This post is not about an issue or question but just a follow up on two earlier posts that I visited here, that is about the calculation of van der Eijk’s index of agreement on ordered rating scales with the user community contributed package agrm, presented by Alejandro Ecker at the 2010 German Stata Users Group meeting (University of Mannheim),
and here, that is about 'a measure of consensus/agreement (un-dispersion) for categorical variables, ordinal ones in particular' (citing Mike Lacey from #4 who discusses his user community contributed package ordvar, a module to calculate measures of ordinal consensus and dispersion).
agrm calculates Cees van der Eijk's alternative measure of agreement "A" as published in: Van der Eijk, C. (2001). Measuring agreement in ordered rating scales. Quality and Quantity, 35, 325-341. https://doi.org/10.1023/A:1010374114305
The package agrm can be installed using:
But the latest version is available from the Github page of the author, Alejandro Ecker, using:
Besides reading the help file, more information about using Cees van der Eijk's alternative measure of agreement is provided by Alejandro Ecker on his Github page, like some words of caution:
According to Google Scholar Van der Eijk (2001) was, as of today, cited in about 255 papers but the recent paper that triggered my exploration of this subject is: Teney, C., Pietrantuono, G., & Wolfram, T. (2024). What polarizes citizens? An explorative analysis of 817 attitudinal items from a non-random online panel in Germany. Plos one, 19(5), e0302446, https://doi.org/10.1371/journal.pone.0302446
and here, that is about 'a measure of consensus/agreement (un-dispersion) for categorical variables, ordinal ones in particular' (citing Mike Lacey from #4 who discusses his user community contributed package ordvar, a module to calculate measures of ordinal consensus and dispersion).
agrm calculates Cees van der Eijk's alternative measure of agreement "A" as published in: Van der Eijk, C. (2001). Measuring agreement in ordered rating scales. Quality and Quantity, 35, 325-341. https://doi.org/10.1023/A:1010374114305
The package agrm can be installed using:
Code:
ssc install agrm which agrm *! agrm v1.1.2 AEcker 25may2011
Code:
net install agrm, from("https://raw.githubusercontent.com/eckerale/agrm/master") replace
which agrm
*! agrm v1.1.4 AEcker 10october2011
h agrm
The measure of agreement is highly sensitve to the number of categories of the ordered rating scale. In order to mitigate the risk of miscounting the number of categories the agrm command takes advantage of all the information provided in the data set (e.g. value labels). The output window reports the number of categories used to calculate the measure of agreement. Please double-check with the actual length of the ordered rating scale. If necessary, use the categories(integer) option to manually adjust the number of categories.
Please note that the agrm command is unable to handle both negative category values and noninteger category values. If you encounter such data in your data set, use the recode command to obtain an ordered rating scale with positive and integer category values.
Please note that the agrm command is unable to handle both negative category values and noninteger category values. If you encounter such data in your data set, use the recode command to obtain an ordered rating scale with positive and integer category values.
