AP Statistics Name:_ANSW OC ner eye fy Chapter 9 - Testing a Claim Date: QJIS[1G Per Section 9.3: Tests About a Population Mean Day 1 READ PG.574 -584! _ Today, you will state and check conditions for performing a significance test about a population mean, You will also perform the significance test “One-Sample 1 Test for a Mean”. | TT, 3 Day 1 Inference about a population mean uses a distribution with m- 1 degrees of freedom (except in rare cases where the population standard deviation ¢ is known). Before performing a significance test about a population Proportion, you must check the conditions. You'll notice that they are exactly the same conditions as for constructing confidence intervals for a population mean jt. CONDITIONS FOR PERFORMING A SIGNIFICANCE TEST ABOUT A MEAN RANDOM: The data come from a well-designed random sample or randomized experiment 10% CONDITION: When sampling without replacement, check that n<;,N. NORMAL/LARGE SAMPLE: This condition is satisfied if =| 1 population has a Normal distribution or the sample size Is large n >30. If the population distribution has an unknown shape and the sample size is small n <30, draw a graph of the sample data to assess the Normality of the population. Do not use + procedures if the graph shows strong skewness or outliers. CALCULATIONS: TEST STATISTIC AND P-VALUE ‘When performing a significance test, we perform calculations assuming the null hypothesis H, is true. The test statistic measures how far the sample result diverges from the arameter value specified by H. 0 ; GEN mpg statistic — parameter 0 Xo! dnt test statistic ne ered standard deviation of statistic For a test of H:H = Ho, our statistic is the sample mean Z and we use the sample Standard deviation a; = = in most cases since we typically don't know the population « to use oy = = ¥ WED 0. The resulting test statistic is: € = ee a a ie ee This statistic has approximately a # distribution with m ~ J degrees of freedom when the Normal/Large Sample condition is met. PROCEDURE NAME: One-Sample f Test for p.DON’T FORGET TO READ PG.574 -584!!1! KSQNEEEE Healthy Streams Performing a significance test about ‘The level of dissolved oxygen (DO) in a stream or river is an important in- dicator of the waters ability to support aquatic life. A researcher measures the DO level at 15 randomly chosen locations along a stream. Here are the results in milligrams per liter (mg/} ems Entel 4m 50am 6m ata 550 4m 440 ITO 542 6 (401 «466 «(287573 555 \ooees perform A dissolved oxygen level below 5 mg/l puts aquatic life at risk. PROBLEM: = (2) Dome have convincing evidence at the c= 0.05 sigalficance level that aquatic Ie in this stream is at risk? (b) Given your conclusion in part (2), which Kind of mistake —a Type lerror or aType I error—could youhave made? Explain what this mistake would meanin context, (a) STATE: We want to test a claim about thetwe man dissolved oXugen wel Ms th this stream Ot the X=0.05 Lovet. Our hypothises re: Hy: M=5 A Ha? M<5 PCAN * We should do a One-Sample t ‘ast for Mu if conditions are mat * Random : The researchar measured Wu DO Wwe! at 1S randomiy . Chasen. Locations clang dhe steam. * 107. Condition : there 8 an mfnite number of Possible Locations ei, along the stream, so the 107: Condition ish now F * Not Large Sample: Sine our samole eii2 ssmall (15 <30 and we don't know rf +he eoowation is Normal, we must en Hu data. the histparum appears rovahly symmetic with no strong skews or outliers, sb itis safe to use + procodureg . pewvel orgie) DO? The sample mean 00 lwyel is K= 4.741 mg/L and sompw standard daviohon is Sx 20.439 from tne 1-vorstats Ratu re on our calc etest statistict: ¥-Ho - 4491-5 : = - 0.4 t-dist. ; Sin OG/ TS 4 ee aes @ P-valut :using cotadator with df 219-1 =14 sy (b) cd $ Clower :-10000 , upper :-0.44, df 214) = 0.192, ao CONCLUDE: We fail to rejeck Hp because +e is greater Han OOF 3h Ri ane tive = Dos" do_not \ si Nan cus cee SOURCE SRS GAB ogres HME Ohigin ter RS wld hove made o. Type HL enor whore we tD reyie (wie it 1S realy alos Raquotic Ufe woubol suffer. we ee‘Try it! Less Music? We uuu $ext he Alsi rock rd station cms to pay average of 50 minutes 6 music every hour. However, it seems that very time you turn to this station, there is a commercial playing, To investigate their claim, you randomly select 12 different hours during the next week and record what the radio station plays in each of the 12 hours. Here are the number of minutes of music in each of these hours: “into Li ¢ dd 44°49 «45 Sl 49 5349 4447504548 STAD F Ha RUC SO eninucts eutry how? TATE: ’ : We want to est a dlaim. about the true moan number of minus of music played every how atthis radio stehor at He %=0.05 sianificancwe toe). Our hypotrses are ? Hy: M250 PLAN: Ha! M250 . We should etform a One-Samole t-Test for M if conditions are, Random: d random sample of hours was sebocied . mer *\D7. Condition : There are more than 10Ua)= 140 hours of music. plaued duning the wwek. q *Normal/Large Sample : We don} know if the population disinbution of music times [s appmxinaelu Normal. Since ue have a small somnt SI22-(12< 2), we will graph the dada to look for any dixpartures from. Norm. . g pm Normality rye dotolot does not show much. 2 sigieeMremeatmee StOWWnUss oF ON ou eS §0 it 18 Sal 44 4G De? sy to USe t -ploadures for Huse date. DO: Timecrainvies) The sample man number of music minuies played 1s = 44.9 min. with a standard dudation of Sx=2.81 using our calculator. thus, He est statistic ts + t= KeMe 2 44-94-50 sa Sn Isl] ; P-value : using He calauactor command with df=1a-1 =|) Led f lower :- 10000, upper :-2.59 ,df:11) = 0.012% e0 (CONCLUDE: 7 we relect the null hupottusis Ho pecause the P-volut 0-012l0 ts less than our significance (ave) aa . ve Convincing evidence. thot the radio N Plous Liss than 50 minudes of musi¢, every hae £-259 °Try it!! Short Subs Abby and Raquel like to eat sub sandwiches. However, they noticed that the lengths of the “6-inch sub” sandwiches they get at their favorite restaurant seemed shorter than the advertised length. To investigate, they randomly selected 24 different times during the next month and ordered a “6-inch” sub, Here are the actual lengths of each of the 24 sandwiches (in inches): 450 475 475 500 500 500 550 5.0 550 550 550 550 575 575 5.75 600 6.00 600 600 600 650 675 675 7.00 Problem: (a) Do these data provide convincing evidence at the a = 0.10 level that the sandwiches at this restaurant are shorter than advertised, on average? STATE: We want to test the following hypotheses at the c = 0.10 significance level Hy a =6 Hy a <6 where = the true mean length of “6-inch” subs from this restaurant. PLAN: If conditions are met, we will perform a one-sample ¢ test for ye + Random: Times were randomly selected to order a sub, so this should be a random sample of 6- inch subs from this restaurant. . * 10% Condition: This restaurant makes more than 10(24) = 240 subs in a month + Normal/Large Sample: We don’t know if the population distribution of lengths is approximately ‘Normal and we don’t have a large sample size, so we will graph the data and look for any departures from Normality. i. The graphs do not show much skewness and there are 3 no outliers, so itis reasonable to use ¢ procedures for etoets these data. 45 0 55 60 65 7K Length (inches) DO: The sample mean length is ¥ = 5.68 inches with a standard deviation of s, = 0.66 inches, 568-6 ‘+ Test statistic = 238. t-dismbution ‘la a? with df= a3 ‘= P-value Using the ¢ distribution with 24 ~ 1 = 23 degrees gem of freedom, the P-value is between 0.01 and 0.02. Using i, technology, P-value = tcdf (lower:~ 10000, upper:-2.38,df£:23) = 0.0130, $252.38 0 CONCLUDE: Because the P-value of 0.0130 is less than «= 0.10, we reject the null hypothesis. There is convincing evidence that the true mean length of “6-inch” subs at this restaurant is less than 6 inches. (b) Given your conclusion in part (a), which kind of mistake—a Type I or a Type II error—could you have made? Explain what this mistake would mean in context. Because we rejected the null hypothesis, itis possible that we made a Type I error. In other words, itis possible that we found convincing evidence that the mean length was less than 6 inches when in reality the mean length is 6 inches.4-dustibution uth df=Go READ PG.5781!!! P-volue Page 578: Alternate Example é More Practice with Table B Problem: . Ge teat (@) Find the P-value for a test of 70: “ = 10 versus “a: > 10 that uses a sample of size 75 and has a test statistic of ¢ = 2.33. © Table & does not indudt a Lune for df= 15-1 = 74 , so we will be conservative and use df =GO from Table B. Move accross the df=6O row ond you'll notice that £22.33 1s between 2.094 and. 2.390, therefore +e P-value ts betweer 0.0] and 0.02. * Using, the calculator: t- cd f Clower: 2.3, upper: loot, df:74) 20.011 (b) Find the P-value for a test of 0: # = 300 versus 4a: # * 300 that uses a sample of size 10 and has a test statistic of =-0.51, >t=0-5| * Smee df=l0-1=4, go 40 row df=4 ih Table B. As we move accidss the row, all of the P-valuus are Larger than t=0.51. This means that the area in the t-curve to the tight of t=0.5l is greater than P= 0.25. Since this 1s a two-sided test, the Pe volut must be at last 0.50 | #-0.51 €=0.51 Cimon ute PST . Using the calulator : feat (lower: 0.51, upper: 10000, d£:4) = 0.3112 & multiply by Q > get P-valut = 0.0283