Note: for continuous distributions, the histogram is. Statistic Introduction with R - Gabriel Baud-Bovy. main="",xlab="dice side",ylab="counts")

Size: px
Start display at page:

Download "Note: for continuous distributions, the histogram is. Statistic Introduction with R - Gabriel Baud-Bovy. main="",xlab="dice side",ylab="counts")"

Transcription

1 Emprcal dtrbuton 5 Dcrete probablty dtrbuton 4 3 # mulated data et for throw of a dce et.eed() x<-ample(:6,ze,replacetrue) Count # count outcome (count<-table(x)) Dce de # barplot barplot(count,la,xlab"dce de",ylab"count") ablne(h/6,lty) 3 # htogram ht(x,breakeq(.5,6.5,),la, man"",xlab"dce de",ylab"count") ablne(h/6,lty) Count Dce de ote: for dcrete dtrbuton, the htogram repreent the dtrbuton of the obervaton IIT Emprcal dtrbuton 8 Contnuou probablty dtrbuton 6 4 tern<-read.data("ternberg.dat") Count # Frequency or ount plot ht(tern$rt,xlab"rt",ylab"count", man"",la) RT # denty plot ht(tern$rt,breakeq(,4,5),freqfalse, xlab"rt",ylab"",man"",la) mtext("count",de,lne4) lne(denty(tern$rt)) ote: for contnuou dtrbuton, the htogram baed on arbtrary bn Count [R] quantle and CDF plot # quantle p<-c(.5,.5,.75) (q<-quantle(tern$rt,c(.5,.5,.75))) # cumulatve probablty plot plot(ort(tern$rt), cumum(rep(/length(tern$rt),length(tern$rt))), type"",xlab"rt",ylab"cdf",yaxt"n") ax(,ateq(,,.5),la) ablne(hp,vq,lty) RT # alternatve method plot(ecdf(tern$rt),xlab"rt",ylab"cdf",man"") IIT RT IIT Cumulatve denty functon. 4 Skewne, kurto ormal dtrbuton (top left) Example of devaton from the aumpton of normalty: modal dtrbuton (combnaton of two normal dtrbuton) Skewne a meaure of aymmetry (ymmetrc) Kurto a meaure of tal length. IIT

2 Probablty plot 5 Probablty plot (alo called rankt plot or theoretcal QQ plot) repreent the ordered data on the vertcal ax agant the correpondng normal core on the horzontal ax. For normally dtrbuted data, pont hould form a traght lne. See graph to ee how to nterpret devaton from the traght lne. IIT [R] qqnorm, qq.plot ormal quantle 6 Make a probablty plot wth the reacton tme of the Sternberg dataet. The hape of the curve ugget data reacton tme are kewed veru the left (th ndeed the cae, ee htogram n a prevou lde) qqnorm(tern$rt,man", xlab"ormal quantle", ylab"rt dtrbuton") or lbrary(car) qq.plot(tern$rt, xlab"ormal quantle", ylab"rt dtrbuton") Make the probablty plot of the log of the reacton tme. IIT RT dtrbuton ormal quantle log(rt) dtrbuton IIT IIT

3 9 Hypothe tetng IIT The ull Hypothe (H ) The frt tep n hypothe to formulate an hypothe about the varable of nteret. In general, the hypothe beng teted called the null hypothe H - counter to what we hope to demontrate. For example, the null hypothe mght be that there are no dfference between the mean µ and µ of two group,.e. H : µ µ, f you are ntereted to prove that there a dfference between the two group. The phloophcal argument ntroduced by Fher for th approach, that we can never prove omethng to be true but we can prove omethng to be fale. IIT Inferental tattc Inferental tattc: The branch of tattc concerned wth method that ue a mall et of data (ample) to make a decon (nference) about a larger et of data (populaton). Hypothe tetng:. Formulaton of the (null) hypothe. Realzaton of the experment 3. Computaton of the tattc of nteret from the ample 4. Tet of the hypothe by comparng the tattc to the theoretcal dtrbuton aumng that the null hypothe true. 5. Rejecton of the null hypothe f the tet tattcally gnfcant In hypothe tetng, the goal to ee f there uffcent tattcal evdence n the data to reject a preumed null hypothe IIT Hypothe tetng The econd tep to formulate a tattc a mathematcal ndcator - that wll allow u to tet the null hypothe. For example, the t-tattc or the F-tattc allow u to realze tet about mean. to compute the theoretcal dtrbuton aocated wth th tattc under the aumpton that the null hypothe true (e.g., the Student or t-dtrbuton for the t-tattc or the Fher dtrbuton for the F-tattc). Th theoretcal dtrbuton wll allow u to quantfy the probablty of obervng the computed value for the tattc of nteret under the aumpton that the null hypothe true. The maller th probablty, the le lkely the null hypothe. The null hypothe ad to be rejected f th probablty maller that ome arbtrary mall threhold value (typcally,.5 or.). Th value, referred to a the level of gnfcance and noted by the Greek letter α, correpond to the probablty of rejectng erroneouly the null hypothe (Type error). IIT

4 Type I and II Error 3 An hypothe tet a tattcal decon; the concluon wll ether be to reject the null hypothe n favor of the alternatve, or to fal to reject the null hypothe. The ultmate decon may be correct or may be n error. Hypothe tetng True tate H true H fale H true Correct Type error (β) Decon H fale Type error (α) Correct There are two type of error, dependng on whch of the hypothee actually true (ee Table): A type error rejectng the null hypothe when t true. The probablty of a Type I error degnated by the Greek letter alpha (α) and called the Type I error rate. A type error falng to reject the null hypothe when t fale. A Type error only an error n the ene that an opportunty to reject the null hypothe correctly wa lot. It not an error n the ene that an ncorrect concluon wa drawn nce no concluon drawn when the null hypothe not rejected. IIT Effect ze 5 A tattcally gnfcant tet gve the ndcaton that there a good probablty of obervng th effect agan f the tudy repeated. In other word, tattcal gnfcance tet ae only the relablty of the oberved effect. However, obervng a tattcally effect doe not mean necearly that th effect relevant or meanngful. Gven large enough ample, any effect (e.g., a dfference between the mean of two group), even very mall effect, can be made to be tattcally gnfcant. It therefore often uefull to gve an ndcaton of the effect. In ome cae, one mght know from the context the meanngfulne of a gnfcant effect. For example, let aume that a det followed by the mother ha the effect of decreang the weght of the babe by 5 g, th certanly a meanngful dfference but t would not be the cae f the tattcally gnfcant dfference wa of 5 g. A typcal way of meaurng the ze of an effect to gve the proporton of the total varance that accounted by the effect. Varou meaure of effect ze ext (e.g., R, η, ω ). IIT Rejectng the null hypothe 4 If the probablty of obervng the computed value of the tattc maller than α, we can reject the null hypothe wth le than α % of chance of makng an error. In th cae, we ay that the tet tattcally gnfcant at the α level. In the context of a tudy where dfferent treatment have been agned to the varou group, we alo ay the effect of the factor of nteret tattcally gnfcant (the factor of nteret the ndependent varable that manpulated by the expermenter, the effect the dfference between the mean value of the dfferent group). We wll ay more on th when we wll tudy lnear model. We can alo ay that the (mean value of dependent) varable depend on (the value of) the factor of nteret. There alway a rk of makng an error when rejectng an hypothe. However, f the tet well done, we can guaranty that th rk below ome arbtrary level. IIT Falng to reject the null hypothe 6 If the probablty of obervng the computed value of the tattc larger than α, we cannot reject the null hypothe. A null hypothe not proved becaue t not rejected. The only vald nterpretaton n th cae that there not enough evdence n the data to reject H wth le than α % of chance of makng an error. For example, the obervaton that ome data et not uffcent to how convncngly that a dfference between mean not zero doe not prove that the dfference zero: o experment can dtnguh between the cae of no dfference between mean and an extremely mall dfference between mean. A tated by Fher, we can never prove that an hypothe true: the fact that data are content wth an hypothe doe not exclude that t mght alo be content wth other other hypothee makng mlar predcton. We can only prove that an hypothe fale. The fnal concluon once the tet ha been carred out alway gven n term of the null hypothe. Even f we reject the null hypothe (H ), th doe not mean that the alternatve hypothe (H ) true. IIT

5 Alternatve Hypothe 7 The alternatve hypothe, H, a tatement of what a tattcal hypothe tet et up to etablh For example, n a clncal tral of a new drug, the alternatve hypothe mght be that the new drug ha a dfferent effect compared to that of the current drug. In that cae, the null hypothe H : µ µ and the alternatve hypothe H :µ µ. To tet th hypothe, need to ue a two-ded tet. The alternatve hypothe mght alo be that the new drug better than the current drug. In th cae, the null hypothe H :µ -µ and the alternatve hypothe H :µ >µ. To tet th hypothe, we need to ue a one-ded tet. ote that thee alternatve hypothe do not pecfy what the dfference between µ and µ. Rejectng the null hypothe rejected. IIT Power and ample ze 9 Once a prece alternatve hypothe ha been formulated, then t alo poble to compute: The probablty (β) of a type error,.e. the probablty of not rejectng H f H true (when H fale). the power of the tet (-β),.e. the probablty of rejectng H o f H true. The power of a tet ncreae wth ample ze. Once a mnmal meanngful dfference ha been fxed, t poble to found out the mnmum ample ze that wll allow a tet to reach the dered level of gnfcance and power. The ample ze mut be large enough to reveal the mnmal meanngful dfference but not too large to yeld a gnfcant reult wth a trval dfference- IIT The mnmal meanngful dfference 8 To compute the probablty of acceptng the null hypothe when t fale (Type error or β), we need to tate a more pecfc alternatve hypothe. For example, the alternatve hypothe mght be that the new drug brng a 5% mprovement of the condton. In practce, however, we often lack a theory to predct the ze of an effect, makng t dffcult to tate pecfc alternatve hypothe. One poble way to determne an alternatve hypothe would be to decded beforehand what would be the mnmal meanngful dfference δ (or effect) that relevant or nteretng for the expermenter: H : µ µ H : µ µ δ. For example, letì aume that the null hypothe that the mean and the mnmal meanng dfference, then t poble to calculate the probabolty of a Type II error β and the power of the tet - β. IIT Power (normal dtrbuton) What the probablty of rejectng the hull hypothe f the dfference between the null and alternatve hypothe δ? Power δ δ β Pr Z < z α Ψ z σ / σ / α where Ψ the cumulatve normal dtrbuton and z - α/ Ψ - (-α/). For unlateral tet, replace z - α/ wth z - α. In the prevou example, the dfference between the oberved mean and the null hypothe The probablty of rejectng the null hypothe aumng that th dfference true almot %:.6 Pr Z <.96 Pr Z 3./ 3 ( <.7 ) Exercce. What the power of the tet f there were only obervaton? Anwer:.8 (8% of chance to reject H H true). IIT / /

6 Parameter etmaton Power analy n R Sample ze plannng for the power analytc approach avalable for certan tet n R by default power.t.tet Power calculaton for one and two ample t-tet power.prop.tet Power calculaton two ample tet for proporton power.anova.tet Power calculaton for balanced one-way analy of varance tet Other pecalzed package pwr package for ome general lnear model tet aypow package for the aymptotc power of lkelhood tet ME package for behavoral and ocal cence IIT Hypothe tetng ummary 3 Expermental (Decrptve tattc) Theoretcal (Inferental tattc) Sample electon Treatment agnment Realzaton of the experment (meaure of the varable of nteret) Computaton of the tattc of nteret (ample tattc) from the meaure (e.g., mean, dfference between the two mean, or coeffcent of correlaton) Aumpton about the dtrbuton of the varable of nteret. ull hypothe: e.g. no dfference or no correlaton between the group. Computaton the theoretcal dtrbuton (or amplng dtrbuton) of the tattc of nteret under the aumpton that the null hypothe (H) true. Compare the ample tattc to the amplng dtrbuton and reject H f the probablty of obervng the ample tattc lower that ome value alpha (uually fxed at.5 or.) Addtonal ue: Effect ze, power IIT Hypothe tetng ummary Sample Group Populaton x x x 3 x 4 x tattc Dtrbuton of the parent populaton ull hypothe H Samplng dtrbuton under H Acceptaton or rejecton of H IIT Addtonal Reference 4 Cohen (994) The earth round (p<.5). Amercan Pychologt, 49():997-3 ckeron () ull Hypothe Sgnfcance Tetng: A Revew of an Old and Contnung Controvery. Pychologcal Method, 5():4-3. ruce Thompon (999) Why "Encouragng" Effect Sze Reportng I ot Workng: the Etology of Reearcher Retance to Changng Practce. Journal of Pychology, Vol. 33, 999 IIT

7 Tetng Mean 5 Parametrc tet Student t tet, one-way AOVA Multple comparon Checkng Aumpton on-parametrc tet IIT One ample tet 7 Objectve: Tetng whether the mean of a ample {x,,x } dfferent from ome theoretcal value. The varance of the parent populaton unknown. Aumpton: The ample normally dtrbuted and meaure are ndependent. ull hypothe: H : µµ > the theoretcal mean µ equal to the value µ. Stattc: The one ample t tattc m µ t / where m and are the ample mean and ample tandard devaton, and µ the theoretcal value of the mean ( x m ) Stattc theoretcal dtrbuton: If the null hypothe and our aumpton are true, then the t-tattc follow a t-dtrbuton wth - degree of freedom. IIT Tetng the Mean Standard devaton of the populaton known (> normal dtrbuton) Standard devaton of the populaton unknown (> Student or t-dtrbuton) One group One ample T tet Two group Independent ample T tet Pared ample T tet Three or more group One-way AOVA One veru two-taled t-tet Two-taled tet: H : µ µ >The alternatve hypothe that the populaton mean µ dfferent from the value µ. The lower and upper crtcal value are repectvely t α/ (-) and t -α/ (-) One-taled tet: H : µ>µ > the alternatve hypothe that the populaton mean µ larger than the value µ. The crtcal value t -α (-). One-taled tet: H : µ<µ > the alternatve hypothe that the populaton mean µ maller than the value µ. The crtcal value t α (-). 6 IIT 8 IIT

8 t.tet 9 t.tet(x, y ULL, alternatve c("two.ded", "le", "greater"), mu, pared FALSE, var.equal FALSE, conf.level.95,...) Decrpton Perform one and two ample t-tet on vector of data. Man argument x a (non-empty) numerc vector of data value.. y an optonal (non-empty) numerc vector of data value.. alternatve pecfy the alternatve hypothe: two-taled "two.ded" (default), one-taled "greater" or "le". mu a number ndcatng the true value of the mean (or dfference n mean f you are performng a two ample tet). pared a logcal ndcatng whether you want a pared t-tet var.equal a logcal varable ndcatng whether to treat the two varance a beng equal. If TRUE then the pooled varance ued to etmate the varance otherwe the Welch (or Satterthwate) approxmaton to the degree of freedom ued. conf.level confdence level of the nterval. IIT Independent-ample T Tet 3 Uage. The Independent-Sample T Tet procedure compare mean for two group of cae. Ideally, for th tet, the ubject hould be randomly agned to two group, o that any dfference n repone due to the treatment (or lack of treatment) and not to other factor. Th not the cae f you compare average ncome for male and female. A peron not randomly agned to be a male or female. In uch tuaton, you hould enure that dfference n other factor are not makng or enhancng a gnfcant dfference n mean. Dfference n average ncome may be nfluenced by factor uch a educaton and not by ex alone. Example. Patent wth hgh blood preure are randomly agned to a placebo group and a treatment group. The placebo ubject receve an nactve pll and the treatment ubject receve a new drug that expected to lower blood preure. After treatng the ubject for two month, the two-ample t tet ued to compare the average blood preure for the placebo group and the treatment group. Each patent meaured once and belong to one group. IIT One ample t-tet 3 > tern<-read.table("ternberg.dat",headertrue) > t.tet(tern$rt,mu5) One Sample t-tet data: tern$rt t , df 99, p-value <.e-6 alternatve hypothe: true mean not equal to 5 95 percent confdence nterval: ample etmate: mean of x 6.6 Manual computaton > (m<-mean(tern$rt)) [] 6.6 > (m<-d(tern$rt)/qrt(length(tern$rt))) [] > (mean(tern$rt)-5)/m [] > mc(-,)*qt(.975,99)*m Tet f the mean reacton tme d dfferent from 5 (.e., H : µ5, two-taled tet). Set µ 5. y default, the t.tet functon doe two-taled tet. value of the tattc t, number of degree of freedom of the theoretcal t-dtrbuton, and p value The P value extremely mall p-value Pr( t > 3.65) therefore the null hypothe can be rejected wth confdence. alternatve hypothe H µ 5, th mean two-taled tet. 95% confdence nterval around the mean m µ IIT Independent ample T tet 3 Objectve: Tetng the dfference between the mean of two ndependent ample {x,,x } and {x,,x } where and are the ze of the frt and econd ample. Aumpton: oth ample are normally dtrbuted and have approxmately the ame varance. ull hypothe: H : µ µ, the two mean are equal. Stattc: The two-ample t tattc t m p / where m and m are the mean of the frt and econd ample, and p the pooled etmate of the tandard devaton. ( x m ) ( x m / m ) ( ) ( ) Theoretcal tattc dtrbuton: If the null hypothe and our aumpton are true, then the t-tattc follow a t-dtrbuton wth - degree of freedom. IIT p t /

9 Welch tet 33 Welch tet a varant of Student tet for ndependent ample when the varance are not equal. In th cae, the t tattc follow a tudent dtrbuton wth ν degree of freedom: ν 4 4 ( ) ( ) where ν an approxmaton to the effectve degree of freedom gven by the Welch Satterthwate equaton. Welch, L (947) "The generalzaton of "Student'" problem when everal dfferent populaton varance are nvolved", ometrka 34 ( ): 8 35, IIT Independent ample T tet (SP) 34 Reacton tme Pared T tet 35 Uage. The Pared-Sample T Tet procedure compare the mean of two varable for a ngle group. It compute the dfference between value of the two varable for each cae and tet whether the average dffer from. Example. In a tudy on hgh blood preure, all patent are meaured at the begnnng of the tudy, gven a treatment, and meaured agan. Thu, each ubject ha two meaure, often called before and after meaure. An alternatve degn for whch th tet ued a matched-par or cae-control tudy. Here, each record n the data fle contan the repone for the patent and alo for h or her matched control ubject. In a blood preure tudy, patent and control mght be matched by age (a 75-year-old patent wth a 75-year-old control group member). IIT > boxplot(tern$rt[tern$tet], tern$rt[tern$tet], ylab"reacton tme",xlab"tet") > # ame but ung a formula > boxplot(rt~tet,tern,ylab"reacton tme",xlab"tet") The boxplot repreent the dtrbuton of the reacton tme when the tet number wa preent (TEST) or mng (TEST) among the prevouly preented et of number. The queton whether the dfference between the two mean tattcally gnfcant. > t.tet(tern$rt[tern$tet],tern$rt[tern$tet], var.equaltrue) Two Sample t-tet data: tern$rt[tern$tet ] and tern$rt[tern$tet ] t , df 98, p-value.3 alternatve hypothe: true dfference n mean not equal to 95 percent confdence nterval: ample etmate: mean of x mean of y m 49(3.79 /5 m 98 / Tet y default, t. tet aume that varance are not equal and ue Welch (or Satterthwate) approxmaton. Settng var.equaltrue wll force - degree of freedom. The null hypothe can be rejected wth confdence becaue the P value (Type error) extremely mall (P, two-taled T tet). IIT Pared T tet 36 Objectve: Tetng the dfference between the mean of two varable meaured on the ame expermental unt {(x, x ),, (x, x )} where the number of expermental unt. Aumpton: The dfference d x -x are ndependent and normally dtrbuted. ull hypothe: H : µ µ, the two mean are equal. ote that th equvalent to ay that the mean m d of the dfference d zero(h : md ) Stattc: Let compute the dfference d x -x for each par of ample (,,). The tattc m d t / where m d and d are the mean and tandard devaton of d repectvely. Theoretcal tattc dtrbuton: If the null hypothe and our aumpton are true, then the t-tattc follow a t-dtrbuton wth - degree of freedom. ote that the procedure for pared t-tet correpond to a one-ample t tet for the dfference d. d IIT t p d p d )

