4 ms·

> If you sample on the interval 1990 to the present, crime appears to have gone down. If you sample from 1960 or any longer intervals to the present, per capita

by glenra 13y ago

> If you sample on the interval 1990 to the present, crime appears to have gone down. If you sample from 1960 or any longer intervals to the present, per capita crime has certainly increased. How about instead of SAMPLING, and without performing any sort of "moving average", you just SHOW the homicide rate, and do so over a range wide enough to INCLUDE 1932. Can you do that? If you do that, I think you will find that you are wrong about this - the homicide rate today is lower than it was in 1932. Because there was a huge peak around the 1930s PRIOR TO the relatively low point in the 1950s. I don't see much point in applying "a reasonable moving average" since the underlying trend is pretty smooth, but if you did, say, a 5-year moving average it would STILL be lower today than in 1932.

lutusp 13y ago
> How about instead of SAMPLING, and without performing any sort of "moving average", you just SHOW the homicide rate ... The answer is simple -- it serves no purpose. If the point is to establish historical trends, and in particular if it's desirable to create a future trend line, the first thing to do is avoid the classic mistake of examining the raw data. (This is true in climate studies also, for the same reasons.) In this case, a moving average meant to confirm or refute the thesis that crime rates are higher now than in the past, is best conducted with a long interval. Such an interval shows that crime per capita is much, much higher now than it was in the past, along with population (which also fluctuates over time). The 1930s were a historical anomaly -- the Great Depression (an increased level of social upheaval and desperation), and prohibition coming to an end (which meant that gangsters had to think of a new way to make a living), arriving at once, make it unrepresentative. Any statistician hoping to establish social trends would treat that period as an obstacle to producing a meaningful trend line. > I don't see much point in applying "a reasonable moving average" Yes, I got that. And I won't be likely to successfully explain to you why it's needed.

by glenra 13y ago

lutusp 13y ago

> How about instead of SAMPLING, and without performing any sort of "moving average", you just SHOW the homicide rate ... The answer is simple -- it serves no purpose. If the point is to establish historical trends, and in particular if it's desirable to create a future trend line, the first thing to do is avoid the classic mistake of examining the raw data. (This is true in climate studies also, for the same reasons.) In this case, a moving average meant to confirm or refute the thesis that crime rates are higher now than in the past, is best conducted with a long interval. Such an interval shows that crime per capita is much, much higher now than it was in the past, along with population (which also fluctuates over time). The 1930s were a historical anomaly -- the Great Depression (an increased level of social upheaval and desperation), and prohibition coming to an end (which meant that gangsters had to think of a new way to make a living), arriving at once, make it unrepresentative. Any statistician hoping to establish social trends would treat that period as an obstacle to producing a meaningful trend line. > I don't see much point in applying "a reasonable moving average" Yes, I got that. And I won't be likely to successfully explain to you why it's needed.