Given the following set of data--value picked;frequency picked;percentage of total--is there anything that can meaningfully be derived from it about the picking process?
c:\test>test.pl 10 : 39663 (3.966%) #### 20 : 41281 (4.128%) #### 30 : 43552 (4.355%) #### 40 : 46839 (4.684%) #### 50 : 50217 (5.022%) ##### 60 : 53097 (5.310%) ##### 70 : 57457 (5.746%) ##### 80 : 61963 (6.196%) ###### 90 : 68065 (6.806%) ###### 100 : 74738 (7.474%) ####### 110 : 68005 (6.801%) ###### 120 : 62216 (6.222%) ###### 130 : 57352 (5.735%) ##### 140 : 53747 (5.375%) ##### 150 : 49963 (4.996%) #### 160 : 46435 (4.644%) #### 170 : 44099 (4.410%) #### 180 : 41758 (4.176%) #### 190 : 39553 (3.955%) ####
It obviously isn't a straight random pick otherwise the distribution would be more even. It's also not quite a classic bell curve.
The process that produces this distribution is random (ie. it uses rand()), but it's more complicated than just $pick = $n[ rand @n ];.
The thing I'm trying to resolve is why is it more complicated? There is a biased towards picking the median value; but is that bias significant?
In reply to [OT]: Statistical significance? by BrowserUk
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |