In this paper, we investigate the effects of galactic extinction close to the south galactic pole and plate--to--plate matching errors. These represent the most probable systematic errors within the Edinburgh/Durham Southern Galaxy Catalogue (EDSGC) which could affect the angular correlation function. The distribution of extinction within the EDSGC area was obtained using the Stark {\it et al.} HI map and the IRAS $100\mu\rm{m}$ flux map. We find that the amplitude of the whole EDSGC correlation function varies by less than ${\rm \delta w}=0.003$ for reddening ratios $\rm{R_V}$ in the range $3.25\rightarrow6$. This corresponds to a range in average extinction of ${\rm A_B}=0.12\rightarrow0.24$. Simulations were carried out to assess the contribution to the correlation function from correlated and uncorrelated plate magnitude errors. The only simulation that affected the large--scale power seen in \w was for correlated plate errors with a systematic plate--to--plate offset of ${\rm \Delta m} =0.02$. This represented an overall $0.4$ magnitude difference between the ends of the EDSGC which was inconsistent with checks carried out with external photometry. All other simulations had an insignificant effect on the large--scale form of the correlation function. These tests suggest that the large--scale power seen in the EDSGC correlation function is due to intrinsic clustering and is not an artifact of the construction or location of the catalogue.