mean vs median

B

bhuvaneshnick

Junior Member
Joined
Dec 18, 2014
Messages
55
i need good example to understand that median is always better than mean.Thank you in advance
 
stapel said:
Who told you that the "median is always better than [the] mean"? :shock:
Click to expand...
take 10 hospitals,out of 9 hospitals patient entry is 10 but on 10th hospital entry is 500,mean of this give 59 which does not close to any of the pattern but median gives 10 but its giving at least the detail of the 9 hospitals.so here probably median works good ,isnt it?
 
bhuvaneshnick said:
take 10 hospitals,out of 9 hospitals patient entry is 10 but on 10th hospital entry is 500,mean of this give 59 which does not close to any of the pattern but median gives 10 but its giving at least the detail of the 9 hospitals.so here probably median works good ,isnt it?
Click to expand...

Suppose you had 10 sacks of apples and wanted to divide them equally between 10 people. If 9 sacks held 10 apples and 1 sack held 20 apples what would you use, the mean or the median?
 
Ishuda said:
Suppose you had 10 sacks of apples and wanted to divide them equally between 10 people. If 9 sacks held 10 apples and 1 sack held 20 apples what would you use, the mean or the median?
Click to expand...


okay i got your intention , for this application mean is best.For some other application median may be best.so it depends on the application. is that okay?
 
bhuvaneshnick said:
okay i got your intention , for this application mean is best.For some other application median may be best.so it depends on the application. is that okay?
Click to expand...

Also depends on the # of specimens.

In general, if the # of specimen is low (<10) then median is better predictor of average behavior. If the # of specimens is more than ~30 then the difference melts away in normal distribution.
 
You must log in or register to reply here.
Top