TESTS OF HYPOTHESIS I

Size: px
Start display at page:

Download "TESTS OF HYPOTHESIS I"

Transcription

1 UNIT 15 STRUCTURE TESTS OF HYPOTHESIS I Tests of Hypothesis I 15.0 Objectives 15.1 Itroductio 15. Poit Estimatio ad Stadard Errors 15.3 Iterval Estimatio 15.4 Cofidece Limits, Cofidece Iterval ad Cofidece Co-efficiet 15.5 Testig Hypothesis Itroductio 15.6 Theory of Testig Hypothesis Level of Sigificace, Type-I ad Type-II Errors ad Power of a Test 15.7 Two-tailed ad Oe-tailed Tests 15.8 Steps to Follow for Testig Hypothesis 15.9 Tests of Sigificace for Populatio Mea Z-test for variables Tests of Sigificace for Populatio Proportio Z-test for Attributes Let Us Sum Up 15.1 Key Words ad Symbols Used Aswers to Self Assessmet Exercises Termial Questios/Exercises Further Readig 15.0 OBJECTIVES After studyig this uit, you should be able to: l l l l l l estimate populatio characteristics (parameters) o the basis of a sample, get familiar with the criteria of a good estimator, differetiate betwee a poit estimator ad a iterval estimator, comprehed the cocept of statistical hypothesis, perform tests of sigificace of populatio mea ad populatio proportio, ad make decisios o the basis of testig hypothesis INTRODUCTION Let us suppose that we have take a radom sample from a populatio with a view to kowig its characteristics, also kow as its parameters. We are the cofroted with the problem of drawig ifereces about the populatio o the basis of the kow sample draw from it. We may look at two differet scearios. I the first case, the populatio is completely ukow ad we would like to throw some light o its parameters with the help of a radom sample draw from the populatio. Thus, if µ deotes the populatio mea, the we ited to make a guess about it o the basis of a radom sample. This is kow as estimatio. For example, oe may be iterested to kow the average icome of people livig i the city of Delhi or the average life i burig hours of a fluorescet tube light produced by Idia Electrical or proportio of people sufferig from T.B. i city B or the percetage of smokers i tow C ad so o. A somewhat differet situatio may arise whe some iformatio about a parameter is either kow or specified ad we would like to verify whether that iformatio holds good for the sample draw from the populatio as well. 5 5

2 Probability ad Hypothesis Testig This is kow as problem of testig of hypothesis. I the previous examples, we may be iterested of i testig whether the average icome i the city of Delhi is, say, Rs.,000 per moth. I the secod example, we may like to verify whether the claims made by Idia Electrical, that their fluorescet lamps would last 5,000 hours, is justified. Some social workers may believe that 0% of the populatio i city B suffers from T.B. We would like to make our commet after a test of hypothesis. I the last example, some huma activists, cocered about the hazards of passive smokig, assert that 30% of the people stayig i tow C are smokers. We may share their opiio oce we have satisfied ourselves after peformig a statistical test of hypothesis. It may be oted that testig of hypothesis plays a vital role i decisio-makig. I the first example, the statisticia may be cocered about whether to bracket Delhi with the top metropolito cities depedig o the average icome based o his/her recommedatios. If o the basis of a statistical test, it is foud that the claim made by the maufacturer of Idia Electrical is justified, the the sales of his lamps would icrease. I the third example if there is evidece, agai o the basis of testig hypothesis, that the social worker is right about his statemet, suitable steps may be udertake to improve the livig coditios of the margialized sectio i the city so that the percetage of people sufferig from T.B. is reduced. Some strict legislatio baig smokig or reducig smokig to a desirable level may be eacted o the basis of a hypothesis tested i the last example. 15. POINT ESTIMATION AND STANDARD ERRORS Estimatio is a itegral part of our daily lives. I order to costruct a ew house or reovate a old house or flat, we demad a estimate of the cost ivolved. A studet estimates his/her chace of success before appearig for a expesive competitive examiatio. 5 6 Now we shall cosider estimatio from the viewpoit of a statisticia. As we discussed i Uit 4: samplig, is the meas to fid the true value of the parameter which ca be correctly obtaied oly through cesus study. I may cases it is ot practicable due to various costraits. Therefore, the alterative approach is to select some items as a sample from the populatio ad collect the data ad aalyse the data, the estimate the chracteristics of the populatio. This is called estimatio. Poit estimate is oe type of estimate. It is a sigle umber which is used as a estimate of the ukow populatio parameter. Let us assume that we have take a radom sample of observatios, x 1, x, x 3 x, from a populatio characterized by a parameter θ (read theta). This symbol θ is used to deote a parameter that could be mea, mode or some measure of variatio etc. Thus θ may be the mea (µ)of a ormal distributio or the probability of success (P) of a biomial distributio with parameters ad p ad so o. I theory of estimatio, we try to fid a statistic (i.e., a fuctio of sample observatios) T which estimates the ukow parameter θ. Thus the sample mea ( x) x i /, x 1, x, x 3 x beig the icome per moth of persos selected at radom from the city of Delhi, may be cosidered to be the estimate of the average icome per moth (µ) of the people of Delhi. This is deoted by µ ˆ x i.e., the estimate of µ is x.

3 To be more precise, x is kow as a poit estimator of µ as we try to estimate the populatio mea (µ) by a sigle value, amely, the sample mea. O the basis of a radom sample of icomes from Delhi, if it is foud that the sample mea is Rs.,000/-, the oe may coclude that the estimate of average icome per moth of the people livig i that city is Rs.,000/-. Tests of Hypothesis I As opposed to a poit estimate, oe may thik of a iterval estimate that is supposed to cotai the average icome of the people of Delhi per moth. This would be discussed i Sectio At this jucture, we must make a distictio betwee the two terms Estimator ad estimate. T is defied to be a estimator of a parameter θ, if T esimtates θ. Thus T is a statistic ad its value may differ from oe sample to aother sample. I other words, T may be cosidered as a radom variable. The probability distributio of T is kow as samplig distributio of T. As already discussed, the sample mea x is a estimator of populatio mea µ. The value of the estimator, as obtaied o the basis of a give sample, is kow as its estimate. Thus x is a estimator of µ, the average icome of Delhi, ad the value of x i.e., Rs.,000/-, as obtaied from the sample, is the estimate of µ. Selectio of the best estimator: Our ext edeavour would be to discuss differet criteria for selectig the best estimator. Ubiasedess ad Miimum Variace: A statistic T is defied to be ubiased for a parameter θ if expectatio of T is θ, i.e., E(T) θ. O the other had if E(T) θ + a (θ), the the differece a (θ) E(T) θ is kow as bias. The bias is kow to be positive if a (θ) > 0 ad egative if a (θ) < 0. Our first priority would be to select a ubiased estimator of θ. However, there may be may ubiased estimators of θ. If x 1, x, x deote sample observatios from a populatio with a ukow parameter θ, the ay of the observatios or ay liear fuctio of them would be a ubiased estimator of θ. I order to choose the best estimator amog these estimators alog with ubiasedess, we itroduce a secod criterio, kow as, miimum variace. A statistic T is defied to be a miimum variace ubiased estimator (MVUE) of θ if T is ubiased for θ ad T has miimum variace amog all the ubiased estimators of θ. We may ote that sample mea ( x ) is a MVUE for µ. We kow that xi x (15.1) ( x E(x) E i ) 1.[ E (xi)] 1 [ µ ] 1. µ [x 1, x, x are take from populatio havig as µ populatio mea] E(x) µ Hece ( x) is a ubiased estimator of µ. 5 7

4 Probability ad Hypothesis Testig Further, variace of ( x) is give by : v (x) v ( x i / ) (where v deotes variace) 1.[ v(x i)] Sice x i1 s are idepedet 1 [where is populatio variace] 1.[ ] (15.) It ca be proved that v ( x ) has the miimum variace amog all the ubiased estimators of µ. Cosistecy: If T is a estimator of θ, the it is obvious that T should be i the eighbourhood of θ. T is kow to be cosistet for θ, if the differece betwee T ad θ ca be made as small as we please by icreasig the sample size sufficietly. We ca further add that T would be a cosistet estimator of θ if i) E (T) θ ad ii) V (T) 0 for a very large i.e., as For example, sample mea ( x) is a cosistet estimator of µ as E ( x) µ Ad V (x) 0 as. It may be oted that if T is a cosistet estimator of θ, the ay fuctio of T is also a cosistet estimator of θ. 5 8 Efficiecy: A statistic T is called as a efficiet estimator of θ if it has the miimum stadard error amog all the estimators of θ for a fixed sample size. Both the sample mea ad sample media are cosistet estimators for µ. But stadard error (a term, to be defied ad explaied i this sectio) of sample mea is less tha that of sample media. Hece sample media is oly a cosistet estimator of µ, whereas sample mea is both cosistet ad efficiet estimator of µ. Sufficiecy: A statistic T is kow to be a sufficiet estimator of θ if T cotais sufficiet iformatio about θ so that we do ot have to look for ay other estimator of θ. Sample mea ( x ) is a sufficiet estimator of µ. Now let us cosider the followig poit estimates that are commoly used. A) Estimatig Populatio Mea: It is obvious that sample mea is the best estimator of populatio mea µ. It is a MVUE. It is both cosistet ad efficiet estimator for µ. Further more, x is a sufficiet estimator for µ. Thus we estimate the average icome of the people of Delhi by the sample mea or the average life of bulbs, maufactured by Idia Electricals, by the correspodig sample mea. B) Estimatig Populatio Proportio: If a discrete radom variable x follows biomial distributio with parameters ad P, the we have µ E(x) P v(x) P (1 p) [ deotig the umber of trials ad P deotig the probability of a success].

