I
Im_leary
Guest
I am in complete agreement with vic. After reading the test conditions several times, I honestly do not understand why it appeared in this periodical. I suppose their test criteria is substantially lower than PS magazine.
This is a test that, if done properly, would require significant resources. The number of components, and the real impact of the variations in those components would require a very large sample population in order to the minimize the effects of the other components on the desired result of determining the effect of change ONLY the cartridge case. You must not dismiss the possibility of a synergistic effect of one component or another, with any one of the cartridge cases tested (in other words, for some reason, if one case or another has a real, measurable, affinity to a different brand of powder... the results would be horribly skewed... as just one of many examples I can think of). Considering this, the mathematics of determining the minimum sample size alone would be an ambitious undertaking.
In this test, at face value, the test conditions were far too limited (several test rifles should have been used), the test sample population was ridiculously too low, no base-line was used, and the performance of the actual test (if it was described in it's entirety) was not adequately controlled. Actually, I think vic was being quite kind in his description of the test. My statistics professor would have called the whole test, "...statistically meaningless". There is quite a difference in trying to derive some sort of scientific accuracy in a properly structured test, and simply relating the results of a poll. This sort of test would fit right in with the methodology of some of the Philadelphia gun control polls I have read... it is an opinion and, as such, is flawed from the outset.
Just for the record, I have found Lake City Match (in .308) were the most consistent in construction and resultant accuracy. My sample group was > 10,000.
This is a test that, if done properly, would require significant resources. The number of components, and the real impact of the variations in those components would require a very large sample population in order to the minimize the effects of the other components on the desired result of determining the effect of change ONLY the cartridge case. You must not dismiss the possibility of a synergistic effect of one component or another, with any one of the cartridge cases tested (in other words, for some reason, if one case or another has a real, measurable, affinity to a different brand of powder... the results would be horribly skewed... as just one of many examples I can think of). Considering this, the mathematics of determining the minimum sample size alone would be an ambitious undertaking.
In this test, at face value, the test conditions were far too limited (several test rifles should have been used), the test sample population was ridiculously too low, no base-line was used, and the performance of the actual test (if it was described in it's entirety) was not adequately controlled. Actually, I think vic was being quite kind in his description of the test. My statistics professor would have called the whole test, "...statistically meaningless". There is quite a difference in trying to derive some sort of scientific accuracy in a properly structured test, and simply relating the results of a poll. This sort of test would fit right in with the methodology of some of the Philadelphia gun control polls I have read... it is an opinion and, as such, is flawed from the outset.
Just for the record, I have found Lake City Match (in .308) were the most consistent in construction and resultant accuracy. My sample group was > 10,000.
Last edited by a moderator: