Introduction Hypothesis testing Hypothesis testing is the method of testing whether claims or. hypotheses regarding a population are likely to be true. The purpose of hypothesis testing is to determine whether there is enough statistical evidence in favor of a certain belief, or hypothesis, about a parameter. © Scanned with OKEN ScannerExample ° Is a new drug effective in curing a certain disease? —A sample of patients is randomly selected. Half of them are given the new drug while the other half are given the old drug. The conditions of the patients are then measured and compared. * Hypothesis tests are used heavily in areas like marketing, — To determine whether a certain type of advertisement is effective — Whether a certain group of individuals buys more or less of your product now compared to last year. © Scanned with OKEN ScannerDetermining the range at which a bat can detect an insect by echo Deciding whether hospital carpeting results in more infections Selecting the best means to stop smoking To see which weight-loss program is more effective with a certain group of people © Scanned with OKEN Scanner* A null hypothesis is a type of hypothesis used in statistics that proposes that there is no. difference between certain characteristics of a population * Alternative hypothesis is a type of hypothesis used in statistics that proposes that there is difference between certain characteristics of a population. A hypothesis (often composite) associated with a theory one would like to test. It is denoted by H, * Researchers work to reject, _nullify or disprove the null hypothesis. Alternative Hypothesis also known as Researchers Hypothesis. © Scanned with OKEN Scanner——— Example: To check the water quality of a river for one year, the researchers are doing the observation. As per the null hypothesis, there is no change in water quality in the first half of the year .as compared to the second _ half. But in the alternative hypothesis, the quality of water is poor in the second half when observed. © Scanned with OKEN Scanner* A medical trial is conducted to test whether or not a new medicine reduces cholesterol by 25%. State the null and _ alternative hypotheses. — Ho: The drug reduces cholesterol by 25%. p = 0.25 : The drug does not reduce cholesterol by 25%. p # 0.25 * We want_to test.whether the mean height of eighth graders is more than 66 6 inches. State the null and alternative hypotheses. — Hy : = 66 —H,:>66 ———- © Scanned with OKEN Scanner* To test if college students take less than five years to graduate from college, on the average. — HO: Average time taken by student to graduate from college is 5 years — HO: M=5 years — H1: Average time taken by student to graduate from college is less than 5 years —H1:M<5 years © Scanned with OKEN Scanner* To test if it takes fewer than 45 minutes to teach a lesson plan — HO: Average time to teach a lesson planis 45% minutes fe — HO: M=45 mins — H1: Average time to teach a lesson plan is fewer than 45 minutes —H1:M <45 mins $e . be US mo g Fe8S 11 © Scanned with OKEN Scanner* Inan issue of U.S. News and World Report, an article on school standards stated that about half of all students in France, Germany, and Israel take advanced placement exams and a third pass. The same article stated that 6.6% of U.S. students take advanced placement exams and 4.4% pass. Test if the percentage of U.S. students who take advanced placement exams is more than 6.6%. State the null and alternative hypotheses. HO: The percentage of U.S. students who take advanced placement exams is 6.6% a ¢ = Let HO: P = 6.6% — H1: The percentage of U.S. students who take advanced placement exams is more than 6.6% wep 2oeh H1: P > 6.6% © Scanned with OKEN Scanner* Ona state driver’s test, about 40% pass the test on the first try. We want to test if more than 40% pass on the first try. State null and alternative hypothesis. — HO: Percentage of people pass the driver's test on their first try is 40% He $ p= Uo% — HO: P = 40% — H1: Percentage of people pass the driver's test on their first try is more than 40% 40%> Ky 2 p> 10% cen 13 © Scanned with OKEN Scanner| Gr oy, f Cevdreadlfche? ja greeted ysl * cadradisln | * Null hypothesis: 1. > "SS “ia, Hote & \ be Bo how Ay . | l\ — A hypothesis associated with a contradiction to a i || theory one would like to test. It fs denoted by H, —H): p= 66 ¢ Alternative hypothesis: — A hypothesis (often composite) associated with a theory one would like to test. It is denoted by H, ; nc G Gh) girecialid unlicr —H,: > 66 icdey oe fe ee By as ctl yf © Scanned with OKEN Scannera A jon sac mapas oll UE ee “ ce _ Type | Error: ° rota ee a es We reject the null hypothesis when the null is true. Put another way, a = Probability of Type | error = P(rejecting Hy | Hy is true) Typical values chosen for a are 0.05 or 0.1 © Scanned with OKEN ScannerType I Error: remap m (He thee We do not reject the null hypothesis when it is not true. Put another way, B = Probability of Type II error = P(do not reject H, | H, is false) xo Vel treKe fsthess wshur a = safe rho woh ey = Arse one) = &C Typalerss) 19 © Scanned with OKEN Scanner* The Jury is instructed to assume the person is innocent, and only decide that the person is guilty if the evidence convinces them of such. Hence favored assumption is the person is innocent , ¢ When there is a favored assumption, the presumed innocence of the person in this case, and the assumption is true, but the jury decides it is false and declares that the person is guilty, we have a so-called Type | error 21 © Scanned with OKEN Scanner* Conversely, if the favored assumption is false, i.e., the person is really guilty, but the jury ae eee crv mane declares that it is true, that is that the person is innocent, error. then we have a so-called Type II * Thus, 22 © scanned with OKEN Scanner* The favored assumption is that the person is guilty. In this case the roles of the Type | and Type Il errors would reverse to yield the following table © scanned with OKEN Scanner2 ol . re 4 are) * Size: sty a — The size of a test is the probability of incorrectly rejecting the null hypothesis if it is true. Probability of rejecting Hy even when H, is true. Hence we can say that probability of committing a Type | error is size of a test. Size is denoted by a — a = P(Reject H,| H, is true) * Power: sey boa 1 # =F — The power of a test is the probability of correctly rejecting the null hypothesis if it is false. Probability of rejecting Hy even when Hy is true. Hence we can say that probability of not committing a Type Ilerror is power of a test. - +1 -B = P(Rejecting H, | H, is false) roped [ne swftsa) 24 © Scanned with OKEN ScannerRelation between Type II Error and Power We know that Type Il Error = Do not reject Hy when H, is false Hence P(Type II Error) = P( Do not reject Hg | Ho is false) B = P( Do not reject Hy | Hg is false) pe ab (aktfat H. / he Bille) = 1 -P( Reject Hy | Hp is false) A= aeons # Pecan (~ f = 1 - Power = (— PCTyfsd omren) Hence Power = 1-8 | i.e. Power = 1— P(Type Il Error) | 25 © Scanned with OKEN ScannerExample Let p be the probability that a coin will fall, head in single toss. In order to test the hypothesis H, : P=0.5 the coin is tossed 6 times and H, is rejected if more than 4 heads are obtained. Find the size of the test. If H, : P=1/3 find the power 26 © Scanned with OKEN ScannerSolution: Let X denotes number of head when a coin is tossed Given hypothesis: Hy : p = 0.5 and H,: p = 1/3 Rejection condition: 2X > 4 n=6 Size = P(Reject H, | H, is true) P(2X > 4| Hy is true) 2x4 ae ~ PCr rte | Me oo Te) FE H> AH Me ee) 27 © Scanned with OKEN Scanner=P (2X >4| p=0.5) X~ B (1, Po) => IX ~ B(6, Po) P( 2X > 4) = P (X=5)+P(X=6) = §C,p54o54+5C.poq&s = 6197p + Poe = 6(0.5)5(1-0.5)+(0.5)6 =7(0.5)6 =0.109 yr b(t be => 2K 1 (2x? map ( 6)b-) = Sep EPA 2 kar # yu &(os) re 3609) O.(ot host pcxe st rus) © Scanned with OKEN Scanner* Power = P(Reject Hy | H, is true) Get Ht atone) =PQX> 4] pata) fom 8” X* 8 (1, Pa) peaes 2 432 POR De OO => 2X ~ B(6, P,) = ob + S48 pha 4)=P (X=5)+P(X=6) = 6b y ae ec = ®C.pq®54+5C.poqes ae = 6p,°q; + p,° oe fan oo = 6(1/3)°(2/3)+(1/3)® =0.0178 Hence P(Type II Error) = 1— 0.0178 = 0.9822 29 © Scanned with OKEN Scanner5 whidh UL test Abc f& a values | ruyfeatharis » ay ee a be apd Be all Rye * Region of Acceptance — The set of values of the test statistic for which we fail to reject the null hypothesis Region of Relection/ Critica Region — The set of values of the test statistic for which we reject the null hypothesis e ti mshatdl Value delinnrtirg sl pugi Critical ee A eae qd ut Bipatochs, — The threshold value delimiting the region of acceptance and rejection of the null hypothesis 30 © Scanned with OKEN Scanneryee Uwe Rede << Mean, Median, Mode NN a Acceptancé Region Critical ae Region Critical Values 31 © scanned with OKEN Scanner2 dgavodt “poor solaron | ered o glue sad _»te® . oge giddy at Level of significance “<7; Level of significance, or significance level, refers to a criterion of judgment upon which a decision is made regarding the value stated ina null hypothesis. The criterion is based on the probability of obtaining a statistic measured in a sample if the value stated in the null hypothesis were true, _, am ow Sage paEtisics Maximum probability of rejecting Hy even when Hy is true 7 The criterion or level of significance is typically set at 5%. That means we are 95% confident that the results have