10 Ozone data et The ozone data et gve the maxmum ozone concentraton n Stamford and Yonker for each day of a fve month perod. Mng data are coded wth the value Stamford Yonker Day ozone<-read.table("ozone.dat",headertrue) # read data ozone[ozone-999]<-a # replace mng data wth A ozone$month<-ordered(ozone$month, # order factor month levelc("may","june","july","augut","september")) > table(tmf.na(ozone$tmf), # count mng data ykr.na(ozone$ykr)) ykr tmf FALSE TRUE FALSE 3 4 TRUE 6 x<-ozone[,c("tmf","ykr")] # plot tme ere matplot(row(x),x,type"l",xaxt"n",xlab"day",ylab"ozone", la,lty,colc("red","blue")) ax(,atwhch(ozone$day),labpate(ozone$month,ozone$day)[ozone$day]) legend("topleft",c("stamford","yonker"),lty,lwd,colc("red","blue"),bty"n") 37 IIT Ozone May June July Augut September Pared t-tet 38 5 Plot Stamford agan Yonker concentraton 5 > plot(ozone$tmf,ozone$ykr,la,ap xlab"stamford",ylab"yonker") 5 5 Stamford > t.tet(ozone$tmf,ozone$ykr, paredtrue) Pared t-tet data: ozone$tmf and ozone$ykr t 3.44, df 3, p-value <.e-6 alternatve hypothe: true dfference n mean not equal to 95 percent confdence nterval: ample etmate: mean of the dfference Cae (day) where there a mng value for one of the two cte are excluded from th analy. The number of degree of freedom (3) correpond to number of day where the ozone concentraton ha been meaured for both cte mnu. The tet hghly gnfcant (P, two-taled T tet), ndcatng that the mean m d not equal to zero and, therefore, that the two mean are dfferent. IIT Yonker 39 IIT Obervaton n Stamford and Yonker are correlated (whch hould be expected nce they were taken on the ame day). 4 IIT

11 Aanaly of varance (AOVA) One-way AOVA Effect ze Multple Comparon Checkng Aumpton Tranformaton on-parametrc tet One-way AOVA The am of the one-way AOVA to tet whether the mean of two or more group are equal or not. In the one-way AOVA, there a unque ndependent varable or experment factor (A) that manpulated by the expermenter. The dependent varable (Y) meaured n dfferent expermental condton whch correpond to dfferent value or level of the expermental factor. In other word, the value of the dependent varable (or the level of the expermental factor) determne the group to whch the obervaton belong. The expermental unt mut have been agned randomly to each treatment (fully randomzed degn). 4 IIT 43 IIT Expermental degn and the AOVA The Analy of Varance (AOVA) one of the mot common tattcal technque n pychology and medcal cence The AOVA can be ued to analyze the effect of manpulaton (or expermental factor) on the mean value of the obervaton n a large varety of tuaton (or expermental degn). Termnology: Completely randomzed, block randomzed or repeated-meaure degn refer to the way treatment have been agned to the expermental unt One-way, two-way, n-way AOVA refer to the number of expermental factor. Factoral AOVA refer to a partcular way of analyzng expermental degn wth two or more experment factor whch allow one to ae the poble nteracton effect of nteracton. One-way AOVA Objectve: Compare the mean of g ample {y,,y n },, {y g,,y gng } where the ample ze of the th ample (n n g ) Aumpton: Data are ndependent and normally dtrbuted. The varance of each group approxmately equal. ull hypothe: H : µ µ g, the mean of all group are equal. Stattc: The F-tattc the rato F E / / g g MS MS E where the treatment (or between-group) um of quare, E the error (or wthn-group) um of quare. Theoretcal tattc dtrbuton: If the null hypothe and our aumpton are true, then the F-tattc follow a Fher dtrbuton wth g- and -g degree of freedom. 4 IIT 44 IIT

12 Sum of quare 45 y Uncorrected um of quare T ( y y ) Y Total um of quare (alo called corrected um of quare for the y ) R ( y ˆ y ) Explaned um of quare R ( y y ˆ ) ε E Redual um of quare (or um of quare of the error, or unexplaned um of quare) X ( x x) Corrected um of quare for the x SXY ( x x )( y y ) Corrected um of quare of the product IIT Why the AOVA work? 46 It can be proved that the total um of quare (T) equal to the treatment um of quare plu the error um of quare T 3 5 g n ( y m ) j E MS an etmate of the varance of the mean: MS g n m g g ( m ) where m the mean of the th ample and m the general (or grand) mean. ote that MST f H true. The mean quare error MSE an etmate wthn-group varance MS E E g ote that we have aumed that the varance wa the ame for all group. F become larger f the dfference between the mean are mportant and, therefore, the varance between group larger. F become maller f there much uncertanty about the mean (that, a larger wthn group varance) g g n n ( y m ) j IIT j j Reacton tme (/) Example 47 Tet f the mean reacton tme (rt) depend on the number of dgt memorzed. > boxplot(rt~ndgt,tern,outlerfalse, ylab"reacton tme (/)", xlab"umber of dgt") 3 5 > ft<-aov(rt~ndgt,tern) > anova(ft) Analy of Varance Table The AOVA table ndcate that the null hypothe can be rejected wth hgh confdence (F(,97)33.4,P<.). Repone: rt Df Sum Sq Mean Sq F value Pr(>F) ndgt e-4 *** Redual Sgnf. code: ***. **. *.5.. F MS MS E IIT umber of dgt Reacton tme (/) umber of dgt Multple comparon 48 The F-tet n the AOVA tell u that there at leat one group that ha a mean that tattcally dfferent from the mean of the other group. Problem: How can we dentfy whch group have mean that are tattcally dfferent one from another? One approach would be to ue the t-tet to make everal two-by-two (or parwe) comparon. The problem that the probablty of makng a fale dcovery (.e., the rk of fndng a tattcally gnfcant dfference between two group when the null hypothe true) ncreae wth the number of tet. ote that the problem extremely general and occur not only when comparng mean but when one make multple tet. When makng multple tet to fnd out ome effect (data noopng), we alway ncreae the rk of makng a fale dcovery. IIT

13 onferron nequalty 49 Probablty of makng a fale dcovery n a ere of n tet n 3 5 onferron lmt Independent tet onferron nequalty: There a theorem that ay that the probablty α of makng a fale dcovery n a ere of n tet maller than the um of the probablte α of makng a Type I error n the ndvdual tet. α α L α n If all tet are ndependent, then we can compute the probablty of makng a fale dcovery n a ere of n tet: α ( α where α the rk of a Type I error n the ndvdual tet ) n IIT A pror or planned comparon 5 In general, th tuaton nvolve a lmted number of tet. In the cae of comparon between mean, the uual approach, baed on onferron nequalty, to ue t-tet wth an alpha level α for the ndvdual tet equal to the dered alpha level α of protecton dvded by the number of tet α α n where α the adjuted alpha level for th t tet. ote th th method extremely general and could be ued to adjut the alpha level each tme we make multple tet. A partcular type of hypothe that one mght want to tet, for example, whether the dfference between the mean of a ubet of group dfferent from the mean from another ubet of group. In that cae, one mght ue contrat (we wll not ee contrat n th cla). Another poblty to ue a t tet between the two ubet (remember to adjut the alpha accordng to the number of tet performed). IIT Comparng mean 5 There are everal technque to adjut the level of confdence of the ndvdual tet to control for the rk of a fale dcovery when comparng mean between dfferent group. We need however to dtnguh between the followng tuaton: A pror or planned-comparon: One ha dentfed n advance whch par of mean (or combnaton of mean) hould be teted. Pot-hoc comparon: In abence of pecfc hypothe, one want to make all parwe comparon between the mean of everal treatment. ote that the number of comparon (tet) n pot-hoc comparon uually much hgher than n planned-comparon. IIT Parwe comparon 5 A parwe comparon tet examne the dfference of every par of mean. ote that there are g(g-)/ par of mean for g group. Parwe comparon tet: Sheffe method onferron Sgnfcant Dfference (SD) Tukey Honet Sgnfcant Dfference (HSD) Ryan-Eno-Gabrel-Welh-Range (REGWR) Student-ewman-Keul (SK) Leat Sgnfcant Dfference (LSD) The mot ued tet the HSD tet. It more powerful than Sheffe method and SD when all par are condered. The other tet mght be more powerful but provde le protecton agant Type I error (ee defnton of error rate n Stattcal book). IIT

14 Tukey HSD 53 The functon TukeyHSD can be ued to compute Tukey Honet Sgnfcant Dfference > ft<-aov(rt~ndgt,tern) > TukeyHSD(ft, whch"ndgt") Tukey multple comparon of mean 95% famly-we confdence level Ft: aov(formula rt ~ ndgt, data tern) $ndgt dff lwr upr p adj The multcomp package offer a convenent nterface to perform multple comparon n a general context. Reference: retz F, Hothorn T, Wetfall P () Multple Comparon Ung R. Routledge. IIT Cohen d and f 55 One-ample t tet d µ µ m σ µ Two-ample t-tet µ d m µ σ m one-way AOVA σ f σ ( µ ) p µ p σ p MS MSE MSE p value of f. a mall effect ze, f.5 a medum effect ze, and f.4 or larger a large effect ze. IIT Effect ze 54 A tattcally gnfcant tet doe mean that the dfference acro goup are large, relevant or meanngful, n partcular f the ample ze large (ee Hypothe tetng). Several meaure of the trenght of aocaton of effect ze ext: Cohen d and f R, η, ω The computaton of the effect ze depend on the expermental degn. We conder only the mplet cae. IIT Effect ze (η ) 56 The mot common meaure of effect ze expre treatment magntude a a proporton of the total varablty that aocated wth the effect The mplet meaure of effect ze η (alo called R ): η R T For example, n the prevou example, the tmulu length explan 8% of the varance of the oberved reacton tme (η 957./ ). ote that t poble to compute confdence nterval around effect ze (Smthon, 3) but we won t cover t here (ee Tabachnck & Fddell, 7, for an ntroductory tattcal textbood that cover t) IIT

