Probability Basic Equations PCAIB )=DCAnB ) / PCB) Independence * Neither # PCAVBJ 1- ↳ DCA / B) =DCA ) PIBIA )=DlB ) Hinde pendent PIANBt-PIAHDCBJDCAorBJ-PCBJ-DCAJ-DCANBJPCAIB.sc )=PCA^Bnc ) / NBN ) PCAORBORC )=pCA ) -111 + Tree ex : Mutually Diagram ¥ > PIAORB) Rz ( never •¥B÷¥Bz q¥→Rz Exclusive - - PCAJTPCB) independent) - p( A) ( DLB ) ) ( 1- PCAI )( 1- PCB ) ) / PCC )) EEX.FM/--x1lxj+PlX=XJCXz)t SDCX]=F×ÉEF+ . . . . . . . . . . . Vara ]=SD[✗32 ← manipulating random variables * Bz 5-2 SDLCX ]=cSD[X ] varf.CI/J--C2VarCx] E[cX]=cE[x ] parameter ✗ NIEM ,SDHD - vtenormalcdfor normalpdf expected valve interpretation lfwe overt - over again average , E[ X-Y]=E[X ] - EEY ] E[Xt6]=E[X] -16 sD[Xt6]=SD[X ] sD[X±Y]=ÑÉD[YT - randomly select E[XtY]=E[X]tE[Y ] the Var[aXtbY ] =jÉ+ÑY]2 - the selected would be about in - E[X , -1×2-1 . . . . ]=NlE[X ] ) .tn#lSDsxDllnthelonqrun,the expected is average going -10 been - sD[X , -1kt . . . Binomial Random Variable / fixed oftr.ca/s,XisnUMberofsvccesses1mvststatevaviableincontext,ptsq exiPCX-4JBinompdfln.pt ) 1444117 Binomcdtcn ,p,7) * - # aswell Binomcdfcn ,p,z, X-D / Kp ) callbinomonlygoestole.tt# M=np e- Conditions ↳ fixed # Oftrials ↳ Independent trials ↳ only Iovtcomet success / fail : in ↳ model binomial random variable w/ normal model success failure condition - potsvccessissametoreach -1 " " - A. PHON -9=10 More Probability Geometric Random Variable lfixe.cl#otsoicesseslthetirstI,covntstrialD probability x~Glp ) M=kp that he/she first misses or makes geometcdtlpixl ✗ → number of trials sD=rq,p conditions PIK-nt.pl/-pjn-1 ↳ ✗ istriallvntillstsvices , Plant 1- Dlxen ) ↳ trials Independent ↳ only Iovtcomet success / fail __ : ↳ potsvccessissametoreach trial Central Limit Theorem 27a) conditions met • • these 801am represent 27b) asampleforthistypeofcar N-pl2.cl 0.0447 ) , normalcdf 80carswovldbelessthanlo%01-allcarsotth.is type -0.0126 wears The (3,3-1,2-910.0447) probability large enough thatyisbetween3-3.lgmlmiisabout.co emission forcarsare mutually independent 0.0126 • should be ng=My=2 -99M / Mi 05=71--0.tt#Y-mi--o.o447mq1m ; N~PH.cl , 0.0447) 27c) INVNOVM / 0.950,1 , / eft) -4.6449 z=Y¥ 1.6449=0%41-7 4=2.9736 Thereisonlyatolochanlethatthe fleet 's meancolevelisgreaterthan 2.97369mm; 1) conditions met - SRS mentioned 25finchesshovldbes-10%01-entirehovsetinchpop.nl - -25<30 * assuming lnotbigenovghsamplesizeproceedwithcavtion ) lheshapeottheorq.pop.is/-air1ysymmetrictheCLTapp1ies*Mx--- - Ng 1.5min ? pix a> 1.71 .FI#-- 0.18min F- Nlliimin , 0.18min ) =O -1333 " 7min = 1.5min Thereisao -1333 probability that the song duration will be greater than 1.7min however weshovldnitpvttoomvchfaithinthis because Ztisnotalargeenovghsampleandweare assuming that theorq.pop.is symmetric 3) . conditions met - - srhsmentionedil 36hotdoqsa-10%0f-wholepop.othotdoythatmanvfaltormaket.ir enough sampler Cltapplies -36>301 big i. If ,gg18!4g ni=18g 05-1%6=16-9 ? PIIH8-4JI-N1189.t.gs -0.0082 - Themanvfactorsclaimisnotwrongasitisshownthatthe probability forthetatcontenttobe18.4qorgreater.is 0.0082 Which is very unlikely and provides evidence thatthemeantat content's around 18g making the manufacturer 's claim correct . F) conditions met srsmentionedv-36shovldbec.IO/ootallprodvcedo.tboltsV - -36>3011941 enough samples # 0.49in o.g.no -51in Plo -49<1%0.51 ) -0.9976 1-0.9976=0.0024 - iii. 0.5in 0i=0¥=o -0033in I - N / 0.5in 0.0033in ) , The probability the process onanygivendayiso -0024 . will be shutdown confidence Interval 2) Themarginoterrormeansthatthe the true medical researcher believes with some % confidence -1ha -1 , , proportion otchildrenexposedtolead-basedpaintiswithin34ooth.is estimate lollcanconclvdethatweare 95% confident that the true Nebraska Boardofparoleis between 0.5610 and 0.6252 . proportion otdecisionsmadebythe . Bblweare 95% confident that the true proportion of auto accidents that involve teenagers is between But conditions met sprsmentionedv 0-1268 and 0.1860 -582910dm -1140%01-911 accident, ✓ ni > long > 10582194-821--91>-10 ✓ 13495%10 nfidenle means that within these range 58214945821--491>-10 Otvalves Weare 91-40 certain / twill contain the true pop proportion pimp - . - ,F¥s F- Nlp , . 