NZCVS-Cycle-5-How-much-crime.xlsx [xlsx, 91 KB]
...other hand, when the confidence intervals of two estimates do overlap, the difference between the estimates is described as not statistically significant. This is a more conservative approach than a formal statistical test. Don't know/refusal responses Residual response categories such as "don't know" and "don't wish to answer" are generally excluded from analysis. This is generally the case for analysis of victimisation and rates of reporting victimisations...