I know that standard deviation can be measured for a set of repeated experiment but if i have only the data of a single measurement can i measure standard deviation of its slope and intercept? Could any one help me to answer on this question?
The quick answer would seem to be no. Perhaps more information about the measurement might help. With one measurement (or even an infinite number with no change of at least one independent variable) how can you even get a slope or intercept?
thank you so much prof Myron for your kind answer, but i did not mean that i have only one measurement i meant i have a complete experiment without repeating it several time i.e one curve. Can i measure standard deviation of its slope and intercept of these data to measure the error even i did not repeat the experiment several time?
Sure you can, althogh it is not the "standard deviation" but the "standard error" of the parameters (intercept and slope). It is called "standard error" because we are talking about the variability of estimates, not of data. You can see the mathematics here: http://en.wikipedia.org/wiki/Simple_linear_regression
However, any statistics software will give you these values. Even Excel does it if you use the function LINEST (see Excel help for usage and details). If you google for "online linear regression" you will even find free online tools doing these calculations.
I like to use the "Regression" function in the Excel "Data Analysis Toolpack". Does the same thing as LINEST, but better looking output. Almost every Anaytical Chemistry book will have a discussion of the regression analysis.
- Badges
- Science topic