15 Effect ze (ω ) 57 The mplet meaure of effect η a decrptve tattc baed on the proporton of varance accoundted n a partcular ample and tend to overetmate the ze of the effect n the populaton. An alternatve meaure, called ω, an unbaed parameter etmate that generalze to the populaton ˆ ω T df MS MS E df df ( F ( F ) ) where the total number of ubject (both formulae are equvalent). ecaue t adjut the overtmaton of effect ze by η, t alway maller than η. In the prevou example, ˆ ω (33.4 (33.4 ) ) 3.77 IIT The aumpton of the AOVA 59 Independence of the obervaton. Techncally, th aumpton tate the the redual ε j are ndependent: cov(ε j, ε kl ). For any two obervaton wthn an expermental treatment, we aume that knowng how one of thee obervaton tand relatve to the treatment mean tell u nothng about the other obervaton. Th one of the reaon why ubject are randomly agned to group. Volaton of th aumpton can have erou conequence for an analy. Homogenety of the varance. The varance nde each group are equal (Var(ε j ) σ ). ormalty. Obervaton are dtrbuted normally around ther mean: y j ~(µ,σ) or, equvalently, ε j ~(,σ). The number of treatment (group) fxed. It n prncple poble to conder tuaton where the number of group condered n the analy conttute only a ubet of all poble group/treatment. In the followng, we hall however aume that all poble group/treatment have been condered. IIT Effect ze The object returned by anova a data.frame. The element of th table can be elected and manpulated. > x<-anova(aov(rt~ndgt,tern)) >.data.frame(x) [] TRUE Compute Cohen f > df<-x["ndgt","df"] > dfe<-x["redual","df"] > MS<-x["ndgt","Sum Sq"] > MSE<-x["Redual","Mean Sq"] > qrt(df/dfe*(ms-mse)/mse) [] Compute eta > x["ndgt","sum Sq"]/um(x[,"Sum Sq"]) [] Compute omega > df<-x["ndgt","df"] > F<-x["ndgt","F value"] # F rato > <-um(x[,"df"]) > df*(f-)/(df*(f-)) [].7693 Checkng the aumpton The qualty of our nference depend on the valdty of our aumpton. The man aumpton are:. data/error are ndependent. data/error are normally dtrbuted 3. data/error have contant varance (acro group) Independence the mot mportant of thee aumpton, and alo the mot dffcult to accommodate when t fal. ormalty the leat mportant aumpton (for t-tet or AOVA), partcularly for large ample ze. alanced degn (ame number of data n each group) are le uceptble to the effect of non-normalty and non-contant varance. There are two man way of dealng wth volaton of the aumpton: Tranformng the data Ung non parametrc tet 58 IIT 6 IIT

16 Independence 6 Independence the mot mportant of thee aumpton, and alo the mot dffcult to accommodate when t fal. Example of dependence eral dependence tme ere, learnng or fatgue effect drft n meaurng ntrument patal aocaton A graphcal method to dentfy eral dependence to plot the data (or the redual) veru ther tme or the order of acquton To deal wth poble order (learnng or fatgue) effect, t good practce to randomze the order of preentaton of tmul and vary th order acro partcpant. IIT Example. Seral dependence 6 Ozone data: plot maxmum ozone concentraton veru the day for Stamford. To plot all cae, Select Graph/Lne Select Value of ndvdual cae and Smple Select Lne Repreent (tmf) and Cae number n Category Labek 3 Auto correlaton plot: Select Graph/Tme Sere/ Autocorrelaton Select Varable (tmf) There ome evdence that conecutve data are potvely correlated at lag (bar above the horzontal lne) Cae umber Syntax IIT 9 Value STMF on-contant varance 63 Several tet ext to tet equalty of varance acro ample (homocedatcty): artlett tet Levene tet rown-forythe tet artlett' tet entve to departure from normalty. That, f the ample come from non-normal dtrbuton, then artlett' tet may mply be tetng for non-normalty. The Levene tet and rown Forythe tet are alternatve to the artlett tet that are le entve to departure from normalty. The mot common devaton from contant varance are thoe were the redual varaton depend on the mean. You can plot the tandard devaton agant the mean for each group to check f there a problem. To tablze the varance when t ncreae wth the mean, you can tranform the varable wth the quare root or logarthm functon. IIT Lag umber GRAPH /LIE(SIMPLE)VALUE( tmf ). ACF VARIALES tmf /OLOG /MXAUTO 6 /SERRORID. levene.tet (lbrary car) 64 levene.tet(y, data,...) levene.tet(y, group,...) Decrpton Tet for homeogenety of varance acro group. Man argument y repone varable for the default method, or formula object. If y a formula, the varable on the rght-hand-de of the model mut all be factor and mut be completely croed. group factor defnng group. data data frame contanng dependent and factor referred to n the formula. ote Th functon part of the lbrary car (J. Fox). IIT

17 Aeng non-normalty ormalty the leat mportant aumpton (for t-tet or AOVA), partcularly for large ample ze. Graphcal method Htogram boxplot (ueful for dentfyng quckly outler). probablty blot (alo named theoretcal quantle-quantle plot or rankt plot) Formal method Ch quare tet Kolmogorov-Smrnov tet Outler are an extreme form of non-normalty and mght have an mportant effect. An outler an obervaton dfferent from the bulk of the data. Revewng all the method to deal wth outler outde the cope of th coure. If ome data unequvocally an outler that need to be deal wth, t poble to replace the outler wth the mean value nde the group to remove the outler from the analy [R] tetng aumpton Tet normalty > k.tet(tern$rt,y"pnorm",meanmean(tern$rt),dd(tern$rt)) One-ample Kolmogorov-Smrnov tet data: tern$rt D.8, p-value.3465 alternatve hypothe: two-ded Warnng meage: In k.tet(tern$rt,y"pnorm", meanmean(tern$rt),d(tern$rt)): cannot compute correct p-value wth te The preence of te generate a warnng, nce contnuou dtrbuton do not generate them. Tet equalty of varance > lbrary(car) > levene.tet(tern$rt, a.factor(tern$ndgt)) Levene' Tet for Homogenety of Varance Df F value Pr(>F) group IIT 67 IIT Kolmogorov-Smrnov tet k.tet 66 k.tet(x, y,..., alternatve c("two.ded", "le", "greater"), exact ULL) Decrpton Perform one or two ample Kolmogorov-Smrnov tet. Man argument x a (non-empty) numerc vector of data value.. y ether a numerc vector of data value, or a character trng namng a cumulatve dtrbuton functon or an actual cumulatve dtrbtuton functon uch a pnorm alternatve pecfy the alternatve hypothe: two-taled "two.ded" (default), one-taled "greater" or "le". Detal If y numerc, a two-ample tet of the null hypothe that x and y were drawn from the ame contnuou dtrbuton performed. Alternatvely, y can be a character trng namng a contnuou (cumulatve) dtrbuton functon, or uch a functon. In th cae, a one-ample tet carred out of the null that the dtrbuton functon whch generated x dtrbuton y wth parameter pecfed by... IIT Tranformaton 68 Score mght be tranformed to better for the followng reaon:. to acheve homogenety of error varance. to acheve normalty of error effect 3. to obtane addtvty effect Poble tranformaton nclude The quare root tranformaton: y qrt(y) the logarthmc tranformaton: y log(y) or y log(y) the recprocal tranformaton: y /x or y /(y) etc. Gven the relatve robutne of the AOVA to devaton from the normalty or homocadetcty aumpton, t not clear whether tranformng thendata derable. Stll, t conventonal to tranfrom ome type of data. IIT

18 Data wth non-normal dtrbuton 69 Data Tranformaton Some data do not have a normal dtrbuton and a varance that depend on the mean. For example, Proporton (nomal) Coeffcent of correlaton arcn log ( p ) r r Proporton Coeffcent of correlaton Poon dtrbuted data (e.g. count, nterpke nterval) Count (Poon) y To perform a t-tet or an AOVA on data of th type, t neceary tranform the varable wth the functon ndcated n the table. For proporton and count, t alo poble to ue generalzed lnear model (logtc regreon, log-lnear model, etc.). IIT Example - Ren Data Set 7 The Ren Data Set: Th data et repreent the reult of a tet of the lfetme (n hour) of an encapulatng ren ued n the contructon of ntegrated chp. Thrty even unt were agn at random to one of fve dfferent temperature tre (temp varable, n Celu). The tme varable gve the tme (n hour) and the tme untl the unt faled. Count Tme (hour) Htogram (and probablty plot) how that data are not normally dtrbuted and boxplot how that varance vary wth the mean. The ame plot wth the log of the lfetme how that the data atfy better the aumpton of the AOVA (next lde). The AOVA on log of the lfetme how that the null hypothe can be rejected wth confdence. Tme (hour) Ren Data Set Tme log(hour) IIT AOVA LOGTIME Temperature (Celu) etween Group Wthn Group Total Sum of Square df Mean Square F Sg E IIT Count Tme log(hour) Count 5 5 Tme (hour) Vt duraton dtrbuton Tme (hour) ormal quantle Temperature (Celu) log(vt duraton dtrbuton) Temperature (Celu) - - ormal quantle on-parametrc tet 7 on-parametrc tet can be ued when aumpton of parametrc tet are not atfed. Parametrc on-parametrc* ndependent ample dependent ample ndependent ample T tet Pared T tet One-way AOVA Mann-Whtney U tet Wlconxon tet Sgn tet Krukal-Wall tet Medan tet on-parametrc tet are le powerful than parametrc tet. IIT

