An article provided by Business Report brings speculation to the 2012 census research findings. Business Report states that "statistical analysis must be carried out in a responsible manner. It requires methodological soundness, theoretical coherence and proper articulation of results". The article states that there are numerous oddities in the 2012 census data. Some people seem to be "unaffected" by these oddities showing no concern/ questioning about the results, while others have pointed them out and consider the census to be misinforming, therefore questioning the census 2012 in its entirety.
One of the oddities discussed regards measurement and reporting comparisons in 2012 household income findings by comparing to 2001 census data. The 2012 census refers to the census 2001 (which used different measurements) to claim household income has doubled since 2001 for a particular demographic. This can be described in our book as using different "metrics". In 2001 the census household income was based on personal income and categorized in 12 income classes, whereas census 2012 refers to the household's entirety income. These differences were not stated in the 2012 census, but were realized by various analysts.
This seems to be a fault in research reporting with statistical analysis. The census is interpreting information with different metrics and comparing them to show a relationship. If the 2012 census wanted to accurately report an increase in income, what should have been compared in 2012 are individual incomes to show a reliable and valid relationship over time.
We talked about reporting research in our textbook and how it needs to be accurate, without bias and without reason to misinterpret data. Based on this article, it seems that the census used different metrics to compare data, which can not be done. It was interesting to learn how analysts caught this "misinformation" and that the general public would be more likely to not notice the fault in 2012 census reporting. This article shows how research institutions can sometimes bias the information to prove a point when the actual way of reporting relationships is not valid. In this case it was in terms of metrics, using different formats of data collection to draw a conclusion and show a relationship, when the formats of data collection can not be compared.