Characterizing the dependence of the orbital architectures and formation environments on the eccentricity distribution of planets is vital for understanding planet formation .
In this work , we perform statistical eccentricity studies of transiting exoplanets using transit durations measured via Kepler combined with precise and accurate stellar radii from the California- Kepler Survey and Gaia .
Compared to previous works that characterized the eccentricity distribution from transit durations , our analysis benefits from both high precision stellar radii ( \sim 3 % ) and a large sample of \sim 1000 planets .
We observe that that systems with only a single observed transiting planet have a higher mean eccentricity ( \bar { e } \sim 0.21 ) than systems with multiple transiting planets ( \bar { e } \sim 0.05 ) , in agreement with previous studies .
We confirm the preference for high and low eccentricity subpopulations among the singly transiting systems .
Finally , we show suggestive new evidence that high e planets in the Kepler sample are preferentially found around high metallicity ( [ Fe/H ] > 0 ) stars .
We conclude by discussing the implications on planetary formation theories .