ALBERT VEXLER, SHULING LIU, LE KANG, AND ALAN DAVID HUTSON

Size: px
Start display at page:

Download "ALBERT VEXLER, SHULING LIU, LE KANG, AND ALAN DAVID HUTSON"

Transcription

1 Commuicatios i tatistics imulatio ad Computatio, 38: , 2009 Copyright Taylor & Fracis Group, LLC IN: prit/ olie DOI: / Modificatios of the Empirical Likelihood Iterval Estimatio with Improved Coverage Probabilities ALBERT VEXLER, HULING LIU, LE KANG, AND ALAN DAVID HUTON Departmet of Biostatistics, The tate Uiversity of New York at Buffalo, Buffalo, New York, UA 1. Itroductio The empirical likelihood EL) techique has bee well addressed i both the theoretical ad applied literature i the cotext of powerful oparametric statistical methods for testig ad iterval estimatios. A oparametric versio of Wilks theorem Wilks, 1938) ca usually provide a asymptotic evaluatio of the Type I error of EL ratio-type tests. I this article, we examie the performace of this asymptotic result whe the EL is based o fiite samples that are from various distributios. I the cotext of the Type I error cotrol, we show that the classical EL procedure ad the tudet s t-test have asymptotically a similar structure. Thus, we coclude that modificatios of t-type tests ca be adopted to improve the EL ratio test. We propose the applicatio of the Che 1995) t-test modificatio to the EL ratio test. We display that the Che approach leads to a locatio chage of observed data whereas the classical Bartlett method is kow to be a scale correctio of the data distributio. Fially, we modify the EL ratio test via both the Che ad Bartlett correctios. We support our argumet with theoretical proofs as well as a Mote Carlo study. A real data example studies the proposed approach i practice. Keywords Bartlett Correctio; Empirical likelihood; Iterval estimatio; Likelihood ratio; Modified t-test; Noparametric test; igificace level; Type I error. Mathematics ubject Classificatio 62G10; 62G20. Likelihood methods are commo statistical tools that provide very powerful methods of statistical iferece. The parametric likelihood methodology requires the assumptio that the distributio fuctio of the idividual observatios have kow forms e.g., Vexler et al., 2009). To relax the parametric assumptios, Owe 1988) itroduced the empirical likelihood EL) methodology as a oparametric versio of the parametric likelihood techique. I a wide rage of applicatios, Received March 13, 2009; Accepted August 13, 2009 Address correspodece to Albert Vexler, Departmet of Biostatistics, The tate Uiversity of New York, Farber Hall, Room 246, Buffalo, NY 14214, UA; avexler@buffalo.edu 2171

2 2172 Vexler et al. the oparametric EL ratio tests have comparable advatages to those of the parametric likelihood methodologies. EL methods have bee well accepted i the literature as very powerful statistical procedures i the distributio-free settig relative to parametric approaches; e.g., see DiCiccio et al. 1991), Hall ad cala 1990), Owe 1988, 1990, 1991, 2001), Qi ad Lawless 1994), ad Yu et al. 2009). Other results iclude that of Owe 1990) who proved that the asymptotic distributio fuctio of twice the logarithm of a EL ratio-type test has the form of a chi-square distributio fuctio. This asymptotic result is a oparametric versio of the Wilks theorem e.g., Wilks, 1938), which is kow i the cotext of the asymptotic ull distributio of the maximum parametric likelihood ratio test statistic. However, it is well kow that the distace betwee the actual ull distributio of a test statistic ad its asymptotic approximatio ca be sigificat. To acquire the Type I error cotrol related to the EL ratio test, the Bartlett correctio ca be applied; e.g., see Owe 2001). The Bartlett correctio is made by multiplyig the relevat test thresholds by a scale parameter, values of which ca be chose to improve upo Wilks approximatio. I practice, a fiite sample size strogly restricts the Type I error cotrol that is obtaied through Wilks approximatio. I this case, the applicatio of Bartlett s correctio ca be also difficult. This is due to the fact that this correctio is based upo asymptotic priciples, which i tur deped fuctioally o ukow parameters, such as the 3rd ad 4th ukow momets of the uderlyig populatio, ad hece have to be estimated. Whe the sample size is relatively small, the ukow parameters utilized withi Bartlett s correctio ca be poorly estimated such that the efficiecy of the Bartlett s techique deteriorates. It should also be oted that the EL ratio test statistic i certai situatios is ot Bartlett correctable; e.g., see Che ad Cui 2006) ad Lazar ad Myklad 1999). Corcora 1998), as well as Daviso ad tafford 1998), remarked that the Bartlett correctio applied to EL ratios does ot provide sigificatly effective results i certai practical situatios. I this article, we ivestigate the Bartlett-corrected EL ratio test based o samples with fiite sizes. Data with skewed distributios ca also affect the efficiecy of the EL method. I this article, utilizig both theoretical ad empirical argumets, we show that, i the cotext of the Type I error cotrol, whe the data is o ormally distributed ad skewed especially whe the sample size is relatively small), the EL ratio test ca be corrected i a similar maer to that of a tudet s t-test correctio. That is, if some correctio ca improve tudet s t-type tests, the it ca also be adopted to the EL ratio-type test. Che 1995) proposed a modificatio of the t-test that was show to be more accurate ad more powerful tha both the Johso s modified t-test ad the utto s composite test. The Che approach is very efficiet for small sample sizes. Thus, we propose to apply the Che 1995) method to modify the EL ratio test. This correctio is differet from the Bartlett modificatio that rescales the log EL ratio i order to improve the secod-order approximatio via a chi-square distributio. Alteratively, we show that Che correctio trasforms observatios, chagig the locatio of data. To take advatages of both the scale ad locatio correctios, we propose a ew modificatio of the EL ratio test combiig the Che ad Bartlett techiques. To achieve our goals metioed above, sketch theoretical proofs of the ew modificatios are provided. A Mote Carlo study is used to evaluate the cotrol of Type I error related to the proposed procedures. A real data example is provided i order to illustrate the proposed method. ectio 5 cotais our cocludig remarks.

3 Modified Empirical Likelihood Ratio Tests Methods Assume, for simplicity, we observe idepedet idetically distributed i.i.d.) radom variables X 1 X 2 X with E X 1 3 <. We would like to test the hypothesis H 0 E X 1 = vs. H 1 E X 1 1) Furthermore, before we proceed ote that the hypothesis above ca be exteded easily to the more geeral case, amely, H 0 E g X = vs. H 1 E g X where g is a give fuctio. To test the hypothesis above defied at 1) oparametrically, first defie the EL fuctio i the form of L X 1 X = p i where p i = 1. Here, p i s, i = 1, have a sese of Pr X i = x i The values of p i s are ukow ad should be obtaied uder the ull or the alterative hypothesis. Uder the ull hypothesis, the EL methodology e.g., Owe, 2001; Yu et al., 2009) requires oe to fid the value of the p i s that maximize the EL that satisfy the empirical costraits p i = 1 ad p ix i =. Usig Lagrage multipliers, oe ca show that uder the ull hypothesis, the maximum EL fuctio has the followig form: where is a root of L X 1 X = sup 0<p 1 p 2 p <1 p i =1 p i X i = p i = X i 1 + X i = 0 1 2) 1 + X i Uder the alterative hypothesis, just the empirical costrait p i = 1 is i effect. The L X 1 X = sup 0<p 1 p 2 p <1 p i =1 p i = 1 = ) 1 3) Combiig 2) ad 3), we obtai the log EL ratio, test statistic for 1) i the form of l X 1 X = 2 log L X 1 X L X 1 X = 2 log 1 + X i 4)

4 2174 Vexler et al. Owe 1988) proved that l X 1 X has a asymptotic chi-square distributio, uder the ull hypothesis. Hece, we ca test H 0 at the approximated sigificace level usig the test l X 1 X = 2 where the test threshold C distributio. Note that oe ca show log 1 + X i >C is the % percetile for a chi-square l X 1 X X i 2 X i 2 as 5) where X i 2 / X i 2 has a asymptotic chi-square distributio uder the ull hypothesis give at 1). The asymptotic propositio 5) ad its proof are similar to results that were demostrated ad utilized by Qi ad Lawless 1994) i a differet cotext. For details, see the Appedix. This asymptotic result motivates us to cosider the t-test statistic T X 1 X = X i X i X 2 X = ) X i / which is asymptotically equivalet to 2 log EL ratio, i.e., by virtue of 5) oe ca show that the 2 log EL ratio is approximated by T 2 that i tur coverges i distributio to a chi-square distributio. This idicates that the EL ratio test ad the tudet s t-test have similar properties i the cotext of the Type I error cotrol. Note that Owe 1990) cocluded that large sample sizes strogly improve the accuracy of the Wilks-type approximatios of the distributio of ull EL ratio test statistic. I the cotext of the Type I error cotrol, we demostrate that the classical EL procedure ad tudet s t-test based o observatios from the same distributios have similar properties. That is, whe data are close to beig ormally distributed, the Type I error cotrol related to the tests based o the data is expected to be accurate eve whe, the sample size, is small. However, whe the sample size is large ad the observatios are from a logormal distributio, either the t-test or the EL ratio test cotrol the Type I error well. I this case, both tests ca be affected strogly by the skewess of data. I ec. 3, we cofirm this coclusio via a Mote Carlo study. I certai situatios, the EL ratio test is Bartlett correctable; e.g., see Owe 2001). The Bartlett correctio is a theoretical adjustmet for the secod-order approximatio of a log-likelihood ratio distributio; e.g., see Hall ad cala 1990). However, the Bartlett correctio caot completely solve the problem of the Type I error cotrol whe the data is skewed. Moreover, whe the Bartlett correctio is applied certai ukow parameters eed to be estimated. The accuracy of this estimatio ca affect the overall efficiecy of the Bartlett correctio. Note that Lazar ad Myklad 1999) proved that whe uisace parameters are preset the EL ratio statistic is ot Bartlett correctable. Whe the Bartlett correctio is i effect, the test statistic is multiplied by a coefficiet. This rescalig of the log likelihood ratio is used to reduce the discrepacy betwee a chi-square distributio ad the

5 Modified Empirical Likelihood Ratio Tests 2175 asymptotic ull distributio of the log likelihood ratio. Thus, we refer to this a scale correctio. Alteratively, sice some modificatios of the t-test, differet from the Bartlett methodology, ca improve the relevat Type I error cotrol, we ca cosider a similar modificatio of the EL ratio test followig the asymptotic equivalece betwee the two tests metioed above. We outlie the approach of Che, showig this method chages the locatio of the sample istead of providig a scalig correctio. I the cotext of the hypothesis H 0 E X = vs. H 1 E X > 6) defie ˆ 1 = X i X 3 / 3, where = X i X 2 / 1 ) 0 5. The the Che modified t-test ca be preseted as where T X 1 X + X 1 X 7) [ X 1 X = 1 { } 2 ] 6 ˆ X [ + 1 { } 2 ] X 9 ˆ 2 X This correctio was obtaied usig a Edgeworth expasio applied to the t-test whe the distributio of data is positively skewed Che, 1995). The test statistic 7) ca be represeted by the regular t-test statistic based o observatios X i = X i + X 1 X i = 1 8) Thus, the Che approach chages the locatio of X i s. Therefore, we ca classify the Che method as a locatio correctio. Assume we chage the locatio of X i s ad defie the EL ratio based o relocated data ad derive a optimal chage of the locatio of the X i s. Towards this ed, the EL ratio test is created utilizig the costraits X i + d p i = ad p i = 1 uder the ull H 0, where d is a shift of the locatio. Let us obtai a form of d to correct the EL ratio test statistic followig the Che approach. By virtue of 5), we have the Type I error { } P H0 l X 1 X >C P X i + d 2 H0 X >C i + d 2 = P H0 { R d > C or R d < C } 9)

6 2176 Vexler et al. where R d = X i + d X i + d 2 ) X d = X + i + d 2 / X i + d 2 Usig the Taylor expasio with d aroud 0, we obtai that R d = T X 1 X + d X i 2 + o d2 as d 0 Here, the correspodig remaied term has a order of d 2 ad ca be igored. ice R d > C ad R d T X 1 X + d / X i 2, the by the Che methodology we must set up d 1 d 0 5 X i 2) 0 5 = X 1 X as i 7) i order to correct the Type I error. Thus, we coclude that the optimal chage i the data locatio give the rule R d > C is d = 1 X 1 X By virtue of 9), ad i order to approximate the decisio rule l X >C, we should also cosider R d < C. I a similar maer to the aalysis metioed above, for a egatively skewed populatio, we should set up d = 1 X 1 X Here, whe we cosidered R d < C, to preserve the statemet of 6), we utilized observatios X 1 X keepig the sig > of the alterative hypothesis of 6). That explais the form of d = 0 5 X 1 X, comparig with the locatio chage 0 5 X 1 X give the rule R d > C. Thus, we propose the followig modificatio of the EL ratio test: 1. Calculate values of the statistics l 1 = l X 1 + X 1 X X + X ) 1 X ad l 2 = l X 1 + X 1 X X + X ) 1 X 2. Reject the ull hypothesis of 1) if l Z 1 ) 2 or l Z 1 ) 2

7 Modified Empirical Likelihood Ratio Tests 2177 where Z 1 is the percetile of the stadard ormal distributio. Note that, if we ca presuppose the uderlyig distributio fuctio of observatios is positively or egatively skewed, we the ca defie the test statistic l = l X 1 + X 1 X for positively skewed data; l = l for egatively skewed data, X i + X 1 X X + X 1 X ) X + X 1 X ) respectively, ad reject the ull hypothesis if l 2 1 1, where is the % percetile of the chi-square distributio with the degree of freedom of 1. Hece, we showed that the Che approach ca be also adapted to be applied to the EL ratio test. I accordace with the previously metioed remarks, we ca cosider the Bartlett ad Che techiques as two differet test-modificatios. The Bartlett correctio ca reduce the slope caused by the differece betwee the chi-square distributio ad the log EL ratio distributio ad the Che approach corrects the bias caused by the skewess of a ull distributio of observatios. Thus, we coclude to combie these modificatios ad to propose the ext decisio rule. We suggest rejectig the ull hypothesis at 1) if l Z 1 ) 1 + a 0 5 or l 2 ) Z 1 ) 1 + a ) where a is a estimator of 1 E X EX 4 2 E X EX 2 1 E X EX E X EX 2 3 Note that, by virtue of the relatio betwee the testig ad iterval estimatio, we ca obtai the cofidece iterval estimator of i the form of CI 1 = X T C 1 where T presets a test statistic for the hypothesis 1) ad C 1 is the correspodig threshold, assumig that the omial coverage probability is 1. Therefore, the previoulsy cosidered issue of the Type I error cotrol is also relevat to a problem to improve the accuracy of the iterval estimatio. 3. Mote Carlo tudy I this sectio, we provide a Mote Carlo study to compare the followig five tests: 1) the stadard EL ratio test; 2) the proposed Che-corrected EL ratio test; 3) the Bartlett-corrected EL ratio test; 4) the proposed Che Bartlett-corrected EL ratio test; ad 5) the tudet s t-test. Towards this ed, the ull distributio

8 2178 Vexler et al. fuctios were chose to be i Normal ad Logormal forms, sice the asymptotic result metioed i ec. 2 showed that we ca expect a good approximatio to the test-statistic distributio whe data are from a ormally distributed populatio, whereas the approximatio ca be expected to be worse i the case with logormally distributed data. We also geerated data followig a chi-square distributio to ivestigate the proposed method. Radom variables were draw from the ull distributios with differet parameters. At each baselie distributio ad each set of parameters, we repeated 10,000 times data geeratios. The Mote Carlo Type I errors are preseted i Tables 1 3. As is show i Table 1, the t-test demostrates the actual Type I errors that are closer to the expected 0.05 tha those of the cosidered tests, uder the ormal distributio assumptio. The stadard EL ratio method has well cotrolled Type I errors whe the sample sizes are greater tha 25. The Bartlett ad Che correctios are assumed to reduce a bias caused by the differece betwee the chi-square distributio ad log EL ratio asymptotic distributio. I this case, this estimatio is ot ecessary; however, the proposed Che Bartlett correctio is readily available. Table 2 displays situatios whe data follow a logormal distributio. The otatio LN a b correspods to a logormal distributio where a ad b are the mea ad stadard deviatio of the logarithm. I accordace with the aalysis metioed i ec. 2, we ca expect correctios of the test statistics are eeded i this case. Here, the Mote Carlo study cofirms this coclusio. The Che Bartlettcorrected EL ratio test is the best of all the five test statistics. Especially whe the data are strogly skewed, e.g., i the cases with LN 1 2, LN 2 2, = 25, the Mote Carlo Type I errors for the EL ratio test ad the t-test are greater tha 0.3 we expected the sigificace level at 0.05). Whe the combiatio of Che Table 1 The Mote Carlo Type I Errors of the tests based o data from the ull ormal distributios: X Normal 0 2 ad H 0 EX= 0 The expected Type I error is 0.05) ample size ELR Che Bartlett Che & Bartlett T-test

