UUÂãÁÄÖ±²¥

AMC Statistical Software

To use a Minitab macro, save it in your Minitab macros folder as a .mac file, and call it from a Minitab Session Window with the command <filename> <argument list>
No file extension is needed in the command.
If you save it elsewhere, or with a different extension, you will have to use the full pathname in the command, ie %<pathname> <filename> <argument list>
The % sign (with no space after it) tells Minitab to look for the macro. 

Related Links

Download a free fully functional 30-day demonstration version of Minitab

AMC Excel Add-ins

AMC Software now includes Add-ins for MS Excel, so that you can use apply AMC statistics functions within MS Excel.

FAQs about AMC add-ins

What is an add-in?
An Add-in is a small program that adds new functions to MS Excel. Once installed, the add-in loads automatically every time MS Excel starts up. 

What do AMC add-ins do?
AMC add-ins will typically provide one or both of:

• A feature that can be accessed from the Excel menu, and is controlled by an Excel dialogue box in much the same way as the File...Print command works. These appear on their own extra menu bar, the "AMC Statistics" menu, which is automatically added when you install an AMC add-in. This is the most common way of accessing an AMC add-in.
• Additional spreadsheet functions that can be inserted in the spreadsheet, just like the AVERAGE() or STDEV() functions built in to MS Excel. 

What are the different files for?
Each AMC add-in application typically consists of:

• the add-in itself, a file with the extension .xla
• a Help system, which can be called up from MS Excel. This typically consists of a traditional Windows Help file (.hlp or .chm) and Help Contents file (.cnt), but might be in another form 
• a 'Readme' file; a small text file which describes briefly how to install the add-in and access the Help system

AMC Add-ins are available as compressed 'zip' files or in their normal state. The Zip files contain all the files needed, and are packaged into a smaller space for faster download. You need an application such as PKzip, WinZip or ArjFolder to 'unzip' the files. Alternatively, you can download the required files directly.

How is an add-in installed?
First, download the add-in from the AMC website. If it is in .zip form, unzip it to give the individual files. The instructions for installing each add-in will be included in the text file "Readme" (or "readme.txt"). This file is a simple text file and will open if you simply double-click on the file name. 

A typical add-in installation includes the following steps:

1. Copy the .xla, .hlp and .cnt (executable, Help and help Contents files respectively) to a directory of your choice. You can use the directory you downloaded the application to if you wish. 

2. In MS Excel, add the Add-In to Excel's list. You may need to refer to Excel's Help system; for Excel 97, the procedure is: 

   A) Choose Tools, Add-ins to display the Add-Ins dialogue box
   B) Click Browse, and navigate to your copy of the .xla file
   C) Click Open.   

3. The new add-in should appear in the list of add-ins. Make sure the check box next to its entry is selected. The add-in will now load. In future, it will load automatically when you start Excel. 

Using an AMC add-in from Excel
If you have no AMC Statistics add-ins installed already, the add-in will normally create a new menu entry, AMC Statistics, on the main Excel menu bar, and add the new AMC add-in menu to that. If you already have an AMC statistics menu, the new menu will simply appear as an additional option. Simply click on the menu and select the option you want to use.

Accessing the add-in's help system
The add-in's menu will normally include a direct link to a Windows Help file. If Windows cannot immediately locate the help file, Windows will ask you to find it; simply browse to the directory you copied the add-in files to. After that, windows will be able to find the file by itself.

The help file will contain all the instructions for using the application. Use the Contents button to browse through the help system. Hitting Find will cause a full-text index to be generated in the usual way.

Add-ins in accredited and regulated laboratories
AMC add-ins are provided without warranty of any kind and the user is entirely responsible for assessing their suitability for a particular purpose. To help with this, the following information is included in the Help system or associated documentation:

• Reference to the particular AMC Brief or technical publication which describes the statistical procedures and their properties.
• A statement of the validation checks and tests performed on the application prior to release.
• A listing of any known issues which may affect operation, with solutions if known.

In addition, the AMC software website includes test sets with known answers which can be used to confirm correct operation on the user's system.

Technical support for AMC add-ins
No technical support is provided for AMC software. However, the help system will include a list of any known problems and 'work-rounds', so is worth browsing through if you encounter a problem.

The help system or documentation may also include an email address to which 'bug reports' may be sent, either directly or via the AMC secretary. This is not intended for user support; it allows the software author(s) to collect information on any problems, and may at the author's discretion result in improvements to the software in due course. "Bug reporting" is therefore welcome.

A Minitab local macro to calculate robust mean and standard deviation

This macro calculates Huber's 'H15' estimators for robust mean and standard deviation. 
Written for Release 12, will run on Release 10 and 11.
Date last modified: 05/02/01.
Input data: column of values.
Output: Robust mean and standard deviation written to Session window.
The robust technique is designed for medium sized data sets that are roughly symmetric and unimodal, but are contaminated with outliers.
Results from small sets should be treated with caution.
Misleading results may be obtained if the program is used on data sets that are multimodal or strongly skewed or almost invariant. 
The algorithm is based on the AMC paper, Analyst, 1989, 114, 1693, which also contains the test data set.

Downloadable Files

Macro code

Test data

Robust statistics: a method of coping with outliers

MS EXCEL Add-in for Robust Statistics

Written for Excel 97 and later. The add-in is available as a zipped file (compressed for fast download). All the installation instructions are in the Robust Statistics Read Me file, as well as in the full help system. 
RobStat.xla includes all the functions described in two AMC papers [1, 2]:  

• Median Absolute Deviation (MAD)
• The derived MADe estimate of standard deviation
• SMAD (which returns the mean absolute deviation if MAD=0
• the A15 estimate of the mean
• the H15 estimates of mean and standard deviation
• Robust Reproducibility calculation [2]     

The add-in includes most of the different functions in two forms:

• Menu-driven functions for Summary Robust Statistics and for robust Reproducibility/ANOVA, providing all the above via an excel menu
• Spreadsheet-callable implementations of all the above except the reproducibility/ANOVA calculation     

Zipped file
Includes all the necessary files but requires pkunzip, winzip or similar software to decompress it:

References:

1. Analytical Methods Committee, Analyst, 1989, 114, 1489. 

2. Analytical Methods Committee, Analyst, 1989, 114, 1693. 

Downloadable Files

Robust Statistics

The complete, compressed set of files for the robust stats add-in including Readme and Help

Software for calculating kernel densities

Minitab local macro

'NMODE13 for Minitab versions 10-13'
'NMODE14 for Minitab version 14'

MS EXCEL Add-in 

Written for Excel 97 and later. The add-in is available as a zipped file (compressed for fast download). All the installation instructions are in the Kernel Read Me file, as well as in the full help system. 

Zipped file 
Includes all the necessary files but requires pkunzip, winzip or similar software to decompress it.

Downloadable Files

NMODE13 for Minitab versions 10-13

NMODE14 for Minitab version 14

Test data 1

Test data 2

Kernel zip

The complete, compressed set of files for the kernel density add-in

Representing data distributions with kernel density estimates

Linear Functional Relationship Estimation by Maximum Likelihood

This Excel Add-in was developed in support of AMC technical brief number 10, published by the UUÂãÁÄÖ±²¥ of Chemistry. It uses the algorithms described in the Technical Brief and cited papers to estimate best fit linear relationships between two variables where both variables have significant uncertainties. 

The add-in provides a single menu-driven facility for maximum likelihood fitting. It includes the following features:

• Fitting with or without the intercept forced through zero
• Optional charts including regression and scaled residuals plots
• Optional input data listing
• Goodness-of fit statistics
• Full Help system   

Zipped file
Includes all the necessary files but requires pkunzip, winzip or similar software to decompress it:

1. References: 
   Analytical Methods Committee (AMC) Technical brief No. 10. "Fitting a linear functional relationship to data with error on both variables" (UUÂãÁÄÖ±²¥ of Chemistry 2002). 
2. B D Ripley, M Thompson, Regression techniques for analytical bias, Analyst 1987 112 377-383 

Downloadable Files

The complete, compressed set of files for the FREML add-in

Fitting a linear functional relationship to data with error on both variables

Minitab 14 macro for 'Goldmine'

Goldmine is a strategy game, for analytical chemists and others, that reinforces the importance of a proper consideration of the uncertainty from sampling. It demonstrates, with financial constraints, the interactions between sampling uncertainty, analytical uncertainty, and the cost effectiveness of the outcome. Although the game is based in the context of prospecting for gold, the principles illustrated apply to most applications of chemical analysis.

Downloadable Files

Analytical and sampling strategy, fitness for purpose, and computer games

An explanatory text

Minitab macro

To download Right Click on link and 'Save Target As' gold14.mac

Minitab macros for estimating normal mixture models

Further information: AMC Technical Brief No. 22

Downloadable Files

Minitab macros for estimating normal mixture models

RANOVA2

A stand-alone program, running in Microsoft ™ Excel, to execute robust and classical analysis of variance with nested data. Suitable for both balanced designs (up to 8 samples per target and 8 analyses per sample), and for unbalanced designs (2 samples but with 2 analyses on one sample only). Output is expressed as standard uncertainty, expanded relative uncertainty, and Uncertainty Factor.

Downloadable Files

RANOVA2 Software

RANOVA3

A stand-alone program, running in Microsoft ™ Excel, to execute robust and classical analysis of variance with nested data. Suitable for both balanced designs (up to 8 samples per target and 8 analyses per sample), and for unbalanced designs (2 samples but with 2 analyses on one sample only). Output is expressed as standard uncertainty, expanded relative uncertainty, and Uncertainty Factor. Version 3 includes an option to calculate confidence intervals on uncertainties for balanced designs with 2 samples & 2 analyses.

Downloadable Files

RANOVA3 Software

RANOVA4

A stand-alone program, running in Microsoft™ Excel, to execute robust and classical analysis of variance with nested data. Suitable for balanced designs with up to 8 samples per sampling target and 8 analyses per sample, and also for unbalanced designs with 2 samples, but with 2 analyses on one sample only. Output is expressed as standard uncertainty, expanded relative uncertainty, and Uncertainty Factor. Version 4.0 includes the following for balanced designs with 2 samples and 2 analyses:

1. An option to calculate confidence intervals on the uncertainty values (as in RANOVA 3);
2. An improved estimate of the measurement uncertainty (MU) which uses a coverage factor (k) calculated for the actual number of sampling targets (typically small, i.e. n < < 30). This is more accurate than the commonly used method of applying k = 2 for an approximate 95 % confidence interval, which underestimates the MU in this situation.

Downloadable Files

RANOVA4 software