Given a biconvex converging lens, we have that R = R1 = R2, so the lensmaker's equation simplifies to
1/f = (n-1)(2/R)
1/f = (1/2)(2/R)
1/f=1/R
f=R
So we have R is 15 cm for the experimental value.
Now given the erroneous assumption that fcomputed=R/2, we would have fcomputed= 15cm/2 = 7.5 cm.
This gives fexperimental / fcomputed = 0.5, or exactly half the experimental value.