Lesson 1. Hypothesis is a tentative, testable assertion regarding the occurrence of certain behaviors, phenomena, or events; a prediction of study outcomes, ‘Two types of hypotheses that will be explored here 1. Null hypothesis 2. Alternative hypothesis Null Hypothesis states that the null condition exists; that is, there is nothing new happening. It is a statement of what the researcher believes will be the outcome of an experiment or a study. Before studies are undertaken, business researchers often have Some idea, or theory based on experience or previous work as to how the study will tarn ut, These ideas, theories, or notions established before an experiment or study is conducted are research hypotheses. Some examples of research hypotheses + Older workers are more loyal to a company. + Companies with more than P1 billion in assets spend a higher percent age of their annual budget on advertising than do companies with less that P1 billion i n assets, + The implementation of a six sigma quality approach in manufacturing will result in MODULE 8 __Introduction to Hypothesis Testing greater productivity + The price of scrap metal is good indicator of the industrial production index six months later. + Airline company stock prices are positively correlated with the volume of OPEC wil production + Mathematics performance ofthe students improve after the remediation class has been conducted. 81mal hypothesis structure |i order to scientifically test research hypotheses, a more formal byt u heeds to be set UP using condition statistical hypotheses. esis states that All statistical hypotheses cons feats. sta nothing "ew happening, the old theory is still true, the old st system is in control. of two parts, a null hypoth andard is correct, ar or hand, states that the new the alternative hypothesis, op the other hand, states that ue theory is true, there are new standards, the system is out of control, and/or something is happening. Lesson 2. Level of Significance ‘To establish whether our obtained sample difference is statistically significant ~ the fesult of a real population difference and not just sampling error ~ it is customary to set up a level of significance, which is denoted by the Greek letter (alpha). The alpha value is the level of probability at which the null hypothesis can be rejected with confidence, and the alternative hypothesis can be accepted with confidence, Accordingly, we decide to reject the null hypothesis if the probability is very small (from the example, less than 5 chances out of 100) that the sample difference is a product of sampling error. Conventionally, we symbolize this small probability by p < 05. Levels of significance do not give us an absolute statement as to the correctness of the null hypothesis. Whenever we decide to reject the null hypothesis at a certain level of significance, we open ourselves to the chance of making the wrong decision which could be ‘Type | or Type Il error. Lesson 3. Type! and Type Il Errors ‘Type I error is committed if we rejected the null hy i 'ypothesis when in fact it should be accepted. Type II error is committed if we accept the n ull hypothesis when in fact it should be rejected, Decision Accent Reject H, Reject H, Reality Hy is true Correct decision Type | error His not True Type trot [correct Decision J Lesson 3. Comparing the same sample Measured twice i ed in determi Example, Social researchers are interested in determining the im mobility on fling of neghborines (that is, postive feng ae! Ped residential prerelocation neighborhood and neighbors in the postr neighbors in the statement of hypotheses: ‘cation neighborhood). The H, + (H, = Hy): The degree of neighborliness does Not differ b, relocation efore and after the H,: ((u, #44) : The degree of neighborliness