MSA ed3 page 93 re: Question about Linearity formulas

[mcada]

Hi everyone,
just would like to ask for help regarding the formula for slope a, intercept b and the upper and lower confidence bands...all formulas use the variable gm

1. does gm means g (# of parts) * m(# of trials) ?
if so, i got different results when I try to calculate the example on pages 94-95

...but when i use g only instead of gm i got the correct result

...attached is my sample worksheet, BTW, in the worksheet I created a simple vba function that calculates the significant t value, actually its just sort of an interpolator of XL's TINV function because TINV only worked with integers

 

[Marc]

I rasn across this oldie with no response. i know it's long over due, but anyone want to take a shot at it anyway?

 

[Hershal]

I would love to try it, but don't have the book. Graeme might be the best to address this.

 

[shard923]

I am unsure after looking at your spreadsheet, what are you trying to accomplish here? I looks like you are trying to perform linear regression. IE Trying to interpolate or predict a linear value for non-linear data. Is this the case or am I way off the mark?

 

[Alexandre Bonatto]

About "m" and "g"

I believe that "g" is the number of samples (number of subgroups, clusters, ...) and "m" is the sample size. In bias study (p.85) - item 5 - it says that g=1 and m=n (number of measurements). If data is collected through any kind of chart (Xbar & R, ...), "m" will be a constant (the size of each sample collected) and "g" will be the number of samples registered in the chart.

Do you agree?