Re: Underestimated stats


[Prev][Next][Index][Thread]

Re: Underestimated stats



In article <797nu5$9vl$1@brokaw.wa.com>,
          jcole@halcyon.com ("John C") wrote:

] I have a question.
]
] Why does the TI-83 under estimate the standard dev. for a short list of
] numbers such as: 6, 7, 8, 9,?
]
] I get a sample dev. of 1.29, and a pop. dev. of 1.12, but hand calculated
] the sample dev. is 1.66.
]
] How can I improve the accuracy of the TI-83 when calculating stats.?


Well, I think that your 83 is correct but your hand calculation has gone a
bit awry.:-)  Sorry!  Let's see if I can set it out:

Mean is (6+7+8+9)/4 = 7.5

Variance is ((6-7.5)^2+(7-7.5)^2+(8-7.5)^2+(9-7.5)^2)/4 = 5/4 = 1.25

Population SD is the square root of 1.25 = 1.118, which agrees with your 83.


For the sample SD divide by 3 instead of 4, giving:

Sample SD = \sqrt{5/3} = 1.29, which again agrees with your 83.


I often feel that this is one of the good things about using calculators like
this; it enables you to check your own calculations easily which can promote
understanding.  Good, isn't it?

Dick

PS  \sqrt{} is LaTeX speak for the square root of whatever is in {here}!
    Using LaTeX you can express any maths perfectly well in plain ascii.

Dick


--
=============================================================================
Dick Smith                                           dick@risctex.demon.co.uk
Acorn Risc PC                                  http://www.risctex.demon.co.uk
=============================================================================


Follow-Ups: References: