I have clipped out the block groups that are within the boundaries of select cities. My problem is, some of the block groups extend beyond the city limits and are thus divided in half after performing the clip. This has caused some noticeable discrepancies in the gathered census data. For example, some block groups which have been divided will retain information for the entire block group rather than reflecting values relating only to the city. This has caused some block groups that are in otherwise unpopulated areas to report unusually high values. Has anyone ever had this problem and if so what was your way around this? Some classmates have suggested the use of definition queries or selecting by location but this would exclude some sections of the city that actually contain the desired census data.