5 Hece, it follows that : Tests of Hypothesis I x i P E(p) E P (15.3) V(xi ) ad V (p) V (x i /) P (1 P) P (1 P) (15.4) Thus if we take a radom sample of size from a populatio where the proportio of populatio possessig a certai characteristic is P ad the sample cotais x uits possessig that characteristic, the a estimate of populatio proportio (P) is give by: x Pˆ (15.5) I other words, the estimate of the populatio proportio is give by the correspodig sample estimate i.e., Pˆ p (15.6) From (15.3) E {p} P So p is a ubiased estimator of P. It ca be show that p has the miimum variace amog all the ubiased estimators of p. I other words, p is a MVUE of P. As v (p) P(1 P) 0 as it follows from Eq. (15.4) that p is a cosistet estimator of P. We ca further establish that p is a efficiet as well as a sufficiet estimator of P. Thus we advocate the use of sample proportio to estimate the populatio proportio as p which satisfies all the desirable properties of a estimator. I order to estimate the proportio of people sufferig from T.B. i city B, if we fid the umber of people sufferig from T.B. is x i a radom sample of size, take from city B, the sample estimate p x/ would provide the estimate of the proportio of people i that city sufferig from T.B. Similarly, the percetage of smokers as foud from a radom sample of people of tow C would provide the estimate of the percetage of smokers i tow C. C) Estimatio of Populatio Variace ad Stadard Error: Stadard error of a statistic T, to be deoted by S.E. (T), may be defied as the stadard deviatio of T as obtaied from the samplig distributio of T. I order to compute the stadard error of sample mea, it may be oted that from Eq. (15.) S.E. (x) for simple radom samplig with replacemet (SRSWR). N η S.E. (x) for simple radom samplig without replacemet N 1 (SRSWOR)]. where is the populatio stadard deviatio (S.D.), is sample size, N is N η populatio size ad the factor is kow as fiite populatio corrector N 1 5 9

6 Probability ad Hypothesis Testig (f.p.c.) or fiite populatio multiplier (f.p.m.) which may be igored for a large populatio. I order to fid S.E., it is ecessary to estimate or i case it is ukow. If x 1, x, x deote sample observatios draw from a populatio with mea µ ad variace, the the sample variace: S (x i x) (15.7) may be cosidered to be a estimator of Sice E(x) µ ad V ( z/ ) E (x µ) We have (15.8) s (x x) (from 15-8) i [(xi µ ) (x µ )] [(x µ ) (x µ ) (x µ ) + (x µ ) ] i i (x µ ) (x µ ) (x µ ) + (x µ i i ) (x µ ) (x µ ). (x µ ) + (x µ i ) [sice Σ (x i µ) Σ x i Σµ x µ (x µ )] (x µ ) (x µ ) + (x µ i ) ( x µ ) + (x µ (15.9) i ) As x i is the ith sample observatio from a populatio with µ as mea ad as variace, it follows that : E (x i µ) Ad E (x µ ) v (x) (15.10) From (15-9), E (s ) E (xi µ ).E (x u) 1 E(S ). ( 1) (15.11) Hece S, the sample variace, is a biased estimator of. As 1 E (S ) E 1 s (15.1) 6 0 thus s 1 ( xi x) 1 (s ) is a ubiased estiator of (15.13)

7 so, we use (s ) (x i x) as a estimator of ad 1 Tests of Hypothesis I (x i x) S as a estimator of 1 A estimate of S.E. ( x) is give by: S S.E. (x) for SRSWR > S N for SRSWOR N 1 P (1 P) From (15.4), it follows that v(p) (15.14) P(1 P) S.E.(p) for SRSWR P (1 P) N. for SRSWOR (15.15) N 1 A estimate of stadard error of sample proportio is give by: p (1 p) S.E. (p) for SRSWR > p (1 p). N for SRSWOR (15.16) Let us cosider the followig illustratios to estimate variace from sample ad also estimate the stadard error. Illustratio 1 A sample of 3 fluorescet lights take from Idia Electricals was tested for the lives of the lights i burig hours. The data are preseted below: Table 15.1: The Lives i Hours of 3 Lights Sl. Life Sl. Life Sl. Life No. (Hours) No. (Hours) No. (Hours) Solutio: We are iterested i estimatig the average life of fluorescet lights maufactured by Idia Electricals. As discussed i this sectio, the estimate of the populatio mea (µ) is give by the correspodig sample mea. The 6 1

8 Probability ad Hypothesis Testig µ ˆ x. If we are further iterested i estimatig the stadard error of x, the we are to compute > S.E.(x) s where, s (xi x) η 1 xi x 1 xi ad x, Sample size We igore f.p.c. as the populatio of lights is very large. Table 15.: Computatio of sample mea ad sample S.D. Life i Hours u i x i 5000 u i x i

9 Tests of Hypothesis I Total From the above Table, u 490, u i 374 i ui 490 u As u x 5000 i u x 5000 i Or, X u (approximately) 4985 (s ) (s x ) s (s ) u i u 1 s hece S.E.(x) so the estimate of the average life of lights as maufactured by Idia Electricals is 4985 hours. Estimate of the populatio variace i (hours) ad the stadard error is hours. Illustratio A sample of 350 people from city C cotaied 70 smokers. Fid a estimate of the proportio of smokers i the city. Also fid a estimate of the stadard error of the proportio of smokers i the sample. Solutio: I this case x o. of smokers i the sample 70, 350. x 70 Thus we have p Hece the estimate of the proportios of smokers i the city is 0. or 0%. Further p (1 p) 0. (1 0.) S.E. (p) > The estimate of the stadard error of the proportio of smokers i the sample is

10 Probability ad Hypothesis Testig Self Assessmet Exercise A 1) State, with reasos, whether the followig statemets are true or false. a) Both the statistic ad parameter are fuctios of sample observatios. b) Ay type of samplig would lead to the same iferece about the populatio. c) Statistical iferece is a statistical process to kow about a populatio from the kowledge of a sample draw from it. d) Ay type of estimator ca be used for estimatig a parameter. e) I most cases, decisio-makig depeds o estimatio. f) There may be more tha oe estimator for a parameter. g) Assumptio of ormality is a must for poit estimatio. h) Every cosistet estimator is ecessarily a efficiet estimator. i) A cosistet estimator approaches the parameter with a icrease i sample size. j) Poit estimator is used as a estimate of the ukow populatio parameter. ) Differetiate betwee estimator ad estimate ) I choosig betwee sample mea ad sample media which oe would you prefer? ) The mothly earigs of 0 families, obtaied from a radom sample from a village i West Begal are give below: Sl. Mothly earigs (Rs.) Sl. Mothly earigs (Rs.) No. No Fid a estimate of the average mothly earigs of the village. Also obtai a estimate of the S.E. of the sample estimate

11 Tests of Hypothesis I 5) I a sample of 900 people, 49 people are foud to be cosumers of tea. Estimate the proportio of cosumers of tea i the populatio. Also fid the correspodig stadard error ) Obtai a ubiased estimate of populatio mea ad populatio variace o the basis of the followig sample observatios: 50, 46, 5, 53, 45, 43, 46, 48, INTERVAL ESTIMATION This is aother type of estimatio. As opposed to estimatig a parameter by a sigle value i.e., poit estimatio discussed i the previous sectio, we may thik of a iterval or a rage of values that is supposed to cotai the parameter. A iterval estimate would always be specified by two values i.e., the lower value ad the upper value, withi which the parameter lies. This is kow as Iterval Estimatio. Thus iterval estimatio may be defied as estimatig a iterval to which the ukow parameter θ may belog, i all likelihood. Regardig the estimatio of the average icome of the people of Delhi city, oe may argue that it would be better to provide a iterval which is likely to cotai the populatio mea. Thus, istead of sayig the estimate of the average icome of Delhi is Rs.,000/-, we may suggest that, i all probability, the estimate of the average icome of Delhi would be from Rs. 1,900/- to Rs.,100/-. I the secod example of estimatig the average life of lights produced by Idia Electricals where the estimate came out to be 4985 hours, the poit estimatio may be a boe of cotetio betwee the producer ad the potetial buyer. The buyer may thik that the average life is rather less tha 4985 hours. A iterval estimatio of the life of lights might satisfy both the parties. Figure 15.1 shows some itervals for θ o the basis of differet samples of the same size from a populatio characterized by a parameter θ. A few itervals do ot cotai θ. 6 5