9 Modified Empirical Likelihood Ratio Tests 2179 Table 2 The Mote Carlo Type I errors correspoded to ull logormal distributios: X Logormal 2 exp ad H 0 EX= 0 The expected Type I error is 0.05) ample size ELR Che Bartlett Che & Bartlett T-test 25 0, 1.0) , 1.5) , 2.0) , 1.0) , 1.5) , 2.0) , 2.0) , 1.0) , 1.5) , 2.0) , 1.0) , 1.5) , 2.0) , 2.0) , 1.0) , 1.5) , 2.0) , 1.0) , 1.5) , 2.0) , 2.0) , 1.0) , 1.5) , 2.0) , 1.0) , 1.5) , 2.0) , 2.0) Table 3 The Mote Carlo Type I errors correspoded to ull chi-square distributios: X ad H 0 EX= 0 The expected Type I error is 0.05) ample size ELR Che Bartlett Che & Bartlett T-test

10 2180 Vexler et al. ad Bartlett correctios were applied, the Type I error was about two times better approximated tha that of the two other tests. Therefore, we ca coclude that the proposed method ca sigificatly improve the Type I error cotrol ad the accuracy to cover probabilities of the iterval estimatio. Table 3 is related to the cases where data follow a chi-square distributio. From Table 3, the proposed Che Bartlett-corrected EL ratio test is show to have the best ability to cotrol the Type I error especially whe the sample sizes are relatively small. With the icreasig of the sample size, all the tests ted to perform well. 4. A Glucose Data Example I this sectio, we use the proposed method to costruct the cofidece iterval estimators based o data from a study that evaluates biomarkers related to atherosclerotic coroary heart disease. We cosider the biomarker glucose that ca quatify differet phases of oxidative stress ad atioxidat status process of a idividual e.g., chisterma et al., 2001). A populatio-based sample of radomly selected residets of Erie ad Niagara couties, years of age, was the focus of this ivestigatio chisterma et al., 2001). The samplig frame for adults betwee the ages of 35 ad 65 was from the New York tate Departmet of Motor Vehicle drivers licese rolls, while the elderly sample age 65 79) was radomly chose from the Health Care Fiacig Admiistratio database. A cohort of 917 me ad wome were selected for the aalyses to preset a populatio without myocardial ifarctio. Participats provided a 12-h fastig food specime for biochemical aalysis at baselie, ad a umber of parameters were examied from fresh blood samples. We illustrate the proposed method based o the biomarker glucose that has values with a right-skewed distributio. The strategy was that a sample with size N was radomly selected from the glucose data to estimate the cofidece itervals at the level 95% of the mea, say a b ; the rest of the data was used to calculate the average of the glucose levels, say c. The value of 917-N was chose to be large, ad hece we believe the computed average c is very close to the theoretical mea of the glucose levels. We repeated this strategy 10,000 times calculatig the frequecies of the evet c a b. This bootstrap-type test shows that the Che Bartlett ad Che corrected EL ratio tests provides the coverage probabilities that are closer to the target 0.95 tha those of the EL ratio test ad the Bartlett corrected EL ratio test ad the t-test. Table 4 presets these results for the differet sample sizes of Table 4 The coverage probabilities related to the glucose biomaker The expected coverage probability is 0.95) ample size ELR Che Bartlett Che & Bartlett T-test

11 Modified Empirical Likelihood Ratio Tests 2181 Table 5 The cofidece iterval estimators at the level 95% of the mea of the glucose biomarker measured from 917 patiets without myocardial ifarctio ELR , ) Che , ) Bartlett , ) Che&Bartlett , ) T-test , ) N = 15 25, 35, 50, ad 100. It ca be oticed that whe the sample size is above 50, the actual coverage probabilities is very close to the omial oe. Thus, we propose to use the Che Bartlett ad Che corrected EL ratio statistics to estimate the cofidece iterval for the mea of glucose measuremets. The ivestigated test statistics provided the cofidece itervals at the level 95% of the mea of the glucose biomarker measured from 917 patiets without myocardial ifarctio the healthy populatio). The results are show i Table 5. The ELR, Bartlett corrected ELR, ad T-test show approximately similar results, the use of which ca be misleadig, takig ito accout the outputs of Table 4 ad that glucose measuremets have a right skewed distributio. 5. Coclusios The EL methodology has bee well addressed i the literature as a oparametric versio of its powerful parametric likelihood couterparts. A importat property is that the asymptotic distributio of 2 log EL ratio test statistics follow a chi-square distributio. However, this approximatio ca be biased especially whe the sample size is small. I this article, we illustrated the skewess of the data distributio ca also affect the efficiecy of the EL ratio tests. With respect to improvig the Type I error cotrol of the tests ad coverage probability of the iterval estimatio the EL techique ad the Bartlett correctio of the EL ratio tests are well accepted i the literature. We proved that the EL ratio test ca be corrected i a similar maer to a classic t-test s correctio. Thus, the Che 1995) approach, which was proposed to correct the skewess of the t-test distributio, was adapted to be applied to the EL ratio test. The Bartlett correctio was show to be based o scalig modified data, whereas the Che method was preseted via chagig the locatio. We oted that the Bartlett techique requires certai additioal coditios, e.g., o uisace parameters related to the statemet of problem, Cramér s coditios, which are uecessary whe the Che correctio is assumed to be applied. I this article, we also proposed to combie both the Bartlett ad Che techiques as the ew modificatio of the EL ratio test i order to improve the accuracy of the iterval estimatios. Our argumets are cofirmed via both the theoretical cosideratio ad the Mote Carlo study. The Bartlett Che-corrected EL ratio test was show to perform much better tha the stadard EL ratio test ad also the classic tudet s t-test whe data were from skewed distributio. To demostrate that the proposed

12 2182 Vexler et al. method ca be utilized i practice, we applied the proposed method to costruct the cofidece itervals for the glucose data that is from a coroary heart disease study. The proposed methodology ca also be applied to differet EL-type procedures. Appedix A.1. ketch Proof of the Asymptotic Chi-quare Distributio of l X 1 X We have l X 1 X = 2 log L X 1 X = 2 log 1 + X i where is a root of the equatio g = X i = 0 A1) 1 + X i Applyig the Taylor expasio, we ca expad g with aroud 0. Hece, by A1), we coclude that X i X i 2 as A2) ice l X 1 X = 2 log 1 + X i, ow cosider l X 1 X with Taylor series where is close to 0 ad after pluggig A2) to the statistic 2 log 1 + X i we obtai l X 1 X X i 2 X A3) i 2 Note that the test statistic T X 1 X 2 has a asymptotic chi-square distributio as follows T X 1 X = X i p N 0 1 X i X 2 ad hece l X 1 X T 2 p X 1 X 2 1. This completes the brief proof. Ackowledgmet The authors are grateful to the Editor ad referee for their helpful commets that clearly improved this article. Refereces Che, L. 1995). Testig the mea of skewed distributios. Joural of the America tatistical Associatio 90:

13 Modified Empirical Likelihood Ratio Tests 2183 Che,. X., Cui, H. 2006). O Bartlett correctio of empirical likelihood i the presece of uisace parameters. Biometrika 93: Corcora,. A. 1998). Adjustmet of empirical discrepacy statistics. Biometrika 85: Daviso, A. C., tafford, J. E. 1998). The score fuctio ad a compariso of various adjustmets of the profile likelihood. The Caadia Joural of tatistics 26: DiCiccio, T., Hall, P., Romao, J. 1991). Empirical likelihood is Bartlett-correctable. The Aals of tatistics 19: Hall, P., cala, L. B. 1990). Methodology ad algorithms of empirical likelihood. Iteratioal tatistical Review/Revue Iteratioale de tatistique 58: Lazar, N. A., Myklad, A. P. 1999). Empirical likelihood i the presece of uisace parameters. Biometrika 86: Owe, A. B. 1988). Empirical likelihood ratio cofidece itervals for a sigle fuctioal. Biometrika 75: Owe, A. B. 1990). Empirical likelihood cofidece regios. The Aals of tatistics 18: Owe, A. B. 1991). Empirical likelihood for liear models. The Aals of tatistics 19: Owe, A. B. 2001). Empirical Likelihood. New York: Chapma ad Hall. Qi, J., Lawless, J. 1994). Empirical likelihood ad geeral estimatig equatios. The Aals of tatistics 22: chisterma, E. F., Faraggi, D., Browe, R., Freudeheim, J., Dor, J., Muti, P., Armstrog, D., Reiser, B., Trevisa, M. 2001). TBAR ad cardiovascular disease i a populatio-based sample. Joural of Cardiovascular Risk 8:1 7. Vexler, A., Wu, C., Yu, K. F. 2009). Optimal hypothesis testig: From semi to fully Bayes factors. Metrika I press. DOI /s Wilks, ). The large-sample distributio of the likelihood ratio for testig composite hypotheses. Aals of Mathematical tatistics 9: Yu, J., Vexler, A., Tia, L. 2009). Aalyzig icomplete data subject to a threshold usig empirical likelihood methods: A applicatio to a peumoia risk study i a ICU settig. Biometrics I press. DOI /j x

POWER COMPARISON OF EMPIRICAL LIKELIHOOD RATIO TESTS: SMALL SAMPLE PROPERTIES THROUGH MONTE CARLO STUDIES*

POWER COMPARISON OF EMPIRICAL LIKELIHOOD RATIO TESTS: SMALL SAMPLE PROPERTIES THROUGH MONTE CARLO STUDIES* Kobe Uiversity Ecoomic Review 50(2004) 3 POWER COMPARISON OF EMPIRICAL LIKELIHOOD RATIO TESTS: SMALL SAMPLE PROPERTIES THROUGH MONTE CARLO STUDIES* By HISASHI TANIZAKI There are various kids of oparametric

More information

Goodness-Of-Fit For The Generalized Exponential Distribution. Abstract

Goodness-Of-Fit For The Generalized Exponential Distribution. Abstract Goodess-Of-Fit For The Geeralized Expoetial Distributio By Amal S. Hassa stitute of Statistical Studies & Research Cairo Uiversity Abstract Recetly a ew distributio called geeralized expoetial or expoetiated

More information

A statistical method to determine sample size to estimate characteristic value of soil parameters

A statistical method to determine sample size to estimate characteristic value of soil parameters A statistical method to determie sample size to estimate characteristic value of soil parameters Y. Hojo, B. Setiawa 2 ad M. Suzuki 3 Abstract Sample size is a importat factor to be cosidered i determiig

More information

Confidence Interval for Standard Deviation of Normal Distribution with Known Coefficients of Variation

Confidence Interval for Standard Deviation of Normal Distribution with Known Coefficients of Variation Cofidece Iterval for tadard Deviatio of Normal Distributio with Kow Coefficiets of Variatio uparat Niwitpog Departmet of Applied tatistics, Faculty of Applied ciece Kig Mogkut s Uiversity of Techology

More information

Properties and Hypothesis Testing

Properties and Hypothesis Testing Chapter 3 Properties ad Hypothesis Testig 3.1 Types of data The regressio techiques developed i previous chapters ca be applied to three differet kids of data. 1. Cross-sectioal data. 2. Time series data.

More information

1 Inferential Methods for Correlation and Regression Analysis

1 Inferential Methods for Correlation and Regression Analysis 1 Iferetial Methods for Correlatio ad Regressio Aalysis I the chapter o Correlatio ad Regressio Aalysis tools for describig bivariate cotiuous data were itroduced. The sample Pearso Correlatio Coefficiet

More information

FACULTY OF MATHEMATICAL STUDIES MATHEMATICS FOR PART I ENGINEERING. Lectures

FACULTY OF MATHEMATICAL STUDIES MATHEMATICS FOR PART I ENGINEERING. Lectures FACULTY OF MATHEMATICAL STUDIES MATHEMATICS FOR PART I ENGINEERING Lectures MODULE 5 STATISTICS II. Mea ad stadard error of sample data. Biomial distributio. Normal distributio 4. Samplig 5. Cofidece itervals

More information

Resampling Methods. X (1/2), i.e., Pr (X i m) = 1/2. We order the data: X (1) X (2) X (n). Define the sample median: ( n.

Resampling Methods. X (1/2), i.e., Pr (X i m) = 1/2. We order the data: X (1) X (2) X (n). Define the sample median: ( n. Jauary 1, 2019 Resamplig Methods Motivatio We have so may estimators with the property θ θ d N 0, σ 2 We ca also write θ a N θ, σ 2 /, where a meas approximately distributed as Oce we have a cosistet estimator

More information

ECE 8527: Introduction to Machine Learning and Pattern Recognition Midterm # 1. Vaishali Amin Fall, 2015

ECE 8527: Introduction to Machine Learning and Pattern Recognition Midterm # 1. Vaishali Amin Fall, 2015 ECE 8527: Itroductio to Machie Learig ad Patter Recogitio Midterm # 1 Vaishali Ami Fall, 2015 tue39624@temple.edu Problem No. 1: Cosider a two-class discrete distributio problem: ω 1 :{[0,0], [2,0], [2,2],

More information

Random Variables, Sampling and Estimation

Random Variables, Sampling and Estimation Chapter 1 Radom Variables, Samplig ad Estimatio 1.1 Itroductio This chapter will cover the most importat basic statistical theory you eed i order to uderstad the ecoometric material that will be comig

More information

If, for instance, we were required to test whether the population mean μ could be equal to a certain value μ

If, for instance, we were required to test whether the population mean μ could be equal to a certain value μ STATISTICAL INFERENCE INTRODUCTION Statistical iferece is that brach of Statistics i which oe typically makes a statemet about a populatio based upo the results of a sample. I oesample testig, we essetially

More information

A proposed discrete distribution for the statistical modeling of

A proposed discrete distribution for the statistical modeling of It. Statistical Ist.: Proc. 58th World Statistical Cogress, 0, Dubli (Sessio CPS047) p.5059 A proposed discrete distributio for the statistical modelig of Likert data Kidd, Marti Cetre for Statistical

More information

Confidence interval for the two-parameter exponentiated Gumbel distribution based on record values

Confidence interval for the two-parameter exponentiated Gumbel distribution based on record values Iteratioal Joural of Applied Operatioal Research Vol. 4 No. 1 pp. 61-68 Witer 2014 Joural homepage: www.ijorlu.ir Cofidece iterval for the two-parameter expoetiated Gumbel distributio based o record values

More information

Approximate Confidence Interval for the Reciprocal of a Normal Mean with a Known Coefficient of Variation

Approximate Confidence Interval for the Reciprocal of a Normal Mean with a Known Coefficient of Variation Metodološki zvezki, Vol. 13, No., 016, 117-130 Approximate Cofidece Iterval for the Reciprocal of a Normal Mea with a Kow Coefficiet of Variatio Wararit Paichkitkosolkul 1 Abstract A approximate cofidece

More information

ON BARTLETT CORRECTABILITY OF EMPIRICAL LIKELIHOOD IN GENERALIZED POWER DIVERGENCE FAMILY. Lorenzo Camponovo and Taisuke Otsu.

