The single isotropic scattering model proposed for the origin of seismic coda predicts a regularly decaying coda wave envelope. In real data however, this regular decay is often interrupted by non‐random arrivals (e.g. reflections) coming later in the coda. This additional energy may be due to clustered scatterers and represents a departure from the inherent assumption in the theory that the scatterers are uniformly randomly distributed. We have simulated the effect of clustered scatterers in synthetic seismograms in order to investigate their influence upon the measured coda Q (Qc). We show that the position and amplitude of such arrivals within the otherwise regularly decaying coda give rise to elevated Qc estimates. Often this additional energy in the coda is difficult to detect on a single isolated seismogram as evidenced by an example from the 1994 Northridge California earthquake. We therefore recommended that where possible, the use of seismic arrays (commonly reserved for P‐wave coda studies) be extended to S‐wave coda studies to aid in the detection of clustered scatterers, prior to the calculation of Qc.