12 Probability ad Hypothesis Testig θ Fig.15.1: Cofidece Itervals to θ 15.4 CONFIDENCE LIMITS, CONFIDENCE INTERVAL AND CONFIDENCE CO-EFFICIENT Let us assume that we have take a radom sample of size from a populatio characterized by a parameter θ. Let us further suppose that based o these sample observatios, it is possible to fid two statistics t 1 ad t such that: P (t 1 < θ) α 1 Ad P (t > θ) α Where α 1 ad α are two small positive umbers. Combiig these two coditios, we may write: P (t 1 θ t ) 1 α (15.17) Where α α 1 + α Equatio (15.17) could be iterpreted as the probability that θ lies betwee t 1 ad t is (1 α), whatever may be the value of θ, satisfyig (15.17). The iterval [t 1, t ], t 1 beig less tha t, that cotais the parameter θ is kow Cofidece Iterval to θ, t 1 beig kow as Lower Cofidece limit ad t as Upper Cofidece Limits. (1 α) is kow to be Cofidece Co-efficiet correspodig to the cofidece iterval [t 1, t ]. Oe may like to kow why the term cofidece comes ito the picture. If we choose α 1 ad α such a way that α 0.01, the the probability that θ would belog to the radom iterval [t 1, t ] is I other words, oe feels 99% cofidet that [t 1, t ] would cotai the ukow parameter θ. Similarly if we select α 0.05, the P [t 1 θ t ] 0.95, thereby implyig that we are 95% cofidet that θ lies betwee t 1 ad t. (15.17) suggests that as α decreases, (1 α) icreases ad the probability that the cofidece iterval [t 1, t ] would iclude the parameter θ also icreases. Hece our edeavour would be to reduce α ad thereby icrease the cofidece co-efficiet (1 α). 6 6

13 Referrig to the estimatio of the average life of lights (θ), if we observe that θ lies betwee 4935 hours ad 5035 hours with probability 0.98, the it would imply that if repeated samples of a fixed size (say 3) are take from the populatio of lights, as maufactured by Idia Electricals, the i 98 per cet of cases, the iterval [4935 hours, 5035 hours] would cotai θ, the average life of lights i the populatio while i per cet of cases, the iterval would ot cotai θ. I this case, the cofidece iterval for θ is [4935 hours, 5035 hours]. Lower Cofidece Limit of θ is 4935 hours, Upper Cofidece Limit of θ is 5035 hours, ad the Cofidece Co-efficiet is 98 per cet. Tests of Hypothesis I Selectio of Cofidece Iterval Our ext task would be to select the basis for estimatig cofidece iterval. Let us assume that we have take a radom sample of size from a ormal populatio characterized by the two parameters µ ad, the populatio mea ad stadard deviatio respectively. Thus, i the case of estimatig a Cofidece Iterval for average icome of people dwellig i Delhi city, we assume that the distributio of icome is ormal ad we have take a radom sample from the city. I aother example cocerig average life of fluorescet lights as produced by Idia Electricals, we assume that the life of a fluorescet light is ormally distributed ad we have take a radom sample from the populatio of fluorescet lights maufactured by Idia Electricals. Figure 15. shows percetage of area uder Normal Curve. It ca be show that if a radom sample of size is draw from a ormal populatio with mea µ ad variace, the ( x), the sample mea also follows ormal distributio with µ as mea ad / as variace. Further as we have observed i Sectio 15.. S.E. (x) From the properties of ormal distributio, it follows that the iterval : [ µ S.E. (x), µ + S.E. (x)] covers 68.7% area. The iterval [ µ S.E. (x), µ + S.E.(x)] covers 95.45% area ad the iterval [ µ 3S.E. (x), µ + 3S.E. (x)] covers 99.73% area. Figure 15. depicts this iformatio % % % 13.59%.14% µ.14% % % 99.73% Fig. 15.: Percetages of Area uder a Normal Curve 6 7

14 Probability ad Hypothesis Testig Now let us cosider a situatio where the assumptio of ormality may ot hold. If the sample size is large eough, the the sample mea x follows approximately, i.e., asymptotically ormal distributio with mea as µ ad stadard error as /, µ ad beig the mea ad S.D. of the populatio uder cosideratio. I case is ukow, we ca replace it by the correspodig sample stadard deviatio. Oe may ask the questio as to how large should be. It is rather difficult to specify a exact value of so that the distributio of x would be asymptotically ormal. Larger the value of, the better. However for practical purposes, if exceeds 30, the we may assume that x is asymptotically ormal. Our ext questio may be what would be the cofidece iterval for µ. Will it be µ ± S.E. (x), or µ ± S.E. (x), or µ ± 3 S.E. (x), or some other iterval? Suppose that the Cofidece Iterval to µ is give by : µ ± u S.E. (x) ad we are to determie u such that : P [x u S.E. (x) µ x + u S.E. (x)] 1 α (15.18) Or, x µ P[ u u] 1 α S.E. (x) Or, P [ u Z u] 1 α [where x µ Z is a stadard ormal variable] S.E. (x) Or. Or, Or. φ (u) φ ( u) 1 α [where φ (K) P(Z K), area uder stadard ormal curve from to K]. φ (u) [ 1 φ (u) 1 α φ (u) α Or. φ ( u) 1 ( α / ) (15.19) Puttig α 0.10 i (15.19), we get φ ( u) or, φ ( u) φ ( ) or. u Thus 100 (1 α) % or 100 (1 0.1)% or 90% cofidece iterval to populatio mea µ is : Give by X 1.645, X Puttig α 0.05, 0.0 ad 0.01 respectively i (15.19) ad proceedig i a similar maer, we get 95% Cofidece Iterval to µ 1.96 x 1.96, x + (15.0) % Cofidece Iterval to µ x.33, x +.33 (15.1)

15 ad 99% Cofidece Iterval to µ x.58, x +.58 (15.) Theoretically we may take ay Cofidece iterval by choosig u accordigly. However i a majority of cases, we prefer 95% or 99% Cofidece Iterval. These are show i Figure 15.3 ad Figure 15.4 below. Tests of Hypothesis I.5% of area uder curve 95% of area uder curve.5% of area uder curve x 1.96 x x Fig. 15.3: 95% Cofidece Iterval for Populatio Mea 0.5% of area uder curve 99% of area uder curve 0.5% of area uder curve x.58 x x +.58 Fig. 15.4: 99% Cofidece Iterval for Populatio Mea Next we cosider Iterval Estimatio i the followig cases: Iterval Estimatio of Populatio Mea As suggested i this sectio uder assumptio of ormality, 95% cofidece iterval to µ, the populatio mea, is give by x 1.96, x If the assumptio of ormality does ot hold but is greater tha 30, the above 95% cofidece iterval still may be used for estimatig populatio mea. I case is ukow, it may be replaced by the correspodig ubiased estimate of, amely S, so log as exceeds 30. However, we may face a difficult situatio i case is ukow ad does ot exceed 30. This problem has bee discussed i the ext uit (Uit-16). Similarly, 99% cofidece iterval to µ is give by : x.58, x +.58 I case is ukow. The 99% cofidece iterval to µ is : S S x.58, x +.58 i case is ukow ad > 30. (15.3) 6 9

16 Probability ad Hypothesis Testig Iterval Estimatio of Ukow Populatio Proportio It ca be assumed that whe is large ad either p or (1 p) is small (oe may specify p 5 ad (1 p) 5), the the sample proportio p is P (1 P) asymptotically ormal with mea as P ad S.E.(p), P beig the ukow populatio proportio i which we are iterested. The estimate of S.E. (p) is give by : p (1 p) S.E.(pˆ ) Hece, 95% cofidece iterval to p is give by : > p ( 1 p) p ( 1 p) p 1. 96, p (15.4) ad 99% cofidece iterval to P is : p.58 p(1 p), p +.58 p(1 p) (15.5) Let us cosider the followig illustratios to uderstad the procedure for iterval estimatio. Illustratio 3 I a radom sample of 1,000 families from the city of Delhi, it was foud that the average of icome as obtaied from the sample is Rs.,000/-, it is further kow that populatio S.D. is Rs. 58. Fid 95% as well as 99% cofidece iterval to populatio mea. Solutio: Let x deote icome of the people of Delhi city. If µ deotes average icome of people dwellig i Delhi, the 95% cofidece iterval to µ is: x 1.96, x ad 99% cofidece iterval to µ is : x.58, x.58 Where x Sample mea; Sample size, ad Populatio stadard deviatio. I our case, x Rs. 000, 1000, Rs x 1.96 Rs Rs x Rs Rs

17 58 x.58 Rs Rs Tests of Hypothesis I ad 58 x +.58 Rs Rs Hece we have 95% cofidece iterval to average icome for the people of Delhi [Rs to Rs ] ad 99% cofidece iterval to average icome for the people of Delhi [Rs to Rs. 01]. Illustratio 4 Calculate the 95% ad 99% cofidece limits to the average life of fluorescet lights produced by Idia Electricals. Solutio: Sice, the populatio stadard deviatio is ukow ad 3 (> 30), we replace by S, the sample S.D. with divisor as ( 1) i our previous example ad get 95% cofidece iterval to µ is: x 1.96 S, x S Similarly, 99% cofidece iterval for µ x.58 S, x +.58 S Where, x Sample mea 4985 hours, Sample size 3; ad S Sample S.D. with ( 1) divisio hours (as computed earlier). x 1.96 S hours 3 Illustratio 5 S x hours 3 S x hours 3 S x hours 3 95% Cofidece Iterval to the average life of lights [ hours, hours]. 99% Cofidece Iterval to the average life of lights [ hours, hours]. While iterviewig 350 people i a city, the umber of smokers was foud to 7 1

18 Probability ad Hypothesis Testig be 70. Obtai 99% lower cofidece limit ad the correspodig upper cofidece limit to the proportio of smokers i the city. Solutio: As discussed i the previous sectio, 99% Lower Cofidece Limit to P, the proportio of smokers i the city is give by: p.58 p (1 p) ad 99% Upper Cofidece Limt to P is: p +.58 p (1 p) provided p 5 ad p (1 p) 5. I this case x o. of smokers 70 o. of people iterviewed 350 x 70 p As p ad (1 p) are rather large, we ca apply the formula for 99% Cofidece Limit as metioed already. 99% Lower Cofidece Limit to P is : 0. (1 0.) % Upper Cofidece Limit to P is : 0. (1 0.) Hece 99% Lower Cofidece Limit ad 99% Upper Cofidece Limit for the proportio of smokers i the city are ad 0.14 respectively. Illustratio 6 I a radom sample of people from a tow, 358 people were foud to be sufferig from T.B. With 95% Cofidece as well as 98% Cofidece, fid the limits betwee which the percetage of the populatio of the tow sufferig from T.B. lies. Solutio: Let x be the umber of people sufferig from T.B. i the sample ad as the umber of people who were examied. The the proportio of people sufferig from T.B. i the sample is give by: x 358 p As p x 358 ad (1 p) p x are both very large umbers, we ca apply the formula for fidig Cofidece Iterval as metioed i the previous sectio. Thus 95% Cofidece Iterval to

19 P, the proportio of the populatio of the tow sufferig from T.B., is give by : Tests of Hypothesis I p 1.96 p (1 p), p p (1 p) ( ), ( ) [ , ] [0.1181, 0.17] I a similar way, 98% Cofidece Iterval to P is give by: p.33 p (1 p), p +.33 p (1 p) ( ), ( ) [0.1150, 0.158] Thereby, we ca say with 95% cofidece that the percetrage of populatio i the tow sufferig from T.B. lies betwee ad 1.7 ad with 98% cofidece that the percetage of populatio sufferig from T.B. lies betwee ad Illustratio 7 A famous shoe compay produces 80,000 pairs of shoes daily. From a sample of 800 pairs, 3% are foud to be of poor quality. Fid the limits for the umber of substadard pair of shoes that ca be expected whe the Cofidece Level is Solutio: Let p be the sample proportio of defective shoes as produced by the shoe compay. I this case sample size () is 800 ad populatio size (N) is 80,000. Sice the populatio is very large, we do ot apply fiite populatio correctio. p 3% 0.03 > p (1 p) 0.03 (1 0.03) S.E. (pˆ ) Thus 99% Lower Cofidece Limit to P, the proportio of defective shoes i the daily productio of the shoe compay is : > p.58 S.E. (pˆ ) similarly 99% Upper Cofidece Limit to p is : P +.58 Ŝ.E. (Pˆ ) Hece, the Lower limit to the umber of substadard i.e., defective pairs of shoes at 99% Level of Cofidece N , (approximately)

20 Probability ad Hypothesis Testig The Upper Limit to the umber of substadard, pairs of shoes at 99% Level of Cofidece is 80, (approximately) 3638 Self Assessmet Exercise B 1) State with reasos, whether the followig statemets are true or false. a) Cofidece Iterval provides a rage of values that may ot cotai the parameter. b) Cofidece Iterval is a fuctio of Cofidece Co-efficiet. c) 95% Cofidece Iterval for populatio mea is x ± 1.96 S.E. (x). d) While computig Cofidece Iterval for populatio mea, if the populatio S.D. is ukow, we ca always replace it by the correspodig sample S.D. p (1 p) e) 99% Upper Cofidece Limit for populatio proportio is p f) Cofidece co-efficiet does ot cotai Lower Cofidece Limit ad Upper Cofidece Limit. g) If p 5 ad p (1 p) 5, oe may apply the formula p ± z α p (1 p) for computig Cofidece Iterval for populatio proportio. h) The iterval µ ± 3 S.E. (x) covers 96% area of the ormal curve. ) Differetiate betwee Poit Estimatio ad Iterval Estimatio ) Distiguish betwee Cofidece Limit ad Cofidece Iterval Out of 5,000 customer s ledger accouts, a sample of 800 accouts was take to test the accuracy of postig ad balacig ad 50 mistakes were foud. Assig limits withi which the umber of wrog postigs ca be expected with 99% cofidece A sample of 0 items is draw at radom from a ormal populatio comprisig 00 items ad havig stadard deviatio as 10. If the sample mea is 40, obtai 95% Iterval Estimate of the populatio mea

21 6) A ew variety of potato grow o 400 plots provided a mea yield of 980 quitals per acre with a S.D. of quitals per acre. Fid 99% Cofidece Limits for the mea yield i the populatio Tests of Hypothesis I 15.5 TESTING HYPOTHESIS INTRODUCTION Referrig to the problem of the status to be give to Delhi City, oe of the criteria for determiig the status would be the average icome of the people of Delhi. Let us suppose that if µ, the average icome of the people is Rs. 3,000 per moth, the Delhi would belog to the group of top cities. I order to estimate µ, we take a radom sample of people livig i that city ad compute x, the sample mea. If x is i the eighbourhood of Rs. 3,000, the we have o hesitatio i declarig the status of Delhi as oe belogig to the top grade. But the most importat questio would be as to what differece betwee the sample mea ad Rs (populatio mea) ca be accepted as the differece due to oly samplig fluctuatios. I order to aswer this questio, let us familiarise ourselves with a few terms associated with the problem. A statemet like The average icome of the people belogig to the city of Delhi is Rs. 3,000 per moth is kow as a ull hypothesis. Thus, a ull hypothesis may be described as a assumptio or a statemet regardig a parameter (populatio mea, µ, i this case) or about the form of a populatio. The term ull is used as we test the hypothesis o the assumptio that there is o differece or, to be more precise, o sigificat differece betwee the value of a parameter ad that of a estimator as obtaied from a radom sample take from the populatio. A hypothesis may be simple or composite. A simple hypothesis is oe that specifies the populatio distributio completely. Thus testig µ 3,000 is a simple hypothesis if the populatio stadard deviatio () is kow. A composite hypothesis is oe that does ot specify the populatio completely. Testig µ 3,000 whe is ukow is a composite hypothesis as it does ot specify the populatio completely. A ull hypothesis is deoted by H 0. Thus we may write : H 0 : µ 3,000 i.e., the ull hypothesis is that the populatio mea is Rs. 3,000. Geerally, we write H 0 : µ µ 0 i.e., the ull hypothesis is that the populatio mea is µ, whereas µ 0 may be ay value as specified i a give situatio. Obviously a ull hypothesis (H 0 ) is to be tested agaist a appropriate alterative hypothesis (H 1 ). Ay hypothesis that cotradicts a ull hypothesis is kow as a alterative hypothesis. If the ull hypothesis is rejected, the alterative hypothesis is accepted. Procedures eablig us to 7 5

22 Probability ad Hypothesis Testig decide whether to accept or reject a hypothesis is kow as test of hypothesis or test of sigificace or decisio rule. Thus, the etire process of hypothesis testig is either to reject or accept H 0 oly. I the preset problem, oe may argue that sice may people of Delhi city are livig i the slums ad eve o the pavemets, the average icome should be less tha Rs So oe alterative hypothesis may be : H 1 : µ < 3,000 i.e., the average icome is less tha Rs. 3,000 or symbolically: or, H 1 : µ < µ 0 i.e., the populatio mea (µ) is less tha µ 0. Agai oe may feel that sice there are may multistoried buildigs ad may ew models of vehicles ru through the streets of the city, the average icome must be more tha Rs. 3,000. So aother alterative hypothesis may be : H : µ > 3000 i.e., the average icome is more tha Rs. 3,000. or, H : µ > µ 0 i.e., the populatio mea is more tha µ o. Lastly, aother group of people may opie that the average icome is sigificatly differet from µ 0. So the third alterative could be : H : µ 3000 i.e., the average icome is aythig but Rs. 3,000. or, H : µ µ 0 i.e., the populatio mea is ot µ THEORY OF TESTING HYPOTHESIS LEVEL OF SIGNIFICANCE, TYPE-I AND TYPE-II ERRORS AND POWER OF A TEST I order to take a decisio about acceptace or rejectio of a ull hypothesis, let us cosider the theory ivolvig testig of hypothesis. Suppose that we have a radom sample of size take from a populatio characterized by a ukow parameter θ. We deote the sample observatios by x (x 1, x, x 3, x ) ad we would like to test H 0 : θ θ 0 agaist H 1 : θ θ If, the x (x 1, x ) ca be represeted as a poit i the -dimesioal plae takig, say x 1, o the horizotal axis ad x o the vertical axis. I a similar way, it is possible to coceive x (x 1, x, x 3, x ) as a poit i the - dimesioal plae. Cosider all the possible samples of a fixed size, i.e., N C i case of SRSWOR ad N i the case of SRSWR, N deotig the populatio size. Next we cosider the sample space formed by all these poits ad let it be deoted by Ω. We divide Ω ito two parts ω ad A Ω ω, the boudary of ω is take withi A. We frame a simple rule which says that if the sample poit x falls o ω, we reject H 0 ad if x falls o A, we accept H 0. ω is kow as the critical regio or rejectio regio ad A, as the acceptace regio. At this jucture, let us make oe poit clear. Acceptace of H 0 does ot mea that H 0 is always true. It just reflects the idea that o the basis of the give data, there is ot eough evidece to support the validity of H 1. I a similar maer rejectio of H 0 idicates the ull hypothesis does ot hold good i the light of the give sample observatios.

23 Type-I ad Type-II Errors Tests of Hypothesis I Now while testig H 0 we are liable to commit two types of errors. I the first case, it may be that H 0 is true but x falls o ω ad as such, we reject H 0. This is kow as type-i error or error of the first kid. Thus type-i error is committed i rejectig a ull hypothesis which is, i fact, true. Secodly, it may be that H 0 is false but x falls o A ad hece we accept H 0. This is kow as type-ii error or error of the secod kid. So type-ii error may be described as the error committed i acceptig a ull hypothesis which is, i fact, false. The two kids of errors are show i Table Table 15.3: Types of Errors i Testig Hypothesis Real Situatio Statistical decisio based o sample H 0 Accepted H 0 Rejected H 0 True Right decisio Type-I error H 0 False Type-II error Right decisio It is obvious that we should take ito accout both types of errors ad must try to reduce them.sice committig these two types of errors may be regarded as radom evets, we may modify our earlier statemet ad suggest that a appropriate test of hypothesis should aim at reducig the probabilities of both types of errors. Let α (read as alpha ) deote the probability of type-i error ad β (read as beta ) the probability of type-ii error. thus by defiitio, we have α The probability of the sample poit fallig o the critical regio whe H 0 is true i.e., the value of θ is θ 0 P (x ω θ 0 ) (15.6) ad β The probability of the sample poit fallig o the critical regio whe H 1 is true, i.e., the value of θ is θ 1 P (x A θ 1 ) (15.7) Surely, our objective would be to reduce both type-i ad type-ii errors. But sice we have take recourse to samplig, it is ot possible to reduce both types of errors simultaeously for a fixed sample size. As we try to reduce α, β icreases ad a reductio i the value of β results i a icrease i the value of α. Thus, we fix α, the probability of type-i error to a give level (say, 5 per cet or 1 per cet) ad subject to that fixatio, we try to reduce β, probability of type-ii error. α is also kow as size of the critical regio. It is further kow as level of sigificace as α costitutes the basis for makig the differece (θ θ 0 ) as sigificat. The selectio of α level of sigificace, depeds o the experimeter. Power of a Test: By defiitio, we have β P (x A θ θ 1 ) [from 15.7] : 1 β 1 P (x A θ θ 1 ) P (x ω θ θ 1 ) [from 15.6] [Sice θ 1 may fall either o ω or A, therefore, P (x ω θ θ 1 ) + P (x A θ θ 1 ) 1 ad we have 1 P (x A θ θ 1 ) P (x ω θ θ 1 )] 7 7

24 Probability ad Hypothesis Testig Now P (x ω θ θ 1 ) is the probability of rejectig H 0 whe H 0 is false ad the alterative hypothesis H 1 is true which should be the desirable property of a appropriate test. It is obvious that a low value of β would esure a high value of (1 β). Hece we try to miimize β, the probability of type-ii error, as the miimizatio of β esures the maximizatio of (1 β). The expressio (1 β) serves as a idicator of the validity of the test as a very high value of (1 β) idicates that the test is doig fie i its edeavour to reject a false hypothesis. Hece (1 β) is kow as power of the test as it tells us how well the test uder cosideratio is performig whe the ull hypothesis is ot true. It is obvious that we should try to make our test as powerful as possible subject to a fixed value of α. Oe may regard power of a test as a fuctio of θ. The fuctio P (θ) 1 β (θ) is kow as the power fuctio of the test. The curve obtaied by plottig P (θ) agaist θ is kow as power curve. Look at the followig figure 15.5 which exhibits a power curve. 1.0 P(θ) 0 θ Fig. 15.5: Power Curve of a Test 15.7 TWO-TAILED AND ONE-TAILED TESTS I order to test the ull hypothesis H 0 : θ θ 0 agaist a plausible alterative hypothesis, let us suppose that we fid a statistic T which is a sufficiet estimator of θ. We assume further that, based o a radom sample take from the populatio characterized by a ukow parameter θ, it is possible to fid a fuctio of T ad θ ad let u u (T, θ) by such a fuctio. T is kow as test statistic for testig H 0 : θ θ 0. Lastly let us assume that whe θ θ 0, u 0 u (T, θ 0 ) i.e., u 0 is the value of u uder H 0 (i.e., assumig the ull hypothesis to be true). Based o the samplig distributio of the test statistic u uder H 0, it may be possible to fid 4 values of u, amely, u α/, u (1-α/), u α ad u (1-α ) for a fixed level of sigificace α, such that : α P (u 0 u α/ ) (15.8) P (u 0 u (1 α/ ) α (15.9) P (u 0 u α ) α (15.30) P (u 0 u (1 α) ) α (15.31) u α may be described as the upper α-poit of the distributio of u ad u (1-α) as the correspodig lower α-poit. Two-tailed test: Addig (15.8) ad (15.9), we get : P (u 0 u α/ ) + P (u 0 u (1-α/) ) α (15.3) 7 8 i.e., the probability that u 0 would exceed u α/ or u 0 is less tha u (1-α/) is α.

25 I order to test H 0 : θ θ 0 agaist H 1 : θ θ 0, if we select a low value of α, say α 0.01, the (15.3) suggests that the probability u 0 is greater tha u a/ or u 0 is less tha u (1-α/) is 0.01 which is pretty low. So o the basis of a radom sample draw from the populatio, if it is foud that u 0 is greater tha u a/ or u 0 is less tha u (1 α/), the we have rather strog evidece that H 0 is ot true. The we reject H 0 : θ θ 0 ad accept the alterative hypothesis H 1 : θ θ 0. As show i the followig Figure 15.6, here the critical regio lies o both tails of the probability distributio of u. Tests of Hypothesis I Critical Regio ω : u 0 u (1-α/) 100 (1 α) % area Acceptace Regio Critical Regio ω : u 0 u α/ 50 α % area u (1- α /) u α / Fig. 15.6: Critical regio of a two-tailed Test 50 α % area If the sample poit x falls o oe of the two tails, we reject H 0 ad accept H 1 : θ θ 0. The statistical test for H 0 : θ θ 0 agaist H 1 : θ θ 0 is kow as both-sided test or two-tailed test as the critical regio, ω lies o both sides of the probability curve, i.e., o the two tails of the curve. The critical regio is ω : u 0 u α/ ad ω : u 0 u (1-α/). It is obvious that a two-tailed test is appropriate whe there are reasos to believe that u differs from θ 0 sigificatly o both the left side ad the right side, i.e., the value of the test statistic u as obtaied from the sample is sigificatly either greater tha or less tha the hypothetical value. For testig the ull hypothesis H 0 : µ 3000, i.e., the average icome of the people of Delhi city is Rs. 3000, oe may thik that the alterative hypothesis would be H 1 : µ 3000 i.e., the average icome is ot Rs ad as such, we may advocate the applicatio of a two-tailed test. Similarly, for testig the ull hypothesis that the average life of lights produced by Idia Electricals is 5,000 hours agaist the alterative hypothesis that the average life is ot 5,000 hours, i.e., for testig H 0 : µ 5,000 agaist H 1 : µ 5,000, we may prescribe a two-tailed test. I the problem cocerig the health of city B, we may be iterested i testig whether 0% of the populatio of city B really suffers from T.B. i.e., testig H 0 : P 0. agaist H 1 : P 0. ad agai a two-tailed test is ecessary ad lastly regardig the harms of smokig, we may like to test H 0 : P 0.3 agaist H 1 : P 0.3. Right-tailed Tests We may thik of testig a ull hypothesis agaist aother pair of alteratives. If we wish to test H 0 : θ θ 0 agaist H 1 : θ > θ 0, the from (15.30) we have P (u 0 u α ) α. This suggests that a low value of α, say α 0.01, implies that the probability that u 0 exceeds u α is So the probability that u 0 exceeds u α is rather small. Thus o the basis of a radom sample draw from this populatio if it is foud that u 0 is greater tha u α, the we have eough evidece to suggest that H 0 is ot true. The we reject H 0 ad accept H 1. This is exhibited i Figure 15.7 as show below: 7 9

26 Probability ad Hypothesis Testig 100 (1 α) % area Acceptace Regio Critical Regio ω : u 0 u α Fig. 15.7: Critical regio of a right-tailed Test u α 100 α % area As show i figure 15.7, the critical regio lies o the right tail of the curve. This is a oe-sided test ad as the critical regio lies o the right tail of the curve, it is kow as right-tailed test or upper-tailed test. We apply a righttailed test whe there is evidece to suggest that the value of the statistic u is sigificatly greater tha the hypothetical value θ 0. I case of testig about the average icome of the citizes of Delhi, if oe has prior iformatio to suggest that the average icome of Delhi is more tha Rs. 3,000, the we would like to test H 0 : µ 3,000 agaist H 1 : µ > 3,000 ad we select the right-tailed test. I a similar maer for testig the hypothesis that the average life of lights by Idia Electricals is more tha 5,000 hours or for testig the hypothesis that more tha 0 per cet suffer from T.B. i city B or for testig the hypothesis that the per cet of smokers i tow C is more tha 30, we apply the righttailed test. Left-tailed test Lastly, we may be iterested to test H 0 : θ θ 0 agaist H : θ < θ 0.From (15.31), we have P (u 0 u 1 α ) α. Choosig α 0.01, this implies that the probability that u 0 would be less tha u α is 0.01, which is surely very low. So, if o the basis of a radom sample take from the populatio, it is foud that u 0 is less tha u 1-α, the we have very serious doubts about the validity of H 0. I this case, we reject H 0 ad accept H : θ < θ 0. This is reflected i Figure 15.8 show below. Critical Regio ω : u 0 u (1-α) 100 (1 α) % area Acceptace Regio 100 α % area u (1- α ) Fig. 15.8: Critical Regio of a Left-tailed Test 8 0 The test for H 0 : θ θ 0 agaist H : θ < θ 0 is aother oe-sided test ad as the critical regio lies o the left tail of the curve, this is kow as a left-

27 tailed test or a lower-tailed test. We apply a left-tailed test whe there is eough idicatio to suggest that the value of the test statistic u is sigificatly less tha the hypothetical value. The for determiig the status of Delhi city, if somebody suggests with evidece that the average icome is less tha Rs. 3,000 ad as such Delhi should ot be regarded as a top grade city, the we are to test H 0 : µ 3000 agaist H 1 : µ < 3000, which is a left-tailed test. We may further ote that we apply left-tailed test whe we would like to test the hypothesis that the average life of lights of Idia Electricals is less tha 5,000 hours or less tha 0 per cet are sufferig from T.B. i city B or less tha 30 per cet are smokers i tow C. Tests of Hypothesis I 15.8 STEPS TO FOLLOW FOR TESTING HYPOTHESIS While testig hypothesis, oe must go through the followig steps. 1) Set up the ull hypothesis H 0 : θ θ 0 ad oe of the alterative hypothesis H : θ θ 0 or H 1 : θ > θ 0 or H : θ < θ 0 depedig upo the problem. Selectig the proper alterative plays a sigificat role i decisio makig i coectio with testig of hypothesis. ) Choose the appropriate test statistic u ad samplig distributio of u uder H 0. I most cases u follows a stadard ormal distributio uder H 0 ad hece Z-test ca be recommeded i such a case. 3) Select α, the level of sigificace of the test if it is ot provided i the give problem. I most cases, we choose α 0.05 ad α 0.01 which are kow as 5% level of sigificace ad 1% level of sigificace. 4) Defie critical regio ω, based o the alterative hypothesis. For testig H 0 : θ θ 0 agaist both-sided alterative H 1 : θ θ 0, the critical regio is give by ω : u 0 u α/ ad ω : u 0 u (1-α/). Similarly, the critical regio for the rightsided alterative is give by ω : u 0 u α ad the critical regio for the left-sided is give by ω : u 0 u 1-α. 5) Obtai the value of u 0 o the basis of the give sample observatios. 6) Reject H 0 if u 0 falls o ω. Otherwise accept H 0. 7) Draw your ow coclusio i very simple laguage which should be uderstood eve by a layma TESTS OF SIGNIFICANCE FOR POPULATION MEAN Z-TEST FOR VARIABLES Let us assume that we have take a radom sample of size from a ormal populatio with mea as µ ad stadard deviatio as. Let the sample observatios be deoted by x 1, x, x 3, x. While testig for the ukow populatio mea µ, we are to cosider the followig cases. Case 1: Whe the stadard deviatio is kow. We wat to test H 0 : µ µ 0 agaist oe of the followig alterative hypothesis. H : µ µ 0 or, H 1 : µ > µ 0 or, H : µ < µ

28 Probability ad Hypothesis Testig As we have discussed i Sectio 15., the best statistic for the parameter µ is x. It has bee proved i that Sectio, E (x) µ. S.E. (x) As such the test statistic : x E(x) z S.E.(x) x µ / is a stadard ormal variable. Uder H 0, i.e., assumig the ull hypothesis to be true, (x µ 0) z0 is a stadard ormal variable. As such, the test is kow as stadard ormal variate test or stadard ormal deviate test or Z-test. I order to fid the critical regio for testig H 0 agaist H from (15.8) ad (15.9), we fid that : ad P (u P (u 0 0 u u ( α /, (1 α /, α ) α ) If we deote the stadard ormal variate by Z, ad the upper α-poit of the stadard ormal distributio by Z α, ad by Z (1 α/) Z α/, (as the stadard ormal distributio is symmetrical about 0), the lower α-poit of the stadard ormal distributio, the the above two equatios are reduced to : α P (Z0 Z α / ) (15.33) α ad P (Z0 Zα / ) (15.34) From (15.33), we have: 1 α P (Z0 < Z( α / ) or or α 1 φ (Zα / ) α φ ( Zα/ ) 1 choosig α.05, φ (Z ) or, φ Z ) (1.96 ) [from Sectio 15.4] ( φ Thus, Z Hece from (15.33) ad (15.34), we have P (Z ) 0.05 ad P (Z )

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

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

1036: Probability & Statistics

1036: Probability & Statistics 036: Probability & Statistics Lecture 0 Oe- ad Two-Sample Tests of Hypotheses 0- Statistical Hypotheses Decisio based o experimetal evidece whether Coffee drikig icreases the risk of cacer i humas. A perso

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

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

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

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

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

Data Analysis and Statistical Methods Statistics 651

Data Analysis and Statistical Methods Statistics 651 Data Aalysis ad Statistical Methods Statistics 651 http://www.stat.tamu.edu/~suhasii/teachig.html Suhasii Subba Rao Review of testig: Example The admistrator of a ursig home wats to do a time ad motio

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

MBACATÓLICA. Quantitative Methods. Faculdade de Ciências Económicas e Empresariais UNIVERSIDADE CATÓLICA PORTUGUESA 9. SAMPLING DISTRIBUTIONS

MBACATÓLICA. Quantitative Methods. Faculdade de Ciências Económicas e Empresariais UNIVERSIDADE CATÓLICA PORTUGUESA 9. SAMPLING DISTRIBUTIONS MBACATÓLICA Quatitative Methods Miguel Gouveia Mauel Leite Moteiro Faculdade de Ciêcias Ecoómicas e Empresariais UNIVERSIDADE CATÓLICA PORTUGUESA 9. SAMPLING DISTRIBUTIONS MBACatólica 006/07 Métodos Quatitativos

More information

MATH 320: Probability and Statistics 9. Estimation and Testing of Parameters. Readings: Pruim, Chapter 4

MATH 320: Probability and Statistics 9. Estimation and Testing of Parameters. Readings: Pruim, Chapter 4 MATH 30: Probability ad Statistics 9. Estimatio ad Testig of Parameters Estimatio ad Testig of Parameters We have bee dealig situatios i which we have full kowledge of the distributio of a radom variable.

More information

Chapter 8: Estimating with Confidence

Chapter 8: Estimating with Confidence Chapter 8: Estimatig with Cofidece Sectio 8.2 The Practice of Statistics, 4 th editio For AP* STARNES, YATES, MOORE Chapter 8 Estimatig with Cofidece 8.1 Cofidece Itervals: The Basics 8.2 8.3 Estimatig

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

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

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

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

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

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

Estimation of a population proportion March 23,

Estimation of a population proportion March 23, 1 Social Studies 201 Notes for March 23, 2005 Estimatio of a populatio proportio Sectio 8.5, p. 521. For the most part, we have dealt with meas ad stadard deviatios this semester. This sectio of the otes

More information

Recall the study where we estimated the difference between mean systolic blood pressure levels of users of oral contraceptives and non-users, x - y.

Recall the study where we estimated the difference between mean systolic blood pressure levels of users of oral contraceptives and non-users, x - y. Testig Statistical Hypotheses Recall the study where we estimated the differece betwee mea systolic blood pressure levels of users of oral cotraceptives ad o-users, x - y. Such studies are sometimes viewed

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

Section 9.2. Tests About a Population Proportion 12/17/2014. Carrying Out a Significance Test H A N T. Parameters & Hypothesis

Section 9.2. Tests About a Population Proportion 12/17/2014. Carrying Out a Significance Test H A N T. Parameters & Hypothesis Sectio 9.2 Tests About a Populatio Proportio P H A N T O M S Parameters Hypothesis Assess Coditios Name the Test Test Statistic (Calculate) Obtai P value Make a decisio State coclusio Sectio 9.2 Tests

More information

Final Examination Solutions 17/6/2010

Final Examination Solutions 17/6/2010 The Islamic Uiversity of Gaza Faculty of Commerce epartmet of Ecoomics ad Political Scieces A Itroductio to Statistics Course (ECOE 30) Sprig Semester 009-00 Fial Eamiatio Solutios 7/6/00 Name: I: Istructor:

More information

- E < p. ˆ p q ˆ E = q ˆ = 1 - p ˆ = sample proportion of x failures in a sample size of n. where. x n sample proportion. population proportion

- E < p. ˆ p q ˆ E = q ˆ = 1 - p ˆ = sample proportion of x failures in a sample size of n. where. x n sample proportion. population proportion 1 Chapter 7 ad 8 Review for Exam Chapter 7 Estimates ad Sample Sizes 2 Defiitio Cofidece Iterval (or Iterval Estimate) a rage (or a iterval) of values used to estimate the true value of the populatio parameter

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

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

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

Math 140 Introductory Statistics

Math 140 Introductory Statistics 8.2 Testig a Proportio Math 1 Itroductory Statistics Professor B. Abrego Lecture 15 Sectios 8.2 People ofte make decisios with data by comparig the results from a sample to some predetermied stadard. These

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

Lecture 6 Simple alternatives and the Neyman-Pearson lemma

Lecture 6 Simple alternatives and the Neyman-Pearson lemma STATS 00: Itroductio to Statistical Iferece Autum 06 Lecture 6 Simple alteratives ad the Neyma-Pearso lemma Last lecture, we discussed a umber of ways to costruct test statistics for testig a simple ull

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

MA238 Assignment 4 Solutions (part a)

MA238 Assignment 4 Solutions (part a) (i) Sigle sample tests. Questio. MA38 Assigmet 4 Solutios (part a) (a) (b) (c) H 0 : = 50 sq. ft H A : < 50 sq. ft H 0 : = 3 mpg H A : > 3 mpg H 0 : = 5 mm H A : 5mm Questio. (i) What are the ull ad alterative

More information

Chapter 4 Tests of Hypothesis

Chapter 4 Tests of Hypothesis Dr. Moa Elwakeel [ 5 TAT] Chapter 4 Tests of Hypothesis 4. statistical hypothesis more. A statistical hypothesis is a statemet cocerig oe populatio or 4.. The Null ad The Alterative Hypothesis: The structure

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

KLMED8004 Medical statistics. Part I, autumn Estimation. We have previously learned: Population and sample. New questions

KLMED8004 Medical statistics. Part I, autumn Estimation. We have previously learned: Population and sample. New questions We have previously leared: KLMED8004 Medical statistics Part I, autum 00 How kow probability distributios (e.g. biomial distributio, ormal distributio) with kow populatio parameters (mea, variace) ca give

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

University of California, Los Angeles Department of Statistics. Hypothesis testing

University of California, Los Angeles Department of Statistics. Hypothesis testing Uiversity of Califoria, Los Ageles Departmet of Statistics Statistics 100B Elemets of a hypothesis test: Hypothesis testig Istructor: Nicolas Christou 1. Null hypothesis, H 0 (claim about µ, p, σ 2, µ

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

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

Chapter 13: Tests of Hypothesis Section 13.1 Introduction

Chapter 13: Tests of Hypothesis Section 13.1 Introduction Chapter 13: Tests of Hypothesis Sectio 13.1 Itroductio RECAP: Chapter 1 discussed the Likelihood Ratio Method as a geeral approach to fid good test procedures. Testig for the Normal Mea Example, discussed

More information

Chapter two: Hypothesis testing

Chapter two: Hypothesis testing : Hypothesis testig - Some basic cocepts: - Data: The raw material of statistics is data. For our purposes we may defie data as umbers. The two kids of umbers that we use i statistics are umbers that result

More information

A quick activity - Central Limit Theorem and Proportions. Lecture 21: Testing Proportions. Results from the GSS. Statistics and the General Population

A quick activity - Central Limit Theorem and Proportions. Lecture 21: Testing Proportions. Results from the GSS. Statistics and the General Population A quick activity - Cetral Limit Theorem ad Proportios Lecture 21: Testig Proportios Statistics 10 Coli Rudel Flip a coi 30 times this is goig to get loud! Record the umber of heads you obtaied ad calculate

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

LESSON 20: HYPOTHESIS TESTING

LESSON 20: HYPOTHESIS TESTING LESSN 20: YPTESIS TESTING utlie ypothesis testig Tests for the mea Tests for the proportio 1 YPTESIS TESTING TE CNTEXT Example 1: supervisor of a productio lie wats to determie if the productio time of

More information

A Confidence Interval for μ

A Confidence Interval for μ INFERENCES ABOUT μ Oe of the major objectives of statistics is to make ifereces about the distributio of the elemets i a populatio based o iformatio cotaied i a sample. Numerical summaries that characterize

More information

6 Sample Size Calculations

6 Sample Size Calculations 6 Sample Size Calculatios Oe of the major resposibilities of a cliical trial statisticia is to aid the ivestigators i determiig the sample size required to coduct a study The most commo procedure for determiig

More information

Chapter 5: Hypothesis testing

Chapter 5: Hypothesis testing Slide 5. Chapter 5: Hypothesis testig Hypothesis testig is about makig decisios Is a hypothesis true or false? Are wome paid less, o average, tha me? Barrow, Statistics for Ecoomics, Accoutig ad Busiess

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

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

Mathacle. PSet Stats, Concepts In Statistics Level Number Name: Date: Confidence Interval Guesswork with Confidence

Mathacle. PSet Stats, Concepts In Statistics Level Number Name: Date: Confidence Interval Guesswork with Confidence PSet ----- Stats, Cocepts I Statistics Cofidece Iterval Guesswork with Cofidece VII. CONFIDENCE INTERVAL 7.1. Sigificace Level ad Cofidece Iterval (CI) The Sigificace Level The sigificace level, ofte deoted

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

Chapter 6 Principles of Data Reduction

Chapter 6 Principles of Data Reduction Chapter 6 for BST 695: Special Topics i Statistical Theory. Kui Zhag, 0 Chapter 6 Priciples of Data Reductio Sectio 6. Itroductio Goal: To summarize or reduce the data X, X,, X to get iformatio about a

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

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

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

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

Topic 18: Composite Hypotheses

Topic 18: Composite Hypotheses Toc 18: November, 211 Simple hypotheses limit us to a decisio betwee oe of two possible states of ature. This limitatio does ot allow us, uder the procedures of hypothesis testig to address the basic questio:

