link re: composite score based on counts of different things for 20 years
Locked
link re: composite score based on counts of different things for 20 years
 
|
I have many times written up a discussion of why it is not valid to simply average percentages (proportions) on very different bases. However, I do not have access to my notes and files at the moment. Also, the point would be emotionally much more acceptable if it came from someone other than the old curmudgeon. Rather than develop a presentation I am looking for a link that explains clearly why one cannot simply create a composite variable by averaging proportions on very different bases. It would be great if the link also suggested what to do instead. say that there are 12 different kinds of crimes and the user wants to get an overall score for each year. She wants to find the average count over the 20 years. denominator= (sum(counts(yr))/20). Then get a yearly measure of a years count over the average count. I.e.,the score for a year is a proportion of the mean count. YearScore(yr) = count(yr)/ denominator. That would be used to show the trend over the 20 years. and so forth for 12 crimes. Criminality(yr) = sum(YearScore(yr)) /12. I need a link in English. A link in Spanish and a link in Russian in addition would also help. -- Art Kendall Social Research Consultants
Art Kendall
Social Research Consultants |
|
Hi Art,
On 29 May 2013 09:00, Art Kendall <[hidden email]> wrote:
|
|
|
Hi Paul: This reference looks fascinating. I never knew it existed. Thank you. --Steve From: SPSSX(r) Discussion [mailto:[hidden email]]
On Behalf Of Paul Cook Hi Art, Is this the type of thing you are looking for? http://www.oecd.org/std/clits/42495745.pdf
On 29 May 2013 09:00, Art Kendall <[hidden email]> wrote:
-- Art Kendall Social Research Consultants Art Kendall View this message in context:
link re: composite score based on counts of different things for 20 years |
Re: link re: composite score based on counts of different things for 20 years
|
Neither did I. The maths is way over my head, but I found a nice quote to use from Endnote 1. (Mansky, Measuring Expectations, Econometrica vol. 72, 2004, 1329-1376) “Economists have long been hostile to subjective data. Caution is prudent, but hostility is not warranted. The empirical evidence . . . .shows that, by and large, persons respond informatively to questions eliciting probabilistic expectations for personally significant events. … The unattractive alternative to measurement is to make unsubstantiated assumptions” I’ve used the quote to open the Subjective Social Indicators section on my website. His name is actually Charles Manski and he’s at Northwestern (http://www.econ.northwestern.edu/people/faculty/charles-manski.html ) I found the quote apposite for two reasons. First, it supports the use of subjective social indicators, echoing the position I took in my article The Quality of Life in Urban Britain: some developments and trends (in Social Trends 7, HMSO, 1976) http://surveyresearch.weebly.com/uploads/2/9/9/8/2998485/hall_1976.pdf Second, it is a riposte to those who accuse us of “positivism” and reminds me of something I apparently said over a pint in a pub years ago to a group of my undergraduate students, It was something like, “If you can’t measure it, you can’t say anything sensible about it” and, “If you can compose a meaningful English sentence about it, you’ve already measured it.” Years later one of them sent me a copy of his book Making Sense of Social Research inscribed, “John, If you can’t measure it . . !“ Malcolm Williams is now Director of the School of Social Sciences at Cardiff. Another apposite quote I must dig out is from the very first page of Karl Marx, Das Kapital, where he says something to the effect that value accrues from the necessity of meeting human needs “whether of the belly or of fancy.” (Food for thought?) John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/spss-without-tears.html From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Salbod, Mr. Stephen Hi Paul: This reference looks fascinating. I never knew it existed. Thank you. --Steve From: SPSSX(r) Discussion [[hidden email]] On Behalf Of Paul Cook Hi Art, Is this the type of thing you are looking for? http://www.oecd.org/std/clits/42495745.pdf On 29 May 2013 09:00, Art Kendall <[hidden email]> wrote:
--Art KendallSocial Research ConsultantsArt Kendall View this message in context: link re: composite score based on counts of different things for 20 years |
Re: link re: composite score based on counts of different things for 20 years
|
|
In reply to this post by Paul Cook
Thank you that would be
an excellent reference for people like
me and many of the regulars on this
list.
I was thinking of something much simpler that shows why one needs to weights to get overall percentages. Perhaps something that shows why one can not simply average percents on different bases and that the grand mean is NOT the simple average of the group means. [and suggests z-scores] . Not only is there some difficulty with the concept, her translator needs to understand it, in order to explain it to her. When I get back home to my files and faster internet, if I cannot find a link I'll find an old write up. BTW There is another reference on indicators. It is meant to be used in evaluation (monitoring) of human rights. http://www.ohchr.org/EN/Issues/Indicators/Pages/HRIndicatorsIndex.aspx As I said to the authors at the UN event introducing it, it has a use of "qualitative" that is not what most people in the social sciences mean by qualitative. The authors consider ordinal level, nominal level, and dichotomies, as "qualitative" If one were to use it in class work one would want to cross walk the vocabulary to other vocabularies. A brief look tells me it would be a great source for people who are familiar with the different dialects of stat and methods. I hope to have time to review it thoroughly later this summer. At that time I'll try to develop notes that elaborate on the vocabulary so that readers can associate the contents with what they already are familiar with. Art Kendall Social Research ConsultantsOn 5/28/2013 6:06 PM, Paul Cook [via SPSSX Discussion] wrote:
Art Kendall
Social Research Consultants |
Re: link re: composite score based on counts of different things for 20 years
|
|
In reply to this post by John F Hall
David Aah, Kelvin of °K. No, they were my own words in answer to the intransigent anti-positivists and anti-empiricists who constantly accused me of "methodological determinism" when I was planning the first and only ever undergraduate degree in applied social research in the UK. The degree was approved in 1977 in spite of them, and was a great success despite constant attempts to undermine it. I (early) retired in 1992, but the degree closed when the last cohort graduated in 1994. Ironically, 20 years later, the ESRC is now funding an initiative on teaching quantitative methods to social science undergraduates: one proposal is for an actual undergraduate degree (See: http://www.heacademy.ac.uk/assets/documents/events/SS_assets/Blog/MacInnes_fullpaper1.pdf http://www.hefce.ac.uk/news/newsarchive/2012/name,75745,en.html http://www.quantitativemethods.ac.uk/ and http://www.ncrm.ac.uk/news/show.php?article=5290 but the link to the ESRC webpage at the end is no longer available.) Who said, "There is only Physics: all the rest is bird-watching."? It was still Kelvin, but what he actually said was actually “In science there is only physics; all the rest is stamp collecting.” as I discovered when I Googled my version and found a fascinating article with more quotes from Kelvin and dozens of others, "The Algorithm: Idiom of Modern Science" by Bernard Chazelle on (http://www.cs.princeton.edu/~chazelle/pubs/algorithm.html) John John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/spss-without-tears.html PS Chuck Manski just sent me a message: “Regarding measurement, you may perhaps be interested that Cemmap , an ESRC research center housed at UCL and IFS in London, ran a conference in 2007 with the title "Measurement Matters." Here is a link to the webpage http://www.cemmap.ac.uk/cemmap/event/id/202. The final session of the meeting was on measuring expectations, which has become well accepted by economists in the UK in recent years.” -----Original Message----- John, Re: > > Second, it is a riposte to those who accuse us of “positivism” and reminds me of something I apparently said over a pint in a pub years ago to a group of my undergraduate students, It was something like,“If you can’t measure it, you can’t say anything sensible about it” and, “If you can compose a meaningful English sentence about it, you’ve already measured it.” > Is this the quote: “When you can measure what you are speaking about, and express it in numbers, you know something about it, when you cannot express it in numbers, your knowledge is of a meager and unsatisfactory kind; it may be the beginning of knowledge, but you have scarely, in your thoughts advanced to the stage of science." - William Thomson (Lord Kelvin) David David Morganstein, V.P. Director, Westat Statistical Staff (O) 301 251 8215 (M) 240 888 5442 |
| Free forum by Nabble | Edit this page |