ON BARTLETT CORRECTABILITY OF EMPIRICAL LIKELIHOOD IN GENERALIZED POWER DIVERGENCE FAMILY. Lorenzo Camponovo and Taisuke Otsu. ON BARTLETT CORRECTABILITY OF EMPIRICAL LIKELIHOOD IN GENERALIZED POWER DIVERGENCE FAMILY By Lorezo Campoovo ad Taisuke Otsu October 011 COWLES FOUNDATION DISCUSSION PAPER NO. 185 COWLES FOUNDATION FOR

More information

Investigating the Significance of a Correlation Coefficient using Jackknife Estimates

Investigating the Significance of a Correlation Coefficient using Jackknife Estimates Iteratioal Joural of Scieces: Basic ad Applied Research (IJSBAR) ISSN 2307-4531 (Prit & Olie) http://gssrr.org/idex.php?joural=jouralofbasicadapplied ---------------------------------------------------------------------------------------------------------------------------

More information

Comparing Two Populations. Topic 15 - Two Sample Inference I. Comparing Two Means. Comparing Two Pop Means. Background Reading

Comparing Two Populations. Topic 15 - Two Sample Inference I. Comparing Two Means. Comparing Two Pop Means. Background Reading Topic 15 - Two Sample Iferece I STAT 511 Professor Bruce Craig Comparig Two Populatios Research ofte ivolves the compariso of two or more samples from differet populatios Graphical summaries provide visual

More information

Lorenzo Camponovo, Taisuke Otsu On Bartlett correctability of empirical likelihood in generalized power divergence family

Lorenzo Camponovo, Taisuke Otsu On Bartlett correctability of empirical likelihood in generalized power divergence family Lorezo Campoovo, Taisuke Otsu O Bartlett correctability of empirical likelihood i geeralized power divergece family Article Accepted versio) Refereed) Origial citatio: Campoovo, Lorezo ad Otsu, Taisuke

More information

2 1. The r.s., of size n2, from population 2 will be. 2 and 2. 2) The two populations are independent. This implies that all of the n1 n2

2 1. The r.s., of size n2, from population 2 will be. 2 and 2. 2) The two populations are independent. This implies that all of the n1 n2 Chapter 8 Comparig Two Treatmets Iferece about Two Populatio Meas We wat to compare the meas of two populatios to see whether they differ. There are two situatios to cosider, as show i the followig examples:

More information

BIOS 4110: Introduction to Biostatistics. Breheny. Lab #9

BIOS 4110: Introduction to Biostatistics. Breheny. Lab #9 BIOS 4110: Itroductio to Biostatistics Brehey Lab #9 The Cetral Limit Theorem is very importat i the realm of statistics, ad today's lab will explore the applicatio of it i both categorical ad cotiuous

More information

1 Review of Probability & Statistics

1 Review of Probability & Statistics 1 Review of Probability & Statistics a. I a group of 000 people, it has bee reported that there are: 61 smokers 670 over 5 960 people who imbibe (drik alcohol) 86 smokers who imbibe 90 imbibers over 5

More information

Double Stage Shrinkage Estimator of Two Parameters. Generalized Exponential Distribution

Double Stage Shrinkage Estimator of Two Parameters. Generalized Exponential Distribution Iteratioal Mathematical Forum, Vol., 3, o. 3, 3-53 HIKARI Ltd, www.m-hikari.com http://dx.doi.org/.9/imf.3.335 Double Stage Shrikage Estimator of Two Parameters Geeralized Expoetial Distributio Alaa M.

More information

Chapter 6 Sampling Distributions

Chapter 6 Sampling Distributions Chapter 6 Samplig Distributios 1 I most experimets, we have more tha oe measuremet for ay give variable, each measuremet beig associated with oe radomly selected a member of a populatio. Hece we eed to

More information

Sample Size Estimation in the Proportional Hazards Model for K-sample or Regression Settings Scott S. Emerson, M.D., Ph.D.

Sample Size Estimation in the Proportional Hazards Model for K-sample or Regression Settings Scott S. Emerson, M.D., Ph.D. ample ie Estimatio i the Proportioal Haards Model for K-sample or Regressio ettigs cott. Emerso, M.D., Ph.D. ample ie Formula for a Normally Distributed tatistic uppose a statistic is kow to be ormally

More information

Because it tests for differences between multiple pairs of means in one test, it is called an omnibus test.

Because it tests for differences between multiple pairs of means in one test, it is called an omnibus test. Math 308 Sprig 018 Classes 19 ad 0: Aalysis of Variace (ANOVA) Page 1 of 6 Itroductio ANOVA is a statistical procedure for determiig whether three or more sample meas were draw from populatios with equal

More information

Table 12.1: Contingency table. Feature b. 1 N 11 N 12 N 1b 2 N 21 N 22 N 2b. ... a N a1 N a2 N ab

Table 12.1: Contingency table. Feature b. 1 N 11 N 12 N 1b 2 N 21 N 22 N 2b. ... a N a1 N a2 N ab Sectio 12 Tests of idepedece ad homogeeity I this lecture we will cosider a situatio whe our observatios are classified by two differet features ad we would like to test if these features are idepedet

More information

PSYCHOLOGICAL RESEARCH (PYC 304-C) Lecture 9

PSYCHOLOGICAL RESEARCH (PYC 304-C) Lecture 9 Hypothesis testig PSYCHOLOGICAL RESEARCH (PYC 34-C Lecture 9 Statistical iferece is that brach of Statistics i which oe typically makes a statemet about a populatio based upo the results of a sample. I

More information

Problem Set 4 Due Oct, 12

Problem Set 4 Due Oct, 12 EE226: Radom Processes i Systems Lecturer: Jea C. Walrad Problem Set 4 Due Oct, 12 Fall 06 GSI: Assae Gueye This problem set essetially reviews detectio theory ad hypothesis testig ad some basic otios

More information

EXAMINATIONS OF THE ROYAL STATISTICAL SOCIETY

EXAMINATIONS OF THE ROYAL STATISTICAL SOCIETY EXAMINATIONS OF THE ROYAL STATISTICAL SOCIETY GRADUATE DIPLOMA, 016 MODULE : Statistical Iferece Time allowed: Three hours Cadidates should aswer FIVE questios. All questios carry equal marks. The umber

More information

This chapter focuses on two experimental designs that are crucial to comparative studies: (1) independent samples and (2) matched pair samples.

This chapter focuses on two experimental designs that are crucial to comparative studies: (1) independent samples and (2) matched pair samples. Chapter 9 & : Comparig Two Treatmets: This chapter focuses o two eperimetal desigs that are crucial to comparative studies: () idepedet samples ad () matched pair samples Idepedet Radom amples from Two

More information

Statistical Inference (Chapter 10) Statistical inference = learn about a population based on the information provided by a sample.

Statistical Inference (Chapter 10) Statistical inference = learn about a population based on the information provided by a sample. Statistical Iferece (Chapter 10) Statistical iferece = lear about a populatio based o the iformatio provided by a sample. Populatio: The set of all values of a radom variable X of iterest. Characterized

More information

MOST PEOPLE WOULD RATHER LIVE WITH A PROBLEM THEY CAN'T SOLVE, THAN ACCEPT A SOLUTION THEY CAN'T UNDERSTAND.

MOST PEOPLE WOULD RATHER LIVE WITH A PROBLEM THEY CAN'T SOLVE, THAN ACCEPT A SOLUTION THEY CAN'T UNDERSTAND. XI-1 (1074) MOST PEOPLE WOULD RATHER LIVE WITH A PROBLEM THEY CAN'T SOLVE, THAN ACCEPT A SOLUTION THEY CAN'T UNDERSTAND. R. E. D. WOOLSEY AND H. S. SWANSON XI-2 (1075) STATISTICAL DECISION MAKING Advaced

More information

Lecture 5: Parametric Hypothesis Testing: Comparing Means. GENOME 560, Spring 2016 Doug Fowler, GS

Lecture 5: Parametric Hypothesis Testing: Comparing Means. GENOME 560, Spring 2016 Doug Fowler, GS Lecture 5: Parametric Hypothesis Testig: Comparig Meas GENOME 560, Sprig 2016 Doug Fowler, GS (dfowler@uw.edu) 1 Review from last week What is a cofidece iterval? 2 Review from last week What is a cofidece

More information

GG313 GEOLOGICAL DATA ANALYSIS

GG313 GEOLOGICAL DATA ANALYSIS GG313 GEOLOGICAL DATA ANALYSIS 1 Testig Hypothesis GG313 GEOLOGICAL DATA ANALYSIS LECTURE NOTES PAUL WESSEL SECTION TESTING OF HYPOTHESES Much of statistics is cocered with testig hypothesis agaist data

More information

7-1. Chapter 4. Part I. Sampling Distributions and Confidence Intervals

7-1. Chapter 4. Part I. Sampling Distributions and Confidence Intervals 7-1 Chapter 4 Part I. Samplig Distributios ad Cofidece Itervals 1 7- Sectio 1. Samplig Distributio 7-3 Usig Statistics Statistical Iferece: Predict ad forecast values of populatio parameters... Test hypotheses

More information

Goodness-Of-Fit For The Generalized Exponential Distribution. Abstract

Goodness-Of-Fit For The Generalized Exponential Distribution. Abstract Goodess-Of-Fit For The Geeralized Expoetial Distributio By Amal S. Hassa stitute of Statistical Studies & Research Cairo Uiversity Abstract Recetly a ew distributio called geeralized expoetial or expoetiated

More information

Tests of Hypotheses Based on a Single Sample (Devore Chapter Eight)

Tests of Hypotheses Based on a Single Sample (Devore Chapter Eight) Tests of Hypotheses Based o a Sigle Sample Devore Chapter Eight MATH-252-01: Probability ad Statistics II Sprig 2018 Cotets 1 Hypothesis Tests illustrated with z-tests 1 1.1 Overview of Hypothesis Testig..........

