Don’t trust summary statistics alone
6 May, 2017 at 08:49 | Posted in Statistics & Econometrics | 1 Comment
When teaching statistics and econometrics yours truly always stress the importance of data visualization. The thirteen datasets shown above is a great illustration of why one should always plot data. They all have the same summary statistics — mean, standard deviation, correlation coefficient — but some are dinosaurs and others are stars …
Conclusion: never trust summary statistics alone!
1 Comment
Sorry, the comment form is closed at this time.
-
Recent Comments
rsm on The ‘Just One More… Jan Milch on Keynes — en ständigt akt… rsm on Brownian motion (student … Nanikore on The total incompetence of peop… Bruce Wilder on The total incompetence of peop… rsm on Ergodicity — a questiona… Edward Fullbrook on Susan Neiman on why left is no… rsm on The non-existence of economic… fredtorssander on The non-existence of economic… Mel on Cutting-edge macroeconomics… fredtorssander on MMT — coming to an econo… Jan Milch on The Swedish for-profit ‘… rsm on The Swedish for-profit ‘… fredtorssander on What’s the use of e… rsm on What’s the use of e… Reading List
Blog at WordPress.com.
Entries and Comments feeds.
LARS P. SYLL 
Nice plots. Even worse for evolving economic time-series data, shapes can be changed depending on time-periods.
Comment by Peiya Liu— 8 May, 2017 #