19 wlcox.tet (R) 73 wlcox.tet(x, y ULL, alternatve c("two.ded", "le", "greater"), mu, pared FALSE, exact ULL, correct TRUE, conf.nt FALSE, conf.level.95,...) Decrpton Perform one and two ample Wlcoxon tet on vector of data; the latter alo known a Mann-Whtney tet.. Man argument y an optonal (non-empty) numerc vector of data value.. alternatve pecfy the alternatve hypothe: two-taled "two.ded" (default), one-taled "greater" or "le". mu a number ndcatng the true value of the mean (or dfference n mean f you are performng a two ample tet). pared a logcal ndcatng whether you want a pared t-tet data frame contanng dependent and factor referred to n the formula. exact a logcal ndcatng whether an exact p-value hould be computed. correct a logcal ndcatng whether to apply contnuty correcton n the normal approxmaton for the p-value. IIT [R] on parametrc tet 75 > wlcox.tet(tern$rt[tern$tet],tern$rt[tern$tet]) Wlcoxon rank um tet wth contnuty correcton data: tern$rt[tern$tet ] and tern$rt[tern$tet ] W 898, p-value 4.83e-5 alternatve hypothe: true locaton hft not equal to > krukal.tet(rt~ndgt,tern) Krukal-Wall rank um tet data: rt by ndgt Krukal-Wall ch-quared , df, p-value <.e-6 IIT krukal.tet (R) 74 krukal.tet(x, g,...) krukal.tet(formula, data, ubet, na.acton,...) Decrpton Perform a Krukal-Wall rank um tet... Man argument x a numerc vector of data value, or a lt of numerc data vector. g a vector or factor object gvng the group for the correpondng element of x. Ignored f x a lt. formula a formula of the form lh ~ rh where lh gve the data value and rh the correpondng group. data an optonal matrx or data frame contanng the varable n the formula. IIT

Specification -- Assumptions of the Simple Classical Linear Regression Model (CLRM) 1. Introduction

Specification -- Assumptions of the Simple Classical Linear Regression Model (CLRM) 1. Introduction ECONOMICS 35* -- NOTE ECON 35* -- NOTE Specfcaton -- Aumpton of the Smple Clacal Lnear Regreon Model (CLRM). Introducton CLRM tand for the Clacal Lnear Regreon Model. The CLRM alo known a the tandard lnear

More information

Additional File 1 - Detailed explanation of the expression level CPD

Additional File 1 - Detailed explanation of the expression level CPD Addtonal Fle - Detaled explanaton of the expreon level CPD A mentoned n the man text, the man CPD for the uterng model cont of two ndvdual factor: P( level gen P( level gen P ( level gen 2 (.).. CPD factor

More information

2.3 Least-Square regressions

2.3 Least-Square regressions .3 Leat-Square regreon Eample.10 How do chldren grow? The pattern of growth vare from chld to chld, o we can bet undertandng the general pattern b followng the average heght of a number of chldren. Here

More information

AP Statistics Ch 3 Examining Relationships

AP Statistics Ch 3 Examining Relationships Introducton To tud relatonhp between varable, we mut meaure the varable on the ame group of ndvdual. If we thnk a varable ma eplan or even caue change n another varable, then the eplanator varable and

More information

Statistical Properties of the OLS Coefficient Estimators. 1. Introduction

Statistical Properties of the OLS Coefficient Estimators. 1. Introduction ECOOMICS 35* -- OTE 4 ECO 35* -- OTE 4 Stattcal Properte of the OLS Coeffcent Etmator Introducton We derved n ote the OLS (Ordnary Leat Square etmator ˆβ j (j, of the regreon coeffcent βj (j, n the mple

More information

Estimation of Finite Population Total under PPS Sampling in Presence of Extra Auxiliary Information

Estimation of Finite Population Total under PPS Sampling in Presence of Extra Auxiliary Information Internatonal Journal of Stattc and Analy. ISSN 2248-9959 Volume 6, Number 1 (2016), pp. 9-16 Reearch Inda Publcaton http://www.rpublcaton.com Etmaton of Fnte Populaton Total under PPS Samplng n Preence

More information

MULTIPLE REGRESSION ANALYSIS For the Case of Two Regressors

MULTIPLE REGRESSION ANALYSIS For the Case of Two Regressors MULTIPLE REGRESSION ANALYSIS For the Cae of Two Regreor In the followng note, leat-quare etmaton developed for multple regreon problem wth two eplanator varable, here called regreor (uch a n the Fat Food

More information

Team. Outline. Statistics and Art: Sampling, Response Error, Mixed Models, Missing Data, and Inference

Team. Outline. Statistics and Art: Sampling, Response Error, Mixed Models, Missing Data, and Inference Team Stattc and Art: Samplng, Repone Error, Mxed Model, Mng Data, and nference Ed Stanek Unverty of Maachuett- Amhert, USA 9/5/8 9/5/8 Outlne. Example: Doe-repone Model n Toxcology. ow to Predct Realzed

More information

Negative Binomial Regression

Negative Binomial Regression STATGRAPHICS Rev. 9/16/2013 Negatve Bnomal Regresson Summary... 1 Data Input... 3 Statstcal Model... 3 Analyss Summary... 4 Analyss Optons... 7 Plot of Ftted Model... 8 Observed Versus Predcted... 10 Predctons...

More information

Statistics for Economics & Business

Statistics for Economics & Business Statstcs for Economcs & Busness Smple Lnear Regresson Learnng Objectves In ths chapter, you learn: How to use regresson analyss to predct the value of a dependent varable based on an ndependent varable

More information

/ n ) are compared. The logic is: if the two

/ n ) are compared. The logic is: if the two STAT C141, Sprng 2005 Lecture 13 Two sample tests One sample tests: examples of goodness of ft tests, where we are testng whether our data supports predctons. Two sample tests: called as tests of ndependence

More information

ANOVA. The Observations y ij

ANOVA. The Observations y ij ANOVA Stands for ANalyss Of VArance But t s a test of dfferences n means The dea: The Observatons y j Treatment group = 1 = 2 = k y 11 y 21 y k,1 y 12 y 22 y k,2 y 1, n1 y 2, n2 y k, nk means: m 1 m 2

More information

Verification of Selected Precision Parameters of the Trimble S8 DR Plus Robotic Total Station

Verification of Selected Precision Parameters of the Trimble S8 DR Plus Robotic Total Station 81 Verfcaton of Selected Precon Parameter of the Trmble S8 DR Plu Robotc Total Staton Sokol, Š., Bajtala, M. and Ježko, J. Slovak Unverty of Technology, Faculty of Cvl Engneerng, Radlnkého 11, 81368 Bratlava,

More information

Chapter 11. Supplemental Text Material. The method of steepest ascent can be derived as follows. Suppose that we have fit a firstorder

Chapter 11. Supplemental Text Material. The method of steepest ascent can be derived as follows. Suppose that we have fit a firstorder S-. The Method of Steepet cent Chapter. Supplemental Text Materal The method of teepet acent can be derved a follow. Suppoe that we have ft a frtorder model y = β + β x and we wh to ue th model to determne

More information

Comparison of Regression Lines

Comparison of Regression Lines STATGRAPHICS Rev. 9/13/2013 Comparson of Regresson Lnes Summary... 1 Data Input... 3 Analyss Summary... 4 Plot of Ftted Model... 6 Condtonal Sums of Squares... 6 Analyss Optons... 7 Forecasts... 8 Confdence

More information

Statistics Chapter 4

Statistics Chapter 4 Statstcs Chapter 4 "There are three knds of les: les, damned les, and statstcs." Benjamn Dsrael, 1895 (Brtsh statesman) Gaussan Dstrbuton, 4-1 If a measurement s repeated many tmes a statstcal treatment

More information

Scattering of two identical particles in the center-of. of-mass frame. (b)

Scattering of two identical particles in the center-of. of-mass frame. (b) Lecture # November 5 Scatterng of two dentcal partcle Relatvtc Quantum Mechanc: The Klen-Gordon equaton Interpretaton of the Klen-Gordon equaton The Drac equaton Drac repreentaton for the matrce α and

More information

Statistics for Managers Using Microsoft Excel/SPSS Chapter 13 The Simple Linear Regression Model and Correlation

Statistics for Managers Using Microsoft Excel/SPSS Chapter 13 The Simple Linear Regression Model and Correlation Statstcs for Managers Usng Mcrosoft Excel/SPSS Chapter 13 The Smple Lnear Regresson Model and Correlaton 1999 Prentce-Hall, Inc. Chap. 13-1 Chapter Topcs Types of Regresson Models Determnng the Smple Lnear

More information

Statistics II Final Exam 26/6/18

Statistics II Final Exam 26/6/18 Statstcs II Fnal Exam 26/6/18 Academc Year 2017/18 Solutons Exam duraton: 2 h 30 mn 1. (3 ponts) A town hall s conductng a study to determne the amount of leftover food produced by the restaurants n the

More information

Confidence intervals for the difference and the ratio of Lognormal means with bounded parameters

Confidence intervals for the difference and the ratio of Lognormal means with bounded parameters Songklanakarn J. Sc. Technol. 37 () 3-40 Mar.-Apr. 05 http://www.jt.pu.ac.th Orgnal Artcle Confdence nterval for the dfference and the rato of Lognormal mean wth bounded parameter Sa-aat Nwtpong* Department

More information

Department of Quantitative Methods & Information Systems. Time Series and Their Components QMIS 320. Chapter 6

Department of Quantitative Methods & Information Systems. Time Series and Their Components QMIS 320. Chapter 6 Department of Quanttatve Methods & Informaton Systems Tme Seres and Ther Components QMIS 30 Chapter 6 Fall 00 Dr. Mohammad Zanal These sldes were modfed from ther orgnal source for educatonal purpose only.

More information

Chapter 11: Simple Linear Regression and Correlation

Chapter 11: Simple Linear Regression and Correlation Chapter 11: Smple Lnear Regresson and Correlaton 11-1 Emprcal Models 11-2 Smple Lnear Regresson 11-3 Propertes of the Least Squares Estmators 11-4 Hypothess Test n Smple Lnear Regresson 11-4.1 Use of t-tests

More information

4 Analysis of Variance (ANOVA) 5 ANOVA. 5.1 Introduction. 5.2 Fixed Effects ANOVA

4 Analysis of Variance (ANOVA) 5 ANOVA. 5.1 Introduction. 5.2 Fixed Effects ANOVA 4 Analyss of Varance (ANOVA) 5 ANOVA 51 Introducton ANOVA ANOVA s a way to estmate and test the means of multple populatons We wll start wth one-way ANOVA If the populatons ncluded n the study are selected

More information

ANSWERS CHAPTER 9. TIO 9.2: If the values are the same, the difference is 0, therefore the null hypothesis cannot be rejected.

ANSWERS CHAPTER 9. TIO 9.2: If the values are the same, the difference is 0, therefore the null hypothesis cannot be rejected. ANSWERS CHAPTER 9 THINK IT OVER thnk t over TIO 9.: χ 2 k = ( f e ) = 0 e Breakng the equaton down: the test statstc for the ch-squared dstrbuton s equal to the sum over all categores of the expected frequency

More information

Chapter 13: Multiple Regression

Chapter 13: Multiple Regression Chapter 13: Multple Regresson 13.1 Developng the multple-regresson Model The general model can be descrbed as: It smplfes for two ndependent varables: The sample ft parameter b 0, b 1, and b are used to

More information

Basic Business Statistics, 10/e

Basic Business Statistics, 10/e Chapter 13 13-1 Basc Busness Statstcs 11 th Edton Chapter 13 Smple Lnear Regresson Basc Busness Statstcs, 11e 009 Prentce-Hall, Inc. Chap 13-1 Learnng Objectves In ths chapter, you learn: How to use regresson

More information

7.1. Single classification analysis of variance (ANOVA) Why not use multiple 2-sample 2. When to use ANOVA

7.1. Single classification analysis of variance (ANOVA) Why not use multiple 2-sample 2. When to use ANOVA Sngle classfcaton analyss of varance (ANOVA) When to use ANOVA ANOVA models and parttonng sums of squares ANOVA: hypothess testng ANOVA: assumptons A non-parametrc alternatve: Kruskal-Walls ANOVA Power

More information

Harmonic oscillator approximation

Harmonic oscillator approximation armonc ocllator approxmaton armonc ocllator approxmaton Euaton to be olved We are fndng a mnmum of the functon under the retrcton where W P, P,..., P, Q, Q,..., Q P, P,..., P, Q, Q,..., Q lnwgner functon

More information

Topic- 11 The Analysis of Variance

Topic- 11 The Analysis of Variance Topc- 11 The Analyss of Varance Expermental Desgn The samplng plan or expermental desgn determnes the way that a sample s selected. In an observatonal study, the expermenter observes data that already

More information

x = , so that calculated

x = , so that calculated Stat 4, secton Sngle Factor ANOVA notes by Tm Plachowsk n chapter 8 we conducted hypothess tests n whch we compared a sngle sample s mean or proporton to some hypotheszed value Chapter 9 expanded ths to

More information

Lecture 6 More on Complete Randomized Block Design (RBD)

Lecture 6 More on Complete Randomized Block Design (RBD) Lecture 6 More on Complete Randomzed Block Desgn (RBD) Multple test Multple test The multple comparsons or multple testng problem occurs when one consders a set of statstcal nferences smultaneously. For

More information

Lecture 4 Hypothesis Testing

Lecture 4 Hypothesis Testing Lecture 4 Hypothess Testng We may wsh to test pror hypotheses about the coeffcents we estmate. We can use the estmates to test whether the data rejects our hypothess. An example mght be that we wsh to

More information

Correlation and Regression. Correlation 9.1. Correlation. Chapter 9

Correlation and Regression. Correlation 9.1. Correlation. Chapter 9 Chapter 9 Correlaton and Regresson 9. Correlaton Correlaton A correlaton s a relatonshp between two varables. The data can be represented b the ordered pars (, ) where s the ndependent (or eplanator) varable,

More information

Adaptive Centering with Random Effects in Studies of Time-Varying Treatments. by Stephen W. Raudenbush University of Chicago.

Adaptive Centering with Random Effects in Studies of Time-Varying Treatments. by Stephen W. Raudenbush University of Chicago. Adaptve Centerng wth Random Effect n Stde of Tme-Varyng Treatment by Stephen W. Radenbh Unverty of Chcago Abtract Of wdepread nteret n ocal cence are obervatonal tde n whch entte (peron chool tate contre

More information

CHAPTER IV RESEARCH FINDING AND ANALYSIS

CHAPTER IV RESEARCH FINDING AND ANALYSIS CHAPTER IV REEARCH FINDING AND ANALYI A. Descrpton of Research Fndngs To fnd out the dfference between the students who were taught by usng Mme Game and the students who were not taught by usng Mme Game

More information

Lecture 6: Introduction to Linear Regression

Lecture 6: Introduction to Linear Regression Lecture 6: Introducton to Lnear Regresson An Manchakul amancha@jhsph.edu 24 Aprl 27 Lnear regresson: man dea Lnear regresson can be used to study an outcome as a lnear functon of a predctor Example: 6

More information

Image Registration for a Series of Chest Radiograph Images

Image Registration for a Series of Chest Radiograph Images Proceedng of the 5th WE Internatonal Conference on gnal Proceng, Itanbul, Turkey, May 7-9, 006 (pp179-184) Image Regtraton for a ere of Chet Radograph Image Omar Mohd. Rjal*, Norlza Mohd. Noor, hee Lee

More information

Chapter 9: Statistical Inference and the Relationship between Two Variables

Chapter 9: Statistical Inference and the Relationship between Two Variables Chapter 9: Statstcal Inference and the Relatonshp between Two Varables Key Words The Regresson Model The Sample Regresson Equaton The Pearson Correlaton Coeffcent Learnng Outcomes After studyng ths chapter,

More information

Root Locus Techniques

Root Locus Techniques Root Locu Technque ELEC 32 Cloed-Loop Control The control nput u t ynthezed baed on the a pror knowledge of the ytem plant, the reference nput r t, and the error gnal, e t The control ytem meaure the output,

More information

Chapter 14 Simple Linear Regression

Chapter 14 Simple Linear Regression Chapter 4 Smple Lnear Regresson Chapter 4 - Smple Lnear Regresson Manageral decsons often are based on the relatonshp between two or more varables. Regresson analss can be used to develop an equaton showng

More information

STAT 511 FINAL EXAM NAME Spring 2001

STAT 511 FINAL EXAM NAME Spring 2001 STAT 5 FINAL EXAM NAME Sprng Instructons: Ths s a closed book exam. No notes or books are allowed. ou may use a calculator but you are not allowed to store notes or formulas n the calculator. Please wrte

More information

Simple Linear Regression Analysis

Simple Linear Regression Analysis LINEAR REGREION ANALYSIS MODULE II Lecture - 5 Smple Lear Regreo Aaly Dr Shalabh Departmet of Mathematc Stattc Ida Ittute of Techology Kapur Jot cofdece rego for A jot cofdece rego for ca alo be foud Such

More information

UNIVERSITY OF TORONTO Faculty of Arts and Science. December 2005 Examinations STA437H1F/STA1005HF. Duration - 3 hours

UNIVERSITY OF TORONTO Faculty of Arts and Science. December 2005 Examinations STA437H1F/STA1005HF. Duration - 3 hours UNIVERSITY OF TORONTO Faculty of Arts and Scence December 005 Examnatons STA47HF/STA005HF Duraton - hours AIDS ALLOWED: (to be suppled by the student) Non-programmable calculator One handwrtten 8.5'' x

More information

Topic 23 - Randomized Complete Block Designs (RCBD)

Topic 23 - Randomized Complete Block Designs (RCBD) Topc 3 ANOVA (III) 3-1 Topc 3 - Randomzed Complete Block Desgns (RCBD) Defn: A Randomzed Complete Block Desgn s a varant of the completely randomzed desgn (CRD) that we recently learned. In ths desgn,

More information

Introduction to Interfacial Segregation. Xiaozhe Zhang 10/02/2015

Introduction to Interfacial Segregation. Xiaozhe Zhang 10/02/2015 Introducton to Interfacal Segregaton Xaozhe Zhang 10/02/2015 Interfacal egregaton Segregaton n materal refer to the enrchment of a materal conttuent at a free urface or an nternal nterface of a materal.

More information

Here is the rationale: If X and y have a strong positive relationship to one another, then ( x x) will tend to be positive when ( y y)

Here is the rationale: If X and y have a strong positive relationship to one another, then ( x x) will tend to be positive when ( y y) Secton 1.5 Correlaton In the prevous sectons, we looked at regresson and the value r was a measurement of how much of the varaton n y can be attrbuted to the lnear relatonshp between y and x. In ths secton,

More information

Y = β 0 + β 1 X 1 + β 2 X β k X k + ε

Y = β 0 + β 1 X 1 + β 2 X β k X k + ε Chapter 3 Secton 3.1 Model Assumptons: Multple Regresson Model Predcton Equaton Std. Devaton of Error Correlaton Matrx Smple Lnear Regresson: 1.) Lnearty.) Constant Varance 3.) Independent Errors 4.) Normalty

More information

Properties of Umass Boston

Properties of Umass Boston Fle name hould be LatName_labNumber.doc or LatName_labNumber.doc.l. 0 pont wll be taken for wrong fle name. Follow the format of report and data heet. Both are poted n the web. MS Word and Ecel 003 format.

More information

[The following data appear in Wooldridge Q2.3.] The table below contains the ACT score and college GPA for eight college students.