More information

A NEW METHOD FOR CONSTRUCTING APPROXIMATE CONFIDENCE INTERVALS FOR M-ESTU1ATES. Dennis D. Boos

A NEW METHOD FOR CONSTRUCTING APPROXIMATE CONFIDENCE INTERVALS FOR M-ESTU1ATES. Dennis D. Boos .- A NEW METHOD FOR CONSTRUCTING APPROXIMATE CONFIDENCE INTERVALS FOR M-ESTU1ATES by Deis D. Boos Departmet of Statistics North Carolia State Uiversity Istitute of Statistics Mimeo Series #1198 September,

More information

Chapter 22. Comparing Two Proportions. Copyright 2010 Pearson Education, Inc.

Chapter 22. Comparing Two Proportions. Copyright 2010 Pearson Education, Inc. Chapter 22 Comparig Two Proportios Copyright 2010 Pearso Educatio, Ic. Comparig Two Proportios Comparisos betwee two percetages are much more commo tha questios about isolated percetages. Ad they are more

More information

Lecture 33: Bootstrap

Lecture 33: Bootstrap Lecture 33: ootstrap Motivatio To evaluate ad compare differet estimators, we eed cosistet estimators of variaces or asymptotic variaces of estimators. This is also importat for hypothesis testig ad cofidece

More information

Expectation and Variance of a random variable

Expectation and Variance of a random variable Chapter 11 Expectatio ad Variace of a radom variable The aim of this lecture is to defie ad itroduce mathematical Expectatio ad variace of a fuctio of discrete & cotiuous radom variables ad the distributio

More information

Topic 9: Sampling Distributions of Estimators

Topic 9: Sampling Distributions of Estimators Topic 9: Samplig Distributios of Estimators Course 003, 2016 Page 0 Samplig distributios of estimators Sice our estimators are statistics (particular fuctios of radom variables), their distributio ca be

More information

Overview. p 2. Chapter 9. Pooled Estimate of. q = 1 p. Notation for Two Proportions. Inferences about Two Proportions. Assumptions

Overview. p 2. Chapter 9. Pooled Estimate of. q = 1 p. Notation for Two Proportions. Inferences about Two Proportions. Assumptions Chapter 9 Slide Ifereces from Two Samples 9- Overview 9- Ifereces about Two Proportios 9- Ifereces about Two Meas: Idepedet Samples 9-4 Ifereces about Matched Pairs 9-5 Comparig Variatio i Two Samples

More information

Econ 325 Notes on Point Estimator and Confidence Interval 1 By Hiro Kasahara

Econ 325 Notes on Point Estimator and Confidence Interval 1 By Hiro Kasahara Poit Estimator Eco 325 Notes o Poit Estimator ad Cofidece Iterval 1 By Hiro Kasahara Parameter, Estimator, ad Estimate The ormal probability desity fuctio is fully characterized by two costats: populatio

More information

Lecture 2: Monte Carlo Simulation

Lecture 2: Monte Carlo Simulation STAT/Q SCI 43: Itroductio to Resamplig ethods Sprig 27 Istructor: Ye-Chi Che Lecture 2: ote Carlo Simulatio 2 ote Carlo Itegratio Assume we wat to evaluate the followig itegratio: e x3 dx What ca we do?

More information

This is an introductory course in Analysis of Variance and Design of Experiments.

This is an introductory course in Analysis of Variance and Design of Experiments. 1 Notes for M 384E, Wedesday, Jauary 21, 2009 (Please ote: I will ot pass out hard-copy class otes i future classes. If there are writte class otes, they will be posted o the web by the ight before class

More information

The standard deviation of the mean

The standard deviation of the mean Physics 6C Fall 20 The stadard deviatio of the mea These otes provide some clarificatio o the distictio betwee the stadard deviatio ad the stadard deviatio of the mea.. The sample mea ad variace Cosider

More information

Additional Notes and Computational Formulas CHAPTER 3

Additional Notes and Computational Formulas CHAPTER 3 Additioal Notes ad Computatioal Formulas APPENDIX CHAPTER 3 1 The Greek capital sigma is the mathematical sig for summatio If we have a sample of observatios say y 1 y 2 y 3 y their sum is y 1 + y 2 +

More information

Output Analysis (2, Chapters 10 &11 Law)

Output Analysis (2, Chapters 10 &11 Law) B. Maddah ENMG 6 Simulatio Output Aalysis (, Chapters 10 &11 Law) Comparig alterative system cofiguratio Sice the output of a simulatio is radom, the comparig differet systems via simulatio should be doe

More information

Chapter 22. Comparing Two Proportions. Copyright 2010, 2007, 2004 Pearson Education, Inc.

Chapter 22. Comparing Two Proportions. Copyright 2010, 2007, 2004 Pearson Education, Inc. Chapter 22 Comparig Two Proportios Copyright 2010, 2007, 2004 Pearso Educatio, Ic. Comparig Two Proportios Read the first two paragraphs of pg 504. Comparisos betwee two percetages are much more commo

More information

Module 1 Fundamentals in statistics

Module 1 Fundamentals in statistics Normal Distributio Repeated observatios that differ because of experimetal error ofte vary about some cetral value i a roughly symmetrical distributio i which small deviatios occur much more frequetly

More information

Frequentist Inference

Frequentist Inference Frequetist Iferece The topics of the ext three sectios are useful applicatios of the Cetral Limit Theorem. Without kowig aythig about the uderlyig distributio of a sequece of radom variables {X i }, for

More information

Control Charts for Mean for Non-Normally Correlated Data

Control Charts for Mean for Non-Normally Correlated Data Joural of Moder Applied Statistical Methods Volume 16 Issue 1 Article 5 5-1-017 Cotrol Charts for Mea for No-Normally Correlated Data J. R. Sigh Vikram Uiversity, Ujjai, Idia Ab Latif Dar School of Studies

More information

April 18, 2017 CONFIDENCE INTERVALS AND HYPOTHESIS TESTING, UNDERGRADUATE MATH 526 STYLE

April 18, 2017 CONFIDENCE INTERVALS AND HYPOTHESIS TESTING, UNDERGRADUATE MATH 526 STYLE April 18, 2017 CONFIDENCE INTERVALS AND HYPOTHESIS TESTING, UNDERGRADUATE MATH 526 STYLE TERRY SOO Abstract These otes are adapted from whe I taught Math 526 ad meat to give a quick itroductio to cofidece

More information

Introductory statistics

Introductory statistics CM9S: Machie Learig for Bioiformatics Lecture - 03/3/06 Itroductory statistics Lecturer: Sriram Sakararama Scribe: Sriram Sakararama We will provide a overview of statistical iferece focussig o the key

More information

GUIDE FOR THE USE OF THE DECISION SUPPORT SYSTEM (DSS)*

GUIDE FOR THE USE OF THE DECISION SUPPORT SYSTEM (DSS)* GUIDE FOR THE USE OF THE DECISION SUPPORT SYSTEM (DSS)* *Note: I Frech SAD (Système d Aide à la Décisio) 1. Itroductio to the DSS Eightee statistical distributios are available i HYFRAN-PLUS software to

More information

Since X n /n P p, we know that X n (n. Xn (n X n ) Using the asymptotic result above to obtain an approximation for fixed n, we obtain

Since X n /n P p, we know that X n (n. Xn (n X n ) Using the asymptotic result above to obtain an approximation for fixed n, we obtain Assigmet 9 Exercise 5.5 Let X biomial, p, where p 0, 1 is ukow. Obtai cofidece itervals for p i two differet ways: a Sice X / p d N0, p1 p], the variace of the limitig distributio depeds oly o p. Use the

More information

Bootstrap Intervals of the Parameters of Lognormal Distribution Using Power Rule Model and Accelerated Life Tests

Bootstrap Intervals of the Parameters of Lognormal Distribution Using Power Rule Model and Accelerated Life Tests Joural of Moder Applied Statistical Methods Volume 5 Issue Article --5 Bootstrap Itervals of the Parameters of Logormal Distributio Usig Power Rule Model ad Accelerated Life Tests Mohammed Al-Ha Ebrahem

More information

Hypothesis Testing. Evaluation of Performance of Learned h. Issues. Trade-off Between Bias and Variance

Hypothesis Testing. Evaluation of Performance of Learned h. Issues. Trade-off Between Bias and Variance Hypothesis Testig Empirically evaluatig accuracy of hypotheses: importat activity i ML. Three questios: Give observed accuracy over a sample set, how well does this estimate apply over additioal samples?

