Sunny Boy wrote:
How could leaving out 26 murders "skew stats" except if a false impression of murder is wanted ? How could the murders of 26 people be an "anomaly" in murder stats ?
They were not left out of the Connecticut report, they were not included in the regular murder stats hence the reason they do not appear in the FBI data. If I recall correctly there was whole section of the report dedicated to the Sandy Hook murders and the heading included something about them. It's not like they were simply swept under the rug.
It becomes an anomaly if it significantly impacts the results of statistical analysis. When analyzing data you need to find anomalies like this, you either adjust or throw it out completely. The FBI crime data is used for a lot of things like realty and I would imagine Connecticut made conscious effort to keep those murders out of the regular stats.
For example the small town my Grandmother lives in might have a murder every decade and if we use the average over many decades it's relatively safe place. They had triple murder from out of towner a few years back. For that year the town was listed in the top 20 most dangerous places in PA. While the raw statistics support it that is just not realistic. It would only be realistic if that trend continued over a decade.
Of course you need to be careful with things like this, that is why I said be very careful with any crime data. Leaving the Sandy Hook murders out the regular murder statistics makes sense generally however if you were doing study on say on how gun laws affect the murder rates it'snot something you would want to leave out.