[The following data appear in Wooldridge Q2.3.] The table below contains the ACT score and college GPA for eight college students. PPOL 59-3 Problem Set Exercses n Smple Regresson Due n class /8/7 In ths problem set, you are asked to compute varous statstcs by hand to gve you a better sense of the mechancs of the Pearson correlaton

More information

Chapter 5 Multilevel Models

Chapter 5 Multilevel Models Chapter 5 Multlevel Models 5.1 Cross-sectonal multlevel models 5.1.1 Two-level models 5.1.2 Multple level models 5.1.3 Multple level modelng n other felds 5.2 Longtudnal multlevel models 5.2.1 Two-level

More information

F statistic = s2 1 s 2 ( F for Fisher )

F statistic = s2 1 s 2 ( F for Fisher ) Stat 4 ANOVA Analyss of Varance /6/04 Comparng Two varances: F dstrbuton Typcal Data Sets One way analyss of varance : example Notaton for one way ANOVA Comparng Two varances: F dstrbuton We saw that the

More information

1. Inference on Regression Parameters a. Finding Mean, s.d and covariance amongst estimates. 2. Confidence Intervals and Working Hotelling Bands

1. Inference on Regression Parameters a. Finding Mean, s.d and covariance amongst estimates. 2. Confidence Intervals and Working Hotelling Bands Content. Inference on Regresson Parameters a. Fndng Mean, s.d and covarance amongst estmates.. Confdence Intervals and Workng Hotellng Bands 3. Cochran s Theorem 4. General Lnear Testng 5. Measures of

More information

January Examinations 2015

January Examinations 2015 24/5 Canddates Only January Examnatons 25 DO NOT OPEN THE QUESTION PAPER UNTIL INSTRUCTED TO DO SO BY THE CHIEF INVIGILATOR STUDENT CANDIDATE NO.. Department Module Code Module Ttle Exam Duraton (n words)

More information

Chapter 11: I = 2 samples independent samples paired samples Chapter 12: I 3 samples of equal size J one-way layout two-way layout

Chapter 11: I = 2 samples independent samples paired samples Chapter 12: I 3 samples of equal size J one-way layout two-way layout Serk Sagtov, Chalmers and GU, February 0, 018 Chapter 1. Analyss of varance Chapter 11: I = samples ndependent samples pared samples Chapter 1: I 3 samples of equal sze one-way layout two-way layout 1

More information

Statistics for Business and Economics

Statistics for Business and Economics Statstcs for Busness and Economcs Chapter 11 Smple Regresson Copyrght 010 Pearson Educaton, Inc. Publshng as Prentce Hall Ch. 11-1 11.1 Overvew of Lnear Models n An equaton can be ft to show the best lnear

More information

First Year Examination Department of Statistics, University of Florida

First Year Examination Department of Statistics, University of Florida Frst Year Examnaton Department of Statstcs, Unversty of Florda May 7, 010, 8:00 am - 1:00 noon Instructons: 1. You have four hours to answer questons n ths examnaton.. You must show your work to receve

More information

x yi In chapter 14, we want to perform inference (i.e. calculate confidence intervals and perform tests of significance) in this setting.

x yi In chapter 14, we want to perform inference (i.e. calculate confidence intervals and perform tests of significance) in this setting. The Practce of Statstcs, nd ed. Chapter 14 Inference for Regresson Introducton In chapter 3 we used a least-squares regresson lne (LSRL) to represent a lnear relatonshp etween two quanttatve explanator

More information

Optimal inference of sameness Supporting information

Optimal inference of sameness Supporting information Optmal nference of amene Supportng nformaton Content Decon rule of the optmal oberver.... Unequal relablte.... Equal relablte... 5 Repone probablte of the optmal oberver... 6. Equal relablte... 6. Unequal

More information

Improvements on Waring s Problem

Improvements on Waring s Problem Improvement on Warng Problem L An-Png Bejng, PR Chna apl@nacom Abtract By a new recurve algorthm for the auxlary equaton, n th paper, we wll gve ome mprovement for Warng problem Keyword: Warng Problem,

More information

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur Analyss of Varance and Desgn of Experment-I MODULE VII LECTURE - 3 ANALYSIS OF COVARIANCE Dr Shalabh Department of Mathematcs and Statstcs Indan Insttute of Technology Kanpur Any scentfc experment s performed

More information

2016 Wiley. Study Session 2: Ethical and Professional Standards Application

2016 Wiley. Study Session 2: Ethical and Professional Standards Application 6 Wley Study Sesson : Ethcal and Professonal Standards Applcaton LESSON : CORRECTION ANALYSIS Readng 9: Correlaton and Regresson LOS 9a: Calculate and nterpret a sample covarance and a sample correlaton

More information

Lecture 9: Linear regression: centering, hypothesis testing, multiple covariates, and confounding

Lecture 9: Linear regression: centering, hypothesis testing, multiple covariates, and confounding Recall: man dea of lnear regresson Lecture 9: Lnear regresson: centerng, hypothess testng, multple covarates, and confoundng Sandy Eckel seckel@jhsph.edu 6 May 8 Lnear regresson can be used to study an

More information

Chapter 6 The Effect of the GPS Systematic Errors on Deformation Parameters

Chapter 6 The Effect of the GPS Systematic Errors on Deformation Parameters Chapter 6 The Effect of the GPS Sytematc Error on Deformaton Parameter 6.. General Beutler et al., (988) dd the frt comprehenve tudy on the GPS ytematc error. Baed on a geometrc approach and aumng a unform

More information

Department of Statistics University of Toronto STA305H1S / 1004 HS Design and Analysis of Experiments Term Test - Winter Solution

Department of Statistics University of Toronto STA305H1S / 1004 HS Design and Analysis of Experiments Term Test - Winter Solution Department of Statstcs Unversty of Toronto STA35HS / HS Desgn and Analyss of Experments Term Test - Wnter - Soluton February, Last Name: Frst Name: Student Number: Instructons: Tme: hours. Ads: a non-programmable

More information

Lecture 9: Linear regression: centering, hypothesis testing, multiple covariates, and confounding

Lecture 9: Linear regression: centering, hypothesis testing, multiple covariates, and confounding Lecture 9: Lnear regresson: centerng, hypothess testng, multple covarates, and confoundng Sandy Eckel seckel@jhsph.edu 6 May 008 Recall: man dea of lnear regresson Lnear regresson can be used to study

More information

Linear Correlation. Many research issues are pursued with nonexperimental studies that seek to establish relationships among 2 or more variables

Linear Correlation. Many research issues are pursued with nonexperimental studies that seek to establish relationships among 2 or more variables Lnear Correlaton Many research ssues are pursued wth nonexpermental studes that seek to establsh relatonshps among or more varables E.g., correlates of ntellgence; relaton between SAT and GPA; relaton

More information

Answers Problem Set 2 Chem 314A Williamsen Spring 2000

Answers Problem Set 2 Chem 314A Williamsen Spring 2000 Answers Problem Set Chem 314A Wllamsen Sprng 000 1) Gve me the followng crtcal values from the statstcal tables. a) z-statstc,-sded test, 99.7% confdence lmt ±3 b) t-statstc (Case I), 1-sded test, 95%

More information

BOOTSTRAP METHOD FOR TESTING OF EQUALITY OF SEVERAL MEANS. M. Krishna Reddy, B. Naveen Kumar and Y. Ramu

BOOTSTRAP METHOD FOR TESTING OF EQUALITY OF SEVERAL MEANS. M. Krishna Reddy, B. Naveen Kumar and Y. Ramu BOOTSTRAP METHOD FOR TESTING OF EQUALITY OF SEVERAL MEANS M. Krshna Reddy, B. Naveen Kumar and Y. Ramu Department of Statstcs, Osmana Unversty, Hyderabad -500 007, Inda. nanbyrozu@gmal.com, ramu0@gmal.com

More information

No! Yes! Only if reactions occur! Yes! Ideal Gas, change in temperature or pressure. Survey Results. Class 15. Is the following possible?

No! Yes! Only if reactions occur! Yes! Ideal Gas, change in temperature or pressure. Survey Results. Class 15. Is the following possible? Survey Reult Chapter 5-6 (where we are gong) % of Student 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% Hour Spent on ChE 273 1-2 3-4 5-6 7-8 9-10 11+ Hour/Week 2008 2009 2010 2011 2012 2013 2014 2015 2017 F17

More information

CHAPTER IV RESEARCH FINDING AND DISCUSSIONS

CHAPTER IV RESEARCH FINDING AND DISCUSSIONS CHAPTER IV RESEARCH FINDING AND DISCUSSIONS A. Descrpton of Research Fndng. The Implementaton of Learnng Havng ganed the whole needed data, the researcher then dd analyss whch refers to the statstcal data

More information

Chapter 3 Describing Data Using Numerical Measures

Chapter 3 Describing Data Using Numerical Measures Chapter 3 Student Lecture Notes 3-1 Chapter 3 Descrbng Data Usng Numercal Measures Fall 2006 Fundamentals of Busness Statstcs 1 Chapter Goals To establsh the usefulness of summary measures of data. The

More information

Statistical tables are provided Two Hours UNIVERSITY OF MANCHESTER. Date: Wednesday 4 th June 2008 Time: 1400 to 1600

Statistical tables are provided Two Hours UNIVERSITY OF MANCHESTER. Date: Wednesday 4 th June 2008 Time: 1400 to 1600 Statstcal tables are provded Two Hours UNIVERSITY OF MNCHESTER Medcal Statstcs Date: Wednesday 4 th June 008 Tme: 1400 to 1600 MT3807 Electronc calculators may be used provded that they conform to Unversty

More information

Small signal analysis

Small signal analysis Small gnal analy. ntroducton Let u conder the crcut hown n Fg., where the nonlnear retor decrbed by the equaton g v havng graphcal repreentaton hown n Fg.. ( G (t G v(t v Fg. Fg. a D current ource wherea

More information

Alpha Risk of Taguchi Method with L 18 Array for NTB Type QCH by Simulation

Alpha Risk of Taguchi Method with L 18 Array for NTB Type QCH by Simulation Proceedng of the World Congre on Engneerng 00 Vol II WCE 00, July -, 00, London, U.K. Alpha Rk of Taguch Method wth L Array for NTB Type QCH by Smulaton A. Al-Refae and M.H. L Abtract Taguch method a wdely

More information

3 Implementation and validation of analysis methods

3 Implementation and validation of analysis methods 3 Implementaton and valdaton of anal method 3. Preface When mplementng new method bacall three cae can be dfferentated: - Implementaton of offcal method (nternatonall approved, valdated method, e.g. method

More information

A Survival-Adjusted Quantal-Response Test for Analysis of Tumor Incidence Rates in Animal Carcinogenicity Studies

A Survival-Adjusted Quantal-Response Test for Analysis of Tumor Incidence Rates in Animal Carcinogenicity Studies Reearch Survval-duted Quantal-Repone Tet for naly of Tumor Incdence Rate n nmal Carcnogencty Stude Shyamal D. Peddada and Grace E. lng otattc ranch Natonal Inttute of Envronmental Health Scence Natonal

More information

Not at Steady State! Yes! Only if reactions occur! Yes! Ideal Gas, change in temperature or pressure. Yes! Class 15. Is the following possible?