More information

Topic 9: Sampling Distributions of Estimators

Topic 9: Sampling Distributions of Estimators Topic 9: Samplig Distributios of Estimators Course 003, 2018 Page 0 Samplig distributios of estimators Sice our estimators are statistics (particular fuctios of radom variables), their distributio ca be

More information

Inferential Statistics. Inference Process. Inferential Statistics and Probability a Holistic Approach. Inference Process.

Inferential Statistics. Inference Process. Inferential Statistics and Probability a Holistic Approach. Inference Process. Iferetial Statistics ad Probability a Holistic Approach Iferece Process Chapter 8 Poit Estimatio ad Cofidece Itervals This Course Material by Maurice Geraghty is licesed uder a Creative Commos Attributio-ShareAlike

More information

Measurement uncertainty of the sound absorption

Measurement uncertainty of the sound absorption Measuremet ucertaity of the soud absorptio coefficiet Aa Izewska Buildig Research Istitute, Filtrowa Str., 00-6 Warsaw, Polad a.izewska@itb.pl 6887 The stadard ISO/IEC 705:005 o the competece of testig

More information

GUIDELINES ON REPRESENTATIVE SAMPLING

GUIDELINES ON REPRESENTATIVE SAMPLING DRUGS WORKING GROUP VALIDATION OF THE GUIDELINES ON REPRESENTATIVE SAMPLING DOCUMENT TYPE : REF. CODE: ISSUE NO: ISSUE DATE: VALIDATION REPORT DWG-SGL-001 002 08 DECEMBER 2012 Ref code: DWG-SGL-001 Issue

More information

Discrete Mathematics for CS Spring 2008 David Wagner Note 22

Discrete Mathematics for CS Spring 2008 David Wagner Note 22 CS 70 Discrete Mathematics for CS Sprig 2008 David Wager Note 22 I.I.D. Radom Variables Estimatig the bias of a coi Questio: We wat to estimate the proportio p of Democrats i the US populatio, by takig

More information

STA Learning Objectives. Population Proportions. Module 10 Comparing Two Proportions. Upon completing this module, you should be able to:

STA Learning Objectives. Population Proportions. Module 10 Comparing Two Proportions. Upon completing this module, you should be able to: STA 2023 Module 10 Comparig Two Proportios Learig Objectives Upo completig this module, you should be able to: 1. Perform large-sample ifereces (hypothesis test ad cofidece itervals) to compare two populatio

More information

Econ 325/327 Notes on Sample Mean, Sample Proportion, Central Limit Theorem, Chi-square Distribution, Student s t distribution 1.

Econ 325/327 Notes on Sample Mean, Sample Proportion, Central Limit Theorem, Chi-square Distribution, Student s t distribution 1. Eco 325/327 Notes o Sample Mea, Sample Proportio, Cetral Limit Theorem, Chi-square Distributio, Studet s t distributio 1 Sample Mea By Hiro Kasahara We cosider a radom sample from a populatio. Defiitio

More information

Convergence of random variables. (telegram style notes) P.J.C. Spreij

Convergence of random variables. (telegram style notes) P.J.C. Spreij Covergece of radom variables (telegram style otes).j.c. Spreij this versio: September 6, 2005 Itroductio As we kow, radom variables are by defiitio measurable fuctios o some uderlyig measurable space

More information

Statistics 511 Additional Materials

Statistics 511 Additional Materials Cofidece Itervals o mu Statistics 511 Additioal Materials This topic officially moves us from probability to statistics. We begi to discuss makig ifereces about the populatio. Oe way to differetiate probability

More information

DS 100: Principles and Techniques of Data Science Date: April 13, Discussion #10

DS 100: Principles and Techniques of Data Science Date: April 13, Discussion #10 DS 00: Priciples ad Techiques of Data Sciece Date: April 3, 208 Name: Hypothesis Testig Discussio #0. Defie these terms below as they relate to hypothesis testig. a) Data Geeratio Model: Solutio: A set

More information

Common Large/Small Sample Tests 1/55

Common Large/Small Sample Tests 1/55 Commo Large/Small Sample Tests 1/55 Test of Hypothesis for the Mea (σ Kow) Covert sample result ( x) to a z value Hypothesis Tests for µ Cosider the test H :μ = μ H 1 :μ > μ σ Kow (Assume the populatio

More information

There is no straightforward approach for choosing the warmup period l.

There is no straightforward approach for choosing the warmup period l. B. Maddah INDE 504 Discrete-Evet Simulatio Output Aalysis () Statistical Aalysis for Steady-State Parameters I a otermiatig simulatio, the iterest is i estimatig the log ru steady state measures of performace.

More information

6.3 Testing Series With Positive Terms

6.3 Testing Series With Positive Terms 6.3. TESTING SERIES WITH POSITIVE TERMS 307 6.3 Testig Series With Positive Terms 6.3. Review of what is kow up to ow I theory, testig a series a i for covergece amouts to fidig the i= sequece of partial

More information

Lecture 8: Non-parametric Comparison of Location. GENOME 560, Spring 2016 Doug Fowler, GS

Lecture 8: Non-parametric Comparison of Location. GENOME 560, Spring 2016 Doug Fowler, GS Lecture 8: No-parametric Compariso of Locatio GENOME 560, Sprig 2016 Doug Fowler, GS (dfowler@uw.edu) 1 Review What do we mea by oparametric? What is a desirable locatio statistic for ordial data? What

More information

Power and Type II Error

Power and Type II Error Statistical Methods I (EXST 7005) Page 57 Power ad Type II Error Sice we do't actually kow the value of the true mea (or we would't be hypothesizig somethig else), we caot kow i practice the type II error

More information

The Sampling Distribution of the Maximum. Likelihood Estimators for the Parameters of. Beta-Binomial Distribution

The Sampling Distribution of the Maximum. Likelihood Estimators for the Parameters of. Beta-Binomial Distribution Iteratioal Mathematical Forum, Vol. 8, 2013, o. 26, 1263-1277 HIKARI Ltd, www.m-hikari.com http://d.doi.org/10.12988/imf.2013.3475 The Samplig Distributio of the Maimum Likelihood Estimators for the Parameters

More information

Sample Size Determination (Two or More Samples)

Sample Size Determination (Two or More Samples) Sample Sie Determiatio (Two or More Samples) STATGRAPHICS Rev. 963 Summary... Data Iput... Aalysis Summary... 5 Power Curve... 5 Calculatios... 6 Summary This procedure determies a suitable sample sie

More information

CHAPTER 4 BIVARIATE DISTRIBUTION EXTENSION

CHAPTER 4 BIVARIATE DISTRIBUTION EXTENSION CHAPTER 4 BIVARIATE DISTRIBUTION EXTENSION 4. Itroductio Numerous bivariate discrete distributios have bee defied ad studied (see Mardia, 97 ad Kocherlakota ad Kocherlakota, 99) based o various methods

More information

10. Comparative Tests among Spatial Regression Models. Here we revisit the example in Section 8.1 of estimating the mean of a normal random

10. Comparative Tests among Spatial Regression Models. Here we revisit the example in Section 8.1 of estimating the mean of a normal random Part III. Areal Data Aalysis 0. Comparative Tests amog Spatial Regressio Models While the otio of relative likelihood values for differet models is somewhat difficult to iterpret directly (as metioed above),

More information

Introduction to Econometrics (3 rd Updated Edition) Solutions to Odd- Numbered End- of- Chapter Exercises: Chapter 3

Introduction to Econometrics (3 rd Updated Edition) Solutions to Odd- Numbered End- of- Chapter Exercises: Chapter 3 Itroductio to Ecoometrics (3 rd Updated Editio) by James H. Stock ad Mark W. Watso Solutios to Odd- Numbered Ed- of- Chapter Exercises: Chapter 3 (This versio August 17, 014) 015 Pearso Educatio, Ic. Stock/Watso