not occurred by chance gaye ey 32 © Scanned with OKEN ScannerP Value 5 joHe porta a Seb “SG ss ral hypo 8 ghuw H ° Ap value is the probability of obtaining a sample outcome, given that the value stated in the null hypothesis is true. The p value for obtaining a sample outcome is compared to the level of significance. * Measure of the strength of evidence the sample data provides against the null hypothesis He eee Lor afin, Se * P—val = p = P(Z > Z,a) 7 Wh ~ bo P(e ™ Zay) Haosurn of Abe Py 1 Gtne daringte late prretes Aquat The mruald teypothud —— P-value = p= P(C2™ Zag) plese) Zo 3 sutcawt, At « © Scanned with OKEN Scanner: a ohwlh Interpretation: pe pened ane * When the p value is less than 5% (p < .05), we reject the null hypothesis. * We will refer to p< .05 as the criterion for deciding to reject the null hypothesis, 2 #< °° epee, * Note that when p = .05, the decision is also to reject the null hypothesis. P00, Ging, Gaciaiow By ales Ke Sypet * When the p value is greater t an 5% (p > .05), we do not reject the null hypothesis. => 0-05 > vet reject, * The decision to reject or retain the null hypothesis is called ance. When the p value is less than .05, we reach significance; the decision is to reject the null hypothesis. When the p value is greater than .05, we fail to reach significance; the decision is to not reject the null hypothesis. Pr oeeS wu visach Ayrifeowa, Ae ition 2040 sefest ths WEE typaitoeAe toh the be ee dan aif Ho Jrs2eth vy 4 7 34 se fail ho Toes Ly Bo" pg rags We melt © Scanned with OKEN Scanner° Level of Significance Value (in the shaded ion) will lead to Hypothesis ¢ P value zobs tejection of Null” Mean’ alpha Xt Critical Value 35, © scanned with OKEN ScannerTypes of test * One Tailed test * A one-tailed test is a statistical test in which the critical area of a distribution is one- sided _so that it is either greater than or less than a certain value, but not both. If the sample being tested falls into the one- sided critical area, the alternative hypothesis will be accepted instead of the null hypothesis. 36 © Scanned with OKEN Scanner* Two-tailed test * A two-tailed test is a method in which tl critical area of a distribution is two- sided and tests whether a sample is greater ,* than or less than a certain range of values. It is cs lorae Petey used i in null-hypothesis testing and testing for statistical significance. 37 © Scanned with OKEN ScannerHow to determine? * The type of test is determined by the Alternative Hypothesis ( H, ) * Left Tailed Test —_ Fi Left-tailed test. — H,: parameter < value * Right Tailed Test i> —H,: parameter > value * Two Tailed Test_ — H,: parameter not equal value BY ” “Two-tailed test b it ago aay sv age SValuaspe Lt es > RigshY 38 U? Bereottircyabe 2 HS pomustaS dee © Scanned with OKEN ScannerSTEP 1: Set up hypotheses and select the level of significance a. — Ho: Null hypothesis (no change, no difference); H,: Research hypothesis (investigator's belief); a =0.05 — The research or alternative hypothesis can take one of three forms. 1 07 (gumporedars Upp al at — Hy: W> Wo, where py is the comparator or null value and an increase is hypothesized - this type of test is called an upper-tailed test; — Hy! 1. < Uy, where a decrease is hypothesized and this : is called a lower-tailed test; or 4° 4+ Ae => ctecrenas 2 hypetiais _ 1b # HB o, where a difference is hypothesized and this is called a two-tailed test. wae Jn hyfettasis.o) thi Caled @ 92 hile est 39 © Scanned with OKEN Scanner* STEP 2: Select the appropriate test statistic. — The test statistic is a single number that summarizes the sample information wo * STEP 3: Set up decision rule. — The decision rule is a statement that tells. under what circumstances to reject the null hypothesis. The decision rule is based on specific values of the test statistic. 40 © Scanned with OKEN Scanner* Rejection Region for Upper-Tailed Test (H,: uU> Uy ) Ay CH 46) The decision rule is: Reject H, if Test statistic value > Critical value " Rejection ait, Becepta Region sf Region ; {1-4 a a h Jes Agia Prramiln le (: . SOR ke al wa gl IP? WE 46 7 D—Feased tual Crd Lotter Hertel feel llr Hy eUL © Scanned with OKEN Scanner* Rejection Region for Lower -Tailed Test (Hy! H< Ho ) The decision rule is: Reject H, if Test statistic value < - Critical value ae, at Ted Sonate = alee Rejection Region Qa Acceptance © Scanned with OKEN Scanner* Rejection Region for Two -Tailed Test (Hy: L# Up ) The decision rule is: Reject Hp if Test statistic value < - Critical value or Test statistic tic value > Critical value i.e; Reject Hy if |statistic value] > wae VHS —Crvetue Critical value a GeeBads = Ci vatce Rejection Region Rejection Region 2 a 2 43 © Scanned with OKEN Scanner