Not at Steady State! Yes! Only if reactions occur! Yes! Ideal Gas, change in temperature or pressure. Yes! Class 15. Is the following possible? Chapter 5-6 (where we are gong) Ideal gae and lqud (today) Dente Partal preure Non-deal gae (next tme) Eqn. of tate Reduced preure and temperature Compreblty chart (z) Vapor-lqud ytem (Ch. 6) Vapor preure

More information

Pythagorean triples. Leen Noordzij.

Pythagorean triples. Leen Noordzij. Pythagorean trple. Leen Noordz Dr.l.noordz@leennoordz.nl www.leennoordz.me Content A Roadmap for generatng Pythagorean Trple.... Pythagorean Trple.... 3 Dcuon Concluon.... 5 A Roadmap for generatng Pythagorean

More information

Chapter 15 - Multiple Regression

Chapter 15 - Multiple Regression Chapter - Multple Regresson Chapter - Multple Regresson Multple Regresson Model The equaton that descrbes how the dependent varable y s related to the ndependent varables x, x,... x p and an error term

More information

SIMPLE LINEAR REGRESSION

SIMPLE LINEAR REGRESSION Smple Lnear Regresson and Correlaton Introducton Prevousl, our attenton has been focused on one varable whch we desgnated b x. Frequentl, t s desrable to learn somethng about the relatonshp between two

More information

ECONOMICS 351*-A Mid-Term Exam -- Fall Term 2000 Page 1 of 13 pages. QUEEN'S UNIVERSITY AT KINGSTON Department of Economics

ECONOMICS 351*-A Mid-Term Exam -- Fall Term 2000 Page 1 of 13 pages. QUEEN'S UNIVERSITY AT KINGSTON Department of Economics ECOOMICS 35*-A Md-Term Exam -- Fall Term 000 Page of 3 pages QUEE'S UIVERSITY AT KIGSTO Department of Economcs ECOOMICS 35* - Secton A Introductory Econometrcs Fall Term 000 MID-TERM EAM ASWERS MG Abbott

More information

experimenteel en correlationeel onderzoek

experimenteel en correlationeel onderzoek expermenteel en correlatoneel onderzoek lecture 6: one-way analyss of varance Leary. Introducton to Behavoral Research Methods. pages 246 271 (chapters 10 and 11): conceptual statstcs Moore, McCabe, and

More information

where I = (n x n) diagonal identity matrix with diagonal elements = 1 and off-diagonal elements = 0; and σ 2 e = variance of (Y X).

where I = (n x n) diagonal identity matrix with diagonal elements = 1 and off-diagonal elements = 0; and σ 2 e = variance of (Y X). 11.4.1 Estmaton of Multple Regresson Coeffcents In multple lnear regresson, we essentally solve n equatons for the p unnown parameters. hus n must e equal to or greater than p and n practce n should e

More information

STATISTICS QUESTIONS. Step by Step Solutions.

STATISTICS QUESTIONS. Step by Step Solutions. STATISTICS QUESTIONS Step by Step Solutons www.mathcracker.com 9//016 Problem 1: A researcher s nterested n the effects of famly sze on delnquency for a group of offenders and examnes famles wth one to

More information

18. SIMPLE LINEAR REGRESSION III

18. SIMPLE LINEAR REGRESSION III 8. SIMPLE LINEAR REGRESSION III US Domestc Beers: Calores vs. % Alcohol Ftted Values and Resduals To each observed x, there corresponds a y-value on the ftted lne, y ˆ ˆ = α + x. The are called ftted values.

More information

ANSWERS. Problem 1. and the moment generating function (mgf) by. defined for any real t. Use this to show that E( U) var( U)

ANSWERS. Problem 1. and the moment generating function (mgf) by. defined for any real t. Use this to show that E( U) var( U) Econ 413 Exam 13 H ANSWERS Settet er nndelt 9 deloppgaver, A,B,C, som alle anbefales å telle lkt for å gøre det ltt lettere å stå. Svar er gtt . Unfortunately, there s a prntng error n the hnt of

More information

28. SIMPLE LINEAR REGRESSION III

28. SIMPLE LINEAR REGRESSION III 8. SIMPLE LINEAR REGRESSION III Ftted Values and Resduals US Domestc Beers: Calores vs. % Alcohol To each observed x, there corresponds a y-value on the ftted lne, y ˆ = βˆ + βˆ x. The are called ftted

More information

Start Point and Trajectory Analysis for the Minimal Time System Design Algorithm

Start Point and Trajectory Analysis for the Minimal Time System Design Algorithm Start Pont and Trajectory Analy for the Mnmal Tme Sytem Degn Algorthm ALEXANDER ZEMLIAK, PEDRO MIRANDA Department of Phyc and Mathematc Puebla Autonomou Unverty Av San Claudo /n, Puebla, 757 MEXICO Abtract:

More information

This column is a continuation of our previous column

This column is a continuation of our previous column Comparson of Goodness of Ft Statstcs for Lnear Regresson, Part II The authors contnue ther dscusson of the correlaton coeffcent n developng a calbraton for quanttatve analyss. Jerome Workman Jr. and Howard

More information

STAT 3008 Applied Regression Analysis

STAT 3008 Applied Regression Analysis STAT 3008 Appled Regresson Analyss Tutoral : Smple Lnear Regresson LAI Chun He Department of Statstcs, The Chnese Unversty of Hong Kong 1 Model Assumpton To quantfy the relatonshp between two factors,

More information

ANALYSIS OF CORRELATED DATA SAMPLING FROM CLUSTERS CLUSTER-RANDOMIZED TRIALS CLUSTERED DATA. Sample size formula (assuming known σ )

ANALYSIS OF CORRELATED DATA SAMPLING FROM CLUSTERS CLUSTER-RANDOMIZED TRIALS CLUSTERED DATA. Sample size formula (assuming known σ ) November 8 ANALYSIS OF CORRELATED DATA SAMPLING FROM CLUSTERS CLUSTER-RANDOMIZED TRIALS Background Independent obervaton: Short revew of well-known fact Comparon of two group contnuou repone Control group:

More information

Predictive Analytics : QM901.1x Prof U Dinesh Kumar, IIMB. All Rights Reserved, Indian Institute of Management Bangalore

Predictive Analytics : QM901.1x Prof U Dinesh Kumar, IIMB. All Rights Reserved, Indian Institute of Management Bangalore Sesson Outlne Introducton to classfcaton problems and dscrete choce models. Introducton to Logstcs Regresson. Logstc functon and Logt functon. Maxmum Lkelhood Estmator (MLE) for estmaton of LR parameters.

More information

Statistical Inference. 2.3 Summary Statistics Measures of Center and Spread. parameters ( population characteristics )

Statistical Inference. 2.3 Summary Statistics Measures of Center and Spread. parameters ( population characteristics ) Ismor Fscher, 8//008 Stat 54 / -8.3 Summary Statstcs Measures of Center and Spread Dstrbuton of dscrete contnuous POPULATION Random Varable, numercal True center =??? True spread =???? parameters ( populaton

More information

BETWEEN-PARTICIPANTS EXPERIMENTAL DESIGNS

BETWEEN-PARTICIPANTS EXPERIMENTAL DESIGNS 1 BETWEEN-PARTICIPANTS EXPERIMENTAL DESIGNS I. Sngle-factor desgns: the model s: y j = µ + α + ε j = µ + ε j where: y j jth observaton n the sample from the th populaton ( = 1,..., I; j = 1,..., n ) µ

More information

Statistics MINITAB - Lab 2

Statistics MINITAB - Lab 2 Statstcs 20080 MINITAB - Lab 2 1. Smple Lnear Regresson In smple lnear regresson we attempt to model a lnear relatonshp between two varables wth a straght lne and make statstcal nferences concernng that

More information

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur Analyss of Varance and Desgn of Experment-I MODULE VIII LECTURE - 34 ANALYSIS OF VARIANCE IN RANDOM-EFFECTS MODEL AND MIXED-EFFECTS EFFECTS MODEL Dr Shalabh Department of Mathematcs and Statstcs Indan

More information

Introduction. Modeling Data. Approach. Quality of Fit. Likelihood. Probabilistic Approach

Introduction. Modeling Data. Approach. Quality of Fit. Likelihood. Probabilistic Approach Introducton Modelng Data Gven a et of obervaton, we wh to ft a mathematcal model Model deend on adutable arameter traght lne: m + c n Polnomal: a + a + a + L+ a n Choce of model deend uon roblem Aroach

More information

Systematic Error Illustration of Bias. Sources of Systematic Errors. Effects of Systematic Errors 9/23/2009. Instrument Errors Method Errors Personal

Systematic Error Illustration of Bias. Sources of Systematic Errors. Effects of Systematic Errors 9/23/2009. Instrument Errors Method Errors Personal 9/3/009 Sstematc Error Illustraton of Bas Sources of Sstematc Errors Instrument Errors Method Errors Personal Prejudce Preconceved noton of true value umber bas Prefer 0/5 Small over large Even over odd

More information

APPROXIMATE FUZZY REASONING BASED ON INTERPOLATION IN THE VAGUE ENVIRONMENT OF THE FUZZY RULEBASE AS A PRACTICAL ALTERNATIVE OF THE CLASSICAL CRI

APPROXIMATE FUZZY REASONING BASED ON INTERPOLATION IN THE VAGUE ENVIRONMENT OF THE FUZZY RULEBASE AS A PRACTICAL ALTERNATIVE OF THE CLASSICAL CRI Kovác, Sz., Kóczy, L.T.: Approxmate Fuzzy Reaonng Baed on Interpolaton n the Vague Envronment of the Fuzzy Rulebae a a Practcal Alternatve of the Clacal CRI, Proceedng of the 7 th Internatonal Fuzzy Sytem

More information

Chapter 8 Indicator Variables

Chapter 8 Indicator Variables Chapter 8 Indcator Varables In general, e explanatory varables n any regresson analyss are assumed to be quanttatve n nature. For example, e varables lke temperature, dstance, age etc. are quanttatve n

More information