumni as a group, a survey has a problem that sets in before anyone answers a single question.
The bias starts with your contact list
Consider who a survey can actually reach.
Institutional email stops working almost as soon as someone graduates or moves on. Personal addresses change. Within a few years, a contact list built during training points mostly nowhere. The people you can still reach are the ones who stayed connected: they updated their details, joined an alumni network, or still collaborate with you.
Those alumni aren’t a random slice. They tend to be the ones still working in your field, often still in academia, and still close to the programs and facilities they came from.
Then comes who replies
Among the alumni you can reach, not everyone answers.
A survey about your program looks relevant to people whose careers still connect to it. A postdoc in the same discipline will recognize themselves in the questions. Someone who moved into finance, policy or a different industry is more likely to assume the survey isn’t meant for them, and move on.
So the sample narrows twice: first to who you can reach, then to who chooses to reply.

The skew runs in one direction
Both filters push the same way. The alumni most likely to end up in your results are the ones whose careers stayed closest to you.
That means a survey can overstate how many of your alumni stayed in the field, and how much they credit your programs for where they landed. Nobody is being dishonest. The skew comes from who ends up answering.
There’s a harder problem underneath. The alumni who are missing are exactly the ones you can’t reach, so you can’t measure how big the skew is. You know your results lean. You can’t say how far.
What surveys are good for
None of this makes surveys useless. Some things only your alumni can tell you: why they chose their career paths (staying in research, moving into industry, taking a role abroad), which skills from their training they actually use, how they felt about the experience, what they earn, and what they’d tell the next cohort. No public record can answer those questions.
The trouble comes when a survey is asked to do a different job: to stand in for where everyone you trained ended up. For that question, there’s a better starting point.
Start from records, not replies
Instead of beginning with the alumni who answer, begin with everyone you trained.
Most organizations can put together a list of names, affiliations, degree levels and the year each person was associated with them. If you don’t have one, a sample can often be rebuilt from publications. From that list, you look for career records people have chosen to publish: professional profiles, university and lab pages, publication records. These show roles, sectors, locations and further degrees. Then you confirm each match against name, institution and degree dates. Nobody has to reply, and nobody has to have stayed in touch.
In our experience, LinkedIn profiles alone typically confirm career records for 40–55% of an organization’s alumni. Adding other public sources raises that further.
“But you only found half of them”
It’s a fair question, and a smaller one than it sounds.
First, 40–55% is a large sample, and that’s from LinkedIn alone. Compare it with the share of your alumni who answered your last survey.
Second, and more important, you can check it. What matters isn’t finding everyone. It’s knowing whether the people you found look like the people you didn’t. Because you started from the full list, you can compare the group you found against your complete list on things you already know: degree level, cohort year, program. Where gender can be estimated for this purpose alone, gender bias can be checked too.
If the group you found mirrors the full list, you can say so, and your results stand on firmer ground. If it doesn’t, you know which way it leans and by how much, and you can account for it.
Public records have gaps of their own, and they vary by career path. Researchers who stay in academia are often consistent about keeping a public profile, but more often on university pages or publication records than on LinkedIn. With a survey, you can only guess who is missing, but starting from the full list lets you better understand and characterize the gap.
If you’re planning an alumni survey
A survey is still the right tool for some questions. A few things make its results more useful.
When you write it:
- Ask what only alumni can tell you. Why they chose their career paths, which skills they’ve used, what they would change about the training, and, if it matters to you, what they earn.
- Ask about experience, not credit. “Did your time with us help your career?” invites a polite yes. “Which skills from your training have you used in the past year?” gives you something more concrete.
- Keep it short, so alumni with the weakest connection to you still get to the end.
- If you want depth, choose who to invite. For interviews or detailed follow-up, pick alumni who reflect the full range of careers, rather than relying on whoever volunteers.
When you read the results:
- Count against everyone you trained. Report how many responded out of your full list, not just out of the people you could contact.
- Check who answered. Compare respondents with your full list on degree level, cohort year and program. If they differ, say so.
- Describe the respondents, not all alumni. Write “among alumni who responded,” not “our alumni.”
- Use surveys to explain, not to count. Surveys are good for telling you why. They’re less reliable for telling you how many.
Used this way, surveys and career records each do the job they’re good at. The records show where your alumni went, and the survey helps explain why.
At Alumni Analyzer, we start from records, not replies: we trace the career outcomes of the people you trained and compare them against a matched group of peers. If you’d like to see what that could show for your organization, book a free consultation at alumnianalyzer.com.