0.01511 %É¥i% . Bdlltcontradiltsthis statement because the politician assaying -1ha -10.201 theallidentsinvolve - - ateenaqerbvtbasedonovrc-1-ithevalveshovldbelessthano.NO 95% -1268,018601 and thatthetrvepop proportion . 95%11=9%-82 ' - ' - " zany ) . 0.1860 Weare sure lies within 0.1268$ ' 1582+-11.91-9963986110.011-1 , (0.1268/0.1860) * Remember 'HsP-hat* Confidence Levetssamplesize affects confidence interval ↳ bigger samples ME .=Z*hF¥ ) .ee/smallermarginoferrorl ↳ half confidence interval M.EE#P-rn ) confidence level The probability thatitthetestwasrepeatedoversover again - obtained would be the , the results same Vponrepeatedsamplingvsinqthesarnesamplesiesmethod otthec.l.pro dviedwillcaptvrethetrve - _#% Hypothesis Test one proportion one tail two tail 2- test 1 proportions) TWO Proportion Making Decisions 2- Interval using test statistics TWO Proportion 2- test 12 groups independent * It iii. out two proportions sample , they are single independent * come from not Critical Value Difference between 2 proportion 2- test $2 proportion z - interval a . Type Type -1-31 # errors lfnvllhypothesisistrvebvtwerejectit-ype2-l tnvllhypothes.is wrong / → but we don't is power → The Probability reject probability otcorrectlyrejectingthetalsetlo getting Type -1# of Type → > - p Xlsamexllevelof power-up 9 sample size $9 ↳ have more into onpop.io there is abetter estimate ottrve pop .la power ) significance will ↳ Willa . significance )J tPlTypeIJ PITYPEI ) tnchanceot rejecting How to PITYPEII) - SENKSD SEM SDM SEM . sD→ -10 measure spread of measurements Interpretation onaveraqe - SEM> measure Margin , - uncertainty Willditterbysfrommean around the estimate ofthemean of error M.E.az#fFFqhorz*fn2waysolevelo1- confidence MET it - out ,nh or isao%C2-a.in 2=1.65 Clf lonovtpvttablesE-F.nl Raise Power 9×9 , samples 're ,lvsD Point Estimate , change Ha ↳ center / middle of CI conditions for Inference procedures T - Distribution 1- sample -1 - interval sometimes might ¥83Bn have to calculate mean → no extreme outliers or use NPPIONCAKI it approx normal it meets the condition . I sample ••µl - t test , SD yourself 2 sample t - test 0 0 2 sample t interval - lxi-i.t #fEFE.lIiIzl~tlx-,SElXi-Xi )) Matched Pairs -1 - test Matched Pairs -1 interval - mean differences Chi Chi - square Square Goodness Of Fit / Do - reflect -1 / consistent ) Chi Square test of Homogeneity expected row total ✗ column total total Chi square test of Residuals for Chi square - Independence Testotln dependence 2cateqor.ua/variab1eareassoiiatedW1oneanother ↳ not causation ↳ observational units are collected categorical variables are observed atrandom $2 lvnliketestofhomoqeneitywheresrsistromh ↳ stat-test-XZ-test-smutrix-edit-enterrowacolvmntorns.us groups separately) → gobackslclickcalc ← ltyoohavemvltipleqvestionsforonegrovp ↳ " association Testot ↳ " " or dependence / independence Homogeneity distribution " Residual forx ' ↳ " " same / procedure c=l0bfÉ¥ homogenous " " Linear Regression -1 Test - actual # Halve lope coefficient te stat g-intercept Rvalve tells - us about strength , directional form of the scatter plot linear Regression T interval - ltlow { - - - - Regression * Reg Have ↳ Igo to vars , stats, eqn Req) Replace , Ly ( Lz Lst Obs Exp 2nd y= , 1st plot keep 4$14 - - is the association ) Linear Plot Residuals put data to list Lin strong - ✗ w/ Li