More information

[412] A TEST FOR HOMOGENEITY OF THE MARGINAL DISTRIBUTIONS IN A TWO-WAY CLASSIFICATION

[412] A TEST FOR HOMOGENEITY OF THE MARGINAL DISTRIBUTIONS IN A TWO-WAY CLASSIFICATION [412] A TEST FOR HOMOGENEITY OF THE MARGINAL DISTRIBUTIONS IN A TWO-WAY CLASSIFICATION BY ALAN STUART Divisio of Research Techiques, Lodo School of Ecoomics 1. INTRODUCTION There are several circumstaces

More information

4. Partial Sums and the Central Limit Theorem

4. Partial Sums and the Central Limit Theorem 1 of 10 7/16/2009 6:05 AM Virtual Laboratories > 6. Radom Samples > 1 2 3 4 5 6 7 4. Partial Sums ad the Cetral Limit Theorem The cetral limit theorem ad the law of large umbers are the two fudametal theorems

More information

Agreement of CI and HT. Lecture 13 - Tests of Proportions. Example - Waiting Times

Agreement of CI and HT. Lecture 13 - Tests of Proportions. Example - Waiting Times Sigificace level vs. cofidece level Agreemet of CI ad HT Lecture 13 - Tests of Proportios Sta102 / BME102 Coli Rudel October 15, 2014 Cofidece itervals ad hypothesis tests (almost) always agree, as log

More information

A goodness-of-fit test based on the empirical characteristic function and a comparison of tests for normality

A goodness-of-fit test based on the empirical characteristic function and a comparison of tests for normality A goodess-of-fit test based o the empirical characteristic fuctio ad a compariso of tests for ormality J. Marti va Zyl Departmet of Mathematical Statistics ad Actuarial Sciece, Uiversity of the Free State,

More information

Product measures, Tonelli s and Fubini s theorems For use in MAT3400/4400, autumn 2014 Nadia S. Larsen. Version of 13 October 2014.

Product measures, Tonelli s and Fubini s theorems For use in MAT3400/4400, autumn 2014 Nadia S. Larsen. Version of 13 October 2014. Product measures, Toelli s ad Fubii s theorems For use i MAT3400/4400, autum 2014 Nadia S. Larse Versio of 13 October 2014. 1. Costructio of the product measure The purpose of these otes is to preset the

More information

A RANK STATISTIC FOR NON-PARAMETRIC K-SAMPLE AND CHANGE POINT PROBLEMS

A RANK STATISTIC FOR NON-PARAMETRIC K-SAMPLE AND CHANGE POINT PROBLEMS J. Japa Statist. Soc. Vol. 41 No. 1 2011 67 73 A RANK STATISTIC FOR NON-PARAMETRIC K-SAMPLE AND CHANGE POINT PROBLEMS Yoichi Nishiyama* We cosider k-sample ad chage poit problems for idepedet data i a

More information

Statistical inference: example 1. Inferential Statistics

Statistical inference: example 1. Inferential Statistics Statistical iferece: example 1 Iferetial Statistics POPULATION SAMPLE A clothig store chai regularly buys from a supplier large quatities of a certai piece of clothig. Each item ca be classified either

More information

A new distribution-free quantile estimator

A new distribution-free quantile estimator Biometrika (1982), 69, 3, pp. 635-40 Prited i Great Britai 635 A ew distributio-free quatile estimator BY FRANK E. HARRELL Cliical Biostatistics, Duke Uiversity Medical Ceter, Durham, North Carolia, U.S.A.

More information

HYPOTHESIS TESTS FOR ONE POPULATION MEAN WORKSHEET MTH 1210, FALL 2018

HYPOTHESIS TESTS FOR ONE POPULATION MEAN WORKSHEET MTH 1210, FALL 2018 HYPOTHESIS TESTS FOR ONE POPULATION MEAN WORKSHEET MTH 1210, FALL 2018 We are resposible for 2 types of hypothesis tests that produce ifereces about the ukow populatio mea, µ, each of which has 3 possible

More information

Class 23. Daniel B. Rowe, Ph.D. Department of Mathematics, Statistics, and Computer Science. Marquette University MATH 1700

Class 23. Daniel B. Rowe, Ph.D. Department of Mathematics, Statistics, and Computer Science. Marquette University MATH 1700 Class 23 Daiel B. Rowe, Ph.D. Departmet of Mathematics, Statistics, ad Computer Sciece Copyright 2017 by D.B. Rowe 1 Ageda: Recap Chapter 9.1 Lecture Chapter 9.2 Review Exam 6 Problem Solvig Sessio. 2

More information

Economics 241B Relation to Method of Moments and Maximum Likelihood OLSE as a Maximum Likelihood Estimator

Economics 241B Relation to Method of Moments and Maximum Likelihood OLSE as a Maximum Likelihood Estimator Ecoomics 24B Relatio to Method of Momets ad Maximum Likelihood OLSE as a Maximum Likelihood Estimator Uder Assumptio 5 we have speci ed the distributio of the error, so we ca estimate the model parameters

More information

MATH/STAT 352: Lecture 15

MATH/STAT 352: Lecture 15 MATH/STAT 352: Lecture 15 Sectios 5.2 ad 5.3. Large sample CI for a proportio ad small sample CI for a mea. 1 5.2: Cofidece Iterval for a Proportio Estimatig proportio of successes i a biomial experimet

More information

Topic 9: Sampling Distributions of Estimators

Topic 9: Sampling Distributions of Estimators Topic 9: Samplig Distributios of Estimators Course 003, 2018 Page 0 Samplig distributios of estimators Sice our estimators are statistics (particular fuctios of radom variables), their distributio ca be

More information

Chapter 11: Asking and Answering Questions About the Difference of Two Proportions

Chapter 11: Asking and Answering Questions About the Difference of Two Proportions Chapter 11: Askig ad Aswerig Questios About the Differece of Two Proportios These otes reflect material from our text, Statistics, Learig from Data, First Editio, by Roxy Peck, published by CENGAGE Learig,

More information

EXAMINATIONS OF THE ROYAL STATISTICAL SOCIETY

EXAMINATIONS OF THE ROYAL STATISTICAL SOCIETY EXAMINATIONS OF THE ROYAL STATISTICAL SOCIETY HIGHER CERTIFICATE IN STATISTICS, 017 MODULE 4 : Liear models Time allowed: Oe ad a half hours Cadidates should aswer THREE questios. Each questio carries

More information

Stat 139 Homework 7 Solutions, Fall 2015

Stat 139 Homework 7 Solutions, Fall 2015 Stat 139 Homework 7 Solutios, Fall 2015 Problem 1. I class we leared that the classical simple liear regressio model assumes the followig distributio of resposes: Y i = β 0 + β 1 X i + ɛ i, i = 1,...,,

More information

CS284A: Representations and Algorithms in Molecular Biology

CS284A: Representations and Algorithms in Molecular Biology CS284A: Represetatios ad Algorithms i Molecular Biology Scribe Notes o Lectures 3 & 4: Motif Discovery via Eumeratio & Motif Represetatio Usig Positio Weight Matrix Joshua Gervi Based o presetatios by

More information

This is a author produced versio of a paper published i Commuicatios i Statistics - Theory ad Methods. This paper has bee peer-reviewed ad is proof-corrected, but does ot iclude the joural pagiatio. Citatio

More information

Economics Spring 2015

Economics Spring 2015 1 Ecoomics 400 -- Sprig 015 /17/015 pp. 30-38; Ch. 7.1.4-7. New Stata Assigmet ad ew MyStatlab assigmet, both due Feb 4th Midterm Exam Thursday Feb 6th, Chapters 1-7 of Groeber text ad all relevat lectures

More information

Improved Class of Ratio -Cum- Product Estimators of Finite Population Mean in two Phase Sampling

Improved Class of Ratio -Cum- Product Estimators of Finite Population Mean in two Phase Sampling Global Joural of Sciece Frotier Research: F Mathematics ad Decisio Scieces Volume 4 Issue 2 Versio.0 Year 204 Type : Double Blid Peer Reviewed Iteratioal Research Joural Publisher: Global Jourals Ic. (USA

More information

CEE 522 Autumn Uncertainty Concepts for Geotechnical Engineering

CEE 522 Autumn Uncertainty Concepts for Geotechnical Engineering CEE 5 Autum 005 Ucertaity Cocepts for Geotechical Egieerig Basic Termiology Set A set is a collectio of (mutually exclusive) objects or evets. The sample space is the (collectively exhaustive) collectio

More information