View Full Version : The Excel error
paraclete
Apr 19, 2013, 03:39 AM
Well, now we know, certain hypotheses are subject to error, modelling errors. Now where have we heard that before? The whole climate change debate has been subject to modelling errors but this is an error of a different kind, one in the assumptions of economists that have sent the world into a tizzy and right wing politics into apolexy. It appears certain economic truths can no more be substantiated that the "science" of climate change and I think a particular person who has made billions from selling a piece of software has a case to answer
The product should come with a a disclaimer
The Excel depression (http://www.smh.com.au/business/world-business/the-excel-depression-20130419-2i486.html)
As a user of this product over many years I can tell you I had to triple check my assumptions and even then I have made some classic blunders like inverting a formula but the consequences are dire when this is used to destroy economies
ebaines
Apr 19, 2013, 08:12 AM
I think a particular person who has made billions from selling a piece of software has a case to answer
You want to blame the makers of Excel for the mistakes made by these two professors? I don't get it. If a child uses pencil and paper to add 1 + 1 and gets 3 should we blame the maker of the pencil?
speechlesstx
Apr 19, 2013, 08:29 AM
Uh yeah, did you miss this part?
Finally, Ms Reinhart and Mr Rogoff allowed researchers at the University of Massachusetts to look at their original spreadsheet - and the mystery of the irreproducible results was solved.
First, they omitted some data; second, they used unusual and highly questionable statistical procedures; and finally, yes, they made an Excel coding error.
Garbage in, garbage out.
excon
Apr 19, 2013, 08:53 AM
Hello,
Did somebody say my name?
excon
speechlesstx
Apr 19, 2013, 09:30 AM
Batosaur?
paraclete
Apr 19, 2013, 03:11 PM
You want to blame the makers of Excel for the mistakes made by these two professors? I don't get it. If a child uses pencil and paper to add 1 + 1 and gets 3 should we blame the maker of the pencil?
No I was pointing out that the era of Excel has produced some classic clangers as far as modelling is concerned and I place the blame on the shoulders of academia who are too ready to believe the results of their computations using a tool that replaces entering the formulas for themselves and on the producers of the product who have not insisted on providing training in the use of their product. Most users of this product are self taught
I think you also need to develop a sense of humour
ebaines
Apr 22, 2013, 06:04 AM
I place the blame ... on the producers of the product who have not insisted on providing training in the use of their product. ... I think you also need to develop a sense of humour
This part didn't sound like a joke:
I think a particular person who has made billions from selling a piece of software has a case to answer.
Yes, Microsoft provides a tool that can be used in error. But it is not incumbent on MS to train people in statistical methods - and certainly economics professors teaching at the most prestigious universities in the world are already highly trained in statistics. Their coding error was apparently in selecting the wrong range of cells in the spreadsheet for a statistics operation, and there is no training in the world that would prevent that.
paraclete
Apr 22, 2013, 07:00 AM
This part didn't sound like a joke:
Yes, Microsoft provides a tool that can be used in error. But it is not incumbent on MS to train people in statistical methods - and certainly economics professors teaching at the most prestigous universities in the world are already highly trained in statistics. Their coding error was apparently in selecting the wrong range of cells in the spreadsheet for a statistics operation, and there is no training in the world that would prevent that.
Excepting training by thedeveloper of the product
ebaines
Apr 22, 2013, 07:38 AM
excepting training by thedeveloper of the product
We'll just have to disagree on that. There's no indication that the professors didn't know how to use the Excel features correctly, so more training on Excel would not have helped. Humans make mistakes, and as with any computer tool if you put garbage in you get garbage out. The professors perhaps should have realized that what they had for results was garbage, which would have driven them to double check and then they might have found their error. The fact that they didn't recognize that the results were wrong is not a problem with insufficient training on Excel.
paraclete
Apr 22, 2013, 04:01 PM
You don't know what you don't know. There is an assumption that the formulas provided in excel are what they seem to be
NeedKarma
Apr 23, 2013, 03:30 AM
Clete,
The error wasn't in the formulas included in Excel, it was in either a macro or Visual Basic code that the user can write to automate tasks.
talaniman
Apr 23, 2013, 05:28 AM
Tailoring the data for a desired outcome doesn't help the accuracy either.
paraclete
Apr 24, 2013, 05:34 PM
Look I know garbage in garbage out I spent years in IT but when you have a product there should be a responsibily to make sure the users know the limitations
NeedKarma
Apr 25, 2013, 01:08 AM
but when you have a product there should be a responsibily to make sure the users know the limitationsLike firearms manufacturers for instance? Or fast food restaurants?
paraclete
Apr 25, 2013, 03:54 AM
Exactly one size fits all doesn't work anywhere
ebaines
Apr 25, 2013, 05:48 AM
look I know garbage in garbage out I spent years in IT but when you have a product there should be a responsibily to make sure the users know the limitations
You have not told us what these so-called "limitations" are, which apparently you think the professors were snagged by. Face it - they screwed up, it was their fault, and they have acknowledged as much. There is no indication at all that some limitation in the spreadsheet led to their mistake. Ever used a calculator? If you want to know the answer to 1 + 1 and you accidentally punch in 1 + 2 do you claim that the resulting error is due to the limitations of the calculator?
excon
Apr 25, 2013, 07:29 AM
Hello:
So, are right wingers WRONG about austerity and Krugman right??
You didn't want us to be LIKE Europe, but you did everything you could to make us be LIKE Europe.
Bwa, ha ha ha ha.
excon
paraclete
Apr 25, 2013, 03:48 PM
Hi Ex you can't be like Europe, the thinking in Europe is different and in any case which Europe are you speaking of, there are three, Northern Europe, Southern Europe and Russia
talaniman
Apr 25, 2013, 03:55 PM
Which one runs the ECB?