Me a sure s o f Va ria b ility L E CT URE 4 Ob je c tive s De fine te rms. Dia g ra m re la tive diffe re nc e s in c e ntra l te nde nc y a nd va ria b ility. Ca lc ula te the ra ng e , sum o f sq ua re s, va ria nc e , a nd sta nda rd de via tio n. Ca lc ula te the dive rsity o f a sa mple o f no mina l o b se rva tio ns. 1
Va ria b ility A me a sure o f va ria b ility is a n indic a tio n o f the spre a d o f me a sure me nts a ro und the c e nte r o f the distrib utio n. Va ria b ility Me a sure s o f va ria b ility a re pa ra me te rs o f the po pula tio n. Sa mple me a sure s tha t e stima te the va ria b ility o f the po pula tio n a re sta tistic s. Ra ng e Va ria nc e Sta nda rd de via tio n 2
Me a sure s o f c e ntra lity a nd va ria b ility Me a sure s o f c e ntra lity a nd va ria b ility 3
Me a sure s o f c e ntra lity a nd va ria b ility Ra ng e T he diffe re nc e b e twe e n the hig he st a nd lo we st me a sure me nts in a g ro up o f da ta . Sta tistic a l ra ng e vs. ma the ma tic a l ra ng e I n a da ta a rra y (sma lle st to la rg e st): Sa mple ra ng e = X n -X 1 4
Ra ng e Give s so me indic a tio n o f the va ria b ility o f da ta , b ut it o nly de pe nds o n the e xtre me va lue s o f the da ta a rra y (la rg e st a nd sma lle st). I t is unlike ly tha t the sa mple will c o nta in the e xtre me va lue s o f the po pula tio n so the sa mple ra ng e will c o nsiste ntly unde re stima te the po pula tio n ra ng e (a b ia se d e stima te ). Bia se d a nd unb ia se d A b ia se d e stima to r will c o nsiste ntly unde r- o r o ve r-e stima te the va lue o f a pa ra me te r. An unb ia se d e stima to r will no t a lwa ys (o r e ve n o fte n) g ive the c o rre c t va lue o f a pa ra me te r, b ut it will o ve r-e stima te the pa ra me te r a s o fte n a s it unde re stima te s the pa ra me te r. 5
Sum o f sq ua re s Sinc e the me a n is a use ful me a sure o f c e ntra l te nde nc y, it is po ssib le to e xpre ss va ria b ility in te rms o f de via tio n fro m the me a n. Sum o f sq ua re s T he sum o f a ll de via tio ns fro m the me a n will a lwa ys e q ua l ze ro . Po sitive de via tio ns a re c a nc e lle d b y ne g a tive de via tio ns. 6
Sum o f sq ua re s Sq ua ring the de via tio ns fro m the me a n is o ne wa y o f e limina ting the sig ns fro m the de via tio ns. T he sum o f the sq ua re s o f the de via tio ns fro m the me a n is c a lle d the sum o f sq ua re s (SS). Sum o f sq ua re s Po pula tio n sum o f sq ua re s Sa mple sum o f sq ua re s 7
Va ria nc e T he me a n sum o f sq ua re s is c a lle d the va ria nc e . Sq ua re d units Po pula tio n va ria nc e Va ria nc e Sa mple va ria nc e No te diffe re nc e in the de no mina to r 8
Sta nda rd de via tio n T he sta nda rd de via tio n is the po sitive sq ua re ro o t o f the va ria nc e . Po pula tio n sta nda rd de via tio n Sta nda rd de via tio n Sa mple sta nda rd de via tio n 9
Co e ffic ie nt o f va ria tio n T he c o e ffic ie nt o f va ria tio n e xpre sse s sa mple va ria b ility re la tive to the sa mple me a n. Dive rsity F o r no mina l sc a le da ta the re is no me a n to se rve a s a re fe re nc e fo r va ria b ility. I nste a d the c o nc e pt o f dive rsity (the distrib utio n o f o b se rva tio ns a mo ng c a te g o rie s) is use d. T he inde x is use d in a re la tive fa shio n (c o mpa riso n no t a b so lute me a sure ) 10
Dive rsity T a ke s into a c c o unt b o th numb e rs o f spe c ie s a nd the e ve nne ss o f the distrib utio n a mo ng c a te g o rie s. Dive rsity n = numb e r o f individua ls in the sa mple fi = numb e r o f o b se rva tio ns in c a te g o ry i 11
E ve nne ss K = numb e r o f c a te g o rie s (spe c ie s) 12
Recommend
More recommend