There are three methods used to detect extrasolar planets.
The first is astrometric, precise stellar positions are determined to try and detect the gravitational tug of an orbiting planet by a shift in the apparent position of the star relative to its neighboring stars.
The second is spectroscopic, where accelerations caused by planetary gravity are detected as Doppler blue or red shifts in the star's spectrum.
The third is the transit, or eclipse method; where a star's brightness dims a tiny amount when its planet passes between us. This method, unfortunately, only works when the ecliptic (orbital) plane of the planet is directly in our line of sight. If the planet passes above or below the stellar disc we will never know its there.
Each of these methods favors certain solar system geometries and requires different observational programs. For example, large planets orbiting very close to small stars are easier to spot. So the types of planets we have discovered so far are not a fair sample of the type of planets that are probably out there.
Earth-type planets orbiting sunlike stars in reasonable, well-behaved orbits are particular difficult to see. Large gas giants in highly elliptical orbits around small stars tend to be represented misproportionally. Another "selection effect".
Space/Science » in reply to Kepler 62f
The third way -- an extrasolar search primer.