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

Understanding Samples

Understanding Samples 1 Will Moroe CS 109 Samplig ad Bootstrappig Lecture Notes #17 August 2, 2017 Based o a hadout by Chris Piech I this chapter we are goig to talk about statistics calculated o samples from a populatio. We

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

Statisticians use the word population to refer the total number of (potential) observations under consideration

Statisticians use the word population to refer the total number of (potential) observations under consideration 6 Samplig Distributios Statisticias use the word populatio to refer the total umber of (potetial) observatios uder cosideratio The populatio is just the set of all possible outcomes i our sample space

More information

Last Lecture. Wald Test

Last Lecture. Wald Test Last Lecture Biostatistics 602 - Statistical Iferece Lecture 22 Hyu Mi Kag April 9th, 2013 Is the exact distributio of LRT statistic typically easy to obtai? How about its asymptotic distributio? For testig

More information

Estimation for Complete Data

Estimation for Complete Data Estimatio for Complete Data complete data: there is o loss of iformatio durig study. complete idividual complete data= grouped data A complete idividual data is the oe i which the complete iformatio of

More information

5. A formulae page and two tables are provided at the end of Part A of the examination PART A

5. A formulae page and two tables are provided at the end of Part A of the examination PART A Istructios: 1. You have bee provided with: (a) this questio paper (Part A ad Part B) (b) a multiple choice aswer sheet (for Part A) (c) Log Aswer Sheet(s) (for Part B) (d) a booklet of tables. (a) I PART

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

(7 One- and Two-Sample Estimation Problem )

(7 One- and Two-Sample Estimation Problem ) 34 Stat Lecture Notes (7 Oe- ad Two-Sample Estimatio Problem ) ( Book*: Chapter 8,pg65) Probability& Statistics for Egieers & Scietists By Walpole, Myers, Myers, Ye Estimatio 1 ) ( ˆ S P i i Poit estimate:

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

Chapter 8: STATISTICAL INTERVALS FOR A SINGLE SAMPLE. Part 3: Summary of CI for µ Confidence Interval for a Population Proportion p

Chapter 8: STATISTICAL INTERVALS FOR A SINGLE SAMPLE. Part 3: Summary of CI for µ Confidence Interval for a Population Proportion p Chapter 8: STATISTICAL INTERVALS FOR A SINGLE SAMPLE Part 3: Summary of CI for µ Cofidece Iterval for a Populatio Proportio p Sectio 8-4 Summary for creatig a 100(1-α)% CI for µ: Whe σ 2 is kow ad paret

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

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

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

AP Statistics Review Ch. 8

AP Statistics Review Ch. 8 AP Statistics Review Ch. 8 Name 1. Each figure below displays the samplig distributio of a statistic used to estimate a parameter. The true value of the populatio parameter is marked o each samplig 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

Direction: This test is worth 150 points. You are required to complete this test within 55 minutes.

Direction: This test is worth 150 points. You are required to complete this test within 55 minutes. Term Test 3 (Part A) November 1, 004 Name Math 6 Studet Number Directio: This test is worth 10 poits. You are required to complete this test withi miutes. I order to receive full credit, aswer each problem

More information

ENGI 4421 Confidence Intervals (Two Samples) Page 12-01

ENGI 4421 Confidence Intervals (Two Samples) Page 12-01 ENGI 44 Cofidece Itervals (Two Samples) Page -0 Two Sample Cofidece Iterval for a Differece i Populatio Meas [Navidi sectios 5.4-5.7; Devore chapter 9] From the cetral limit theorem, we kow that, for sufficietly

More information

( θ. sup θ Θ f X (x θ) = L. sup Pr (Λ (X) < c) = α. x : Λ (x) = sup θ H 0. sup θ Θ f X (x θ) = ) < c. NH : θ 1 = θ 2 against AH : θ 1 θ 2

( θ. sup θ Θ f X (x θ) = L. sup Pr (Λ (X) < c) = α. x : Λ (x) = sup θ H 0. sup θ Θ f X (x θ) = ) < c. NH : θ 1 = θ 2 against AH : θ 1 θ 2 82 CHAPTER 4. MAXIMUM IKEIHOOD ESTIMATION Defiitio: et X be a radom sample with joit p.m/d.f. f X x θ. The geeralised likelihood ratio test g.l.r.t. of the NH : θ H 0 agaist the alterative AH : θ H 1,

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

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

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

Chapter 22: What is a Test of Significance?

Chapter 22: What is a Test of Significance? Chapter 22: What is a Test of Sigificace? Thought Questio Assume that the statemet If it s Saturday, the it s the weeked is true. followig statemets will also be true? Which of the If it s the weeked,

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

Stat 421-SP2012 Interval Estimation Section

Stat 421-SP2012 Interval Estimation Section Stat 41-SP01 Iterval Estimatio Sectio 11.1-11. We ow uderstad (Chapter 10) how to fid poit estimators of a ukow parameter. o However, a poit estimate does ot provide ay iformatio about the ucertaity (possible

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

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

UNIVERSITY OF TORONTO Faculty of Arts and Science APRIL/MAY 2009 EXAMINATIONS ECO220Y1Y PART 1 OF 2 SOLUTIONS

UNIVERSITY OF TORONTO Faculty of Arts and Science APRIL/MAY 2009 EXAMINATIONS ECO220Y1Y PART 1 OF 2 SOLUTIONS PART of UNIVERSITY OF TORONTO Faculty of Arts ad Sciece APRIL/MAY 009 EAMINATIONS ECO0YY PART OF () The sample media is greater tha the sample mea whe there is. (B) () A radom variable is ormally distributed

More information

(6) Fundamental Sampling Distribution and Data Discription

(6) Fundamental Sampling Distribution and Data Discription 34 Stat Lecture Notes (6) Fudametal Samplig Distributio ad Data Discriptio ( Book*: Chapter 8,pg5) Probability& Statistics for Egieers & Scietists By Walpole, Myers, Myers, Ye 8.1 Radom Samplig: Populatio:

More information

Stat 319 Theory of Statistics (2) Exercises

Stat 319 Theory of Statistics (2) Exercises Kig Saud Uiversity College of Sciece Statistics ad Operatios Research Departmet Stat 39 Theory of Statistics () Exercises Refereces:. Itroductio to Mathematical Statistics, Sixth Editio, by R. Hogg, J.

More information

Sample questions. 8. Let X denote a continuous random variable with probability density function f(x) = 4x 3 /15 for

Sample questions. 8. Let X denote a continuous random variable with probability density function f(x) = 4x 3 /15 for Sample questios Suppose that humas ca have oe of three bloodtypes: A, B, O Assume that 40% of the populatio has Type A, 50% has type B, ad 0% has Type O If a perso has type A, the probability that they

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

Chapter 23: Inferences About Means

Chapter 23: Inferences About Means Chapter 23: Ifereces About Meas Eough Proportios! We ve spet the last two uits workig with proportios (or qualitative variables, at least) ow it s time to tur our attetios to quatitative variables. For

More information

Sampling Distributions, Z-Tests, Power

Sampling Distributions, Z-Tests, Power Samplig Distributios, Z-Tests, Power We draw ifereces about populatio parameters from sample statistics Sample proportio approximates populatio proportio Sample mea approximates populatio mea Sample variace

More information

STAT431 Review. X = n. n )

STAT431 Review. X = n. n ) STAT43 Review I. Results related to ormal distributio Expected value ad variace. (a) E(aXbY) = aex bey, Var(aXbY) = a VarX b VarY provided X ad Y are idepedet. Normal distributios: (a) Z N(, ) (b) X N(µ,

More information

Instructor: Judith Canner Spring 2010 CONFIDENCE INTERVALS How do we make inferences about the population parameters?

Instructor: Judith Canner Spring 2010 CONFIDENCE INTERVALS How do we make inferences about the population parameters? CONFIDENCE INTERVALS How do we make ifereces about the populatio parameters? The samplig distributio allows us to quatify the variability i sample statistics icludig how they differ from the parameter

More information

Infinite Sequences and Series

Infinite Sequences and Series Chapter 6 Ifiite Sequeces ad Series 6.1 Ifiite Sequeces 6.1.1 Elemetary Cocepts Simply speakig, a sequece is a ordered list of umbers writte: {a 1, a 2, a 3,...a, a +1,...} where the elemets a i represet

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

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

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

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

Notes on Hypothesis Testing, Type I and Type II Errors

Notes on Hypothesis Testing, Type I and Type II Errors Joatha Hore PA 818 Fall 6 Notes o Hypothesis Testig, Type I ad Type II Errors Part 1. Hypothesis Testig Suppose that a medical firm develops a ew medicie that it claims will lead to a higher mea cure rate.

More information