Since mid-2007 we have carried out a dedicated long-term monitoring programme at 15 GHz using the Owens Valley Radio Observatory 40 meter telescope. One of the main goals of this programme is to study the relation between the radio and gamma-ray emission in blazars and to use it as a tool to locate the site of high energy emission. Using this large sample of objects we are able to characterize the radio variability, and study the significance of correlations between the radio and gamma-ray bands. We find that the radio variability of many sources can be described using a simple power law power spectral density, and that when taking into account the red-noise characteristics of the light curves, cases with significant correlation are rare. We note that while significant correlations are found in few individual objects, radio variations are most often delayed with respect to the gamma-ray variations. This suggests that the gamma-ray emission originates upstream of the radio emission. Because strong flares in most known gamma-ray-loud blazars are infrequent, longer light curves are required to settle the issue of the strength of radio-gamma cross-correlations and establish confidently possible delays between the two. For this reason continuous multiwavelength monitoring over a longer time period is essential for statistical tests of jet emission models.