Since large cities were over sampled in the survey, the posts tratification weight was used for compensating
this bias. The weighted results show a larger sample size (2,215.25) than unweighted sample
size (902), which may adversely affect inferential data and standard errors. We normalized the
weighted sample size to equal the unweighted sample size for correcting this bias, multiplying the
weights by (unweighted sample size)/(weighted sample size).