We present a study on the effect of undetected stellar companions on the derived planetary radii for the Kepler Objects of Interest ( KOIs ) .
The current production of the KOI list assumes that the each KOI is a single star .
Not accounting for stellar multiplicity statistically biases the planets towards smaller radii .
The bias towards smaller radii depends on the properties of the companion stars and whether the planets orbit the primary or the companion stars .
Defining a planetary radius correction factor X _ { R } , we find that if the KOIs are assumed to be single , then , on average , the planetary radii may be underestimated by a factor of \langle X _ { R } \rangle \approx 1.5 .
If typical radial velocity and high resolution imaging observations are performed and no companions are detected , this factor reduces to \langle X _ { R } \rangle \approx 1.2 .
The correction factor \langle X _ { R } \rangle is dependent upon the primary star properties and ranges from \langle X _ { R } \rangle \approx 1.6 for A and F stars to \langle X _ { R } \rangle \approx 1.2 for K and M stars .
For missions like K2 and TESS where the stars may be closer than the stars in the Kepler target sample , observational vetting ( primary imaging ) reduces the radius correction factor to \langle X _ { R } \rangle \approx 1.1 .
Finally , we show that if the stellar multiplicity rates are not accounted for correctly , occurrence rate calculations for Earth-sized planets may overestimate the frequency of small planets by as much as 15 - 20 % .