Limitations
Last but not least, it is necessary to recognize the limitations of the correlations that were
obtained. It is safe to assume that correlation coefficients of close to one in magnitude are no
coincidence, but with r values of .3, it is better to be skeptical. Another problem to take into
account is the fact that the entire population did not answer the long version of the census so
the reported numbers are not complete. Also, because some people may rush through filling
out the census, it is entirely possible that they may misinterpret the questions they are asked.