Behavior Hiring
Research brief

Does onboarding reduce early turnover?

A research brief on the 2025 newcomer socialization meta-analysis: which adjustment factors relate to turnover intentions and actual exits, how strong the evidence is, and what it implies for hiring.

Short answer

Perceived fit has the strongest relationship with turnover intentions of any newcomer adjustment factor (r = −.31), followed by social acceptance (−.27) and role clarity (−.25). The relationships with actual turnover are considerably weaker: perceived fit −.16, social acceptance −.12, task mastery −.10, role clarity −.09.

Key findings
  • Perceived fit has the strongest relationship with turnover intentions of any newcomer adjustment factor (r = −.31), followed by social acceptance (−.27) and role clarity (−.25).
  • The relationships with actual turnover are considerably weaker: perceived fit −.16, social acceptance −.12, task mastery −.10, role clarity −.09.
  • Structured, institutionalized onboarding has a direct negative effect on turnover intentions (β = −.18). Mentoring shows an independent effect (β = −.11).
  • Social acceptance is the main pathway. The authors summarize the finding as "the people make the place."

Onboarding is one of the few hiring-adjacent practices almost everyone agrees is important and almost nobody measures. The 2025 meta-analysis of newcomer socialization by Bauer and colleagues is the most complete picture available, and it is more useful for what it shows about mechanism than for any single headline number.

What the evidence actually supports

The strongest claim this research supports is about intentions, not exits. Newcomers who feel they fit, who are socially accepted, and who understand their role report substantially lower intention to leave. That is a real finding, measured across a large body of studies.

The link to people actually leaving is present but modest, and it rests on far fewer studies. The actual-turnover correlations come from roughly four or five studies, against a much larger base for the attitudinal outcomes. Anyone citing this literature should be clear which of the two they are citing.

This matters because of how onboarding research gets quoted in practice. Figures like "a third of new hires leave within a year" circulate widely and come from survey firms rather than this meta-analysis. They may well be roughly right. They are not what the peer-reviewed socialization literature establishes, and conflating the two is how a reasonable case gets built on a weak foundation.

The mechanism worth noticing

If you read the effect sizes in order, a pattern appears that is more useful than the individual numbers.

Task mastery, the thing most onboarding programmes are actually built around, has the weakest relationship with turnover intentions of the four adjustment factors (−.14). Perceived fit and social acceptance, which most programmes address incidentally if at all, have roughly twice the association.

The practical reading: a new hire who can do the work but doesn't feel they belong is at more risk than one who belongs but is still learning the work. Most onboarding agendas are built the other way round, heavy on systems training and light on deliberate social integration.

Where hiring connects

Perceived fit being the strongest factor has an implication that runs backward into selection.

Fit here is the newcomer's own sense that this is the right place and the right role for them. That perception is formed partly during onboarding, but it is substantially set earlier, by what the candidate was told the job would be. A hiring process that oversells, or that never discusses what the role actually demands day to day, is building a fit problem that onboarding then has to repair.

This is the argument for naming what a role is not during hiring rather than after it. A candidate who hears honestly that the work is solitary, or that the first month is ambiguous, either self-selects out or arrives with accurate expectations. Both outcomes beat discovering it in week three.

It is also the argument for carrying what you learned in hiring into the first 90 days. The behaviors you rated, the gaps you accepted, the areas two interviewers disagreed about: that is a map of where this person's adjustment is most likely to be difficult, and it is usually filed and forgotten the day the offer is signed.

What this does not say

That onboarding programmes reduce turnover by a specific amount. The actual-turnover evidence is too thin for that claim, and anyone offering you a precise percentage is drawing on something other than this literature.

That onboarding can fix a bad hire. Nothing here supports that. The adjustment factors describe how well someone settles into a role they are suited to.

That structure alone is enough. Institutionalized tactics help (β = −.18), but the largest correlations run through social acceptance. A well-documented programme that leaves someone isolated is missing the main mechanism.

What to do with it

Front-load social integration rather than systems training. The evidence points at belonging, not competence, as the lever on early attrition.

Assign a mentor deliberately. It shows an effect independent of general onboarding structure, which is rare in this literature.

Measure intention, not just exits. Turnover is a lagging indicator with a small sample in any one organization. Intention to leave moves earlier, correlates more strongly with everything here, and can be asked about directly at 30 and 60 days.

And be honest during hiring about what the role is. Perceived fit is the strongest factor in the data, and it is largely set before anyone starts.

Method

A narrative review of a published meta-analysis: Bauer, Erdogan, Ellis, Truxillo, Brady and Bodner (2025), Journal of Management. Correlations are as reported there. The actual-turnover estimates rest on a small number of studies (k = 4 to 5) and should be treated as indicative. Turnover in the underlying studies was typically measured over intervals of roughly 15 to 27 weeks, so this evidence speaks to early attrition rather than long-run retention.

Sources

  1. Bauer, T. N., Erdogan, B., Ellis, A. M., Truxillo, D. M., Brady, G., & Bodner, T. (2025). Newcomer socialization: a meta-analysis. Journal of Management.

How to cite this page

Behavior Hiring. (2026). Does onboarding reduce early turnover?. https://behaviorhiring.com/research/onboarding-early-turnover/

Monthly Research Brief

One email a month. New research, a practical tool, and what changed in the field.

No spam. Unsubscribe any time.