Freshman vs. Veteran
Whether time served changes how a member votes.
By Dallas AI Metrics, LLC · Analysis as of 2026-08-04 · Computed from 440 roll calls across 526 members · Methodology
A durable assumption in political commentary is that new members arrive independent and are gradually absorbed — that seniority produces conformity. An equally common assumption runs the opposite way: that newcomers, lacking standing and committee position, have the strongest incentive to follow leadership.
Both are testable against the voting record. Neither survives contact with it particularly well.
How the cohorts are built
Members are grouped by the year they were first elected to either chamber, not by time in their current seat. A long-serving representative who has just moved to the Senate counts as a veteran, because their legislative experience is what the comparison is about.
Three bands: two years or less, three to ten years, and more than ten. The boundaries are conventional rather than derived — there is no natural break point in the data, and different cut-offs would shift the composition of each group somewhat.
| Cohort | Members | Mean defection rate | Mean absence rate |
|---|---|---|---|
| Freshman (2 years or less) | 80 | 4.6% | 1.5% |
| Mid-career (3–10 years) | 228 | 5.8% | 3.5% |
| Veteran (more than 10 years) | 218 | 4.6% | 2.8% |
The result
The differences between cohorts are small. Small enough that the honest summary is that length of service does not, on this evidence, predict how independently a member votes.
That is a genuine finding, and a slightly deflating one. Both of the common narratives predict a clear gradient across these groups, and neither appears. Whatever drives a member's willingness to break with their party, it is not primarily how long they have been in the building.
Why the comparison is weaker than it looks
Even taking the small differences at face value would be a mistake, because the cohorts are not comparable groups that happen to differ in seniority. They differ systematically in other ways.
Survivorship. Veterans are, by definition, members who kept winning. Those who lost are not in the data. If a particular voting style makes re-election harder, that style is progressively filtered out of the senior cohort — and the difference would show up as an effect of seniority when it is really an effect of who survived.
Seat competitiveness. Long-serving members disproportionately hold safe seats, because safe seats are what allow long service. Safe-seat members face different incentives than marginal-seat members regardless of experience. The tenure comparison is partly a competitiveness comparison wearing a disguise.
Chamber and party composition. The cohorts do not contain the same mix of chambers or parties, and both of those affect voting behaviour independently. None of this is controlled for.
Wave elections. Freshman classes arrive in bursts, and a class elected in a wave year shares a political character that has nothing to do with being new. A cohort boundary that happens to enclose one such class inherits its character wholesale.
What would be needed to answer the question properly
Establishing whether seniority changes behaviour requires following individual members over time and observing whether the same person votes differently as they accumulate service. That is a longitudinal question, and this is a cross-sectional snapshot — different people at different career stages, measured once.
The corpus does not currently extend far enough back to support the longitudinal version. Until it does, the appropriate conclusion is the null one: no cohort difference large enough to be worth explaining.
Attendance is reported alongside party unity in the table above, and carries all the same caveats set out on the participation page — it is a rate, not a judgment. Full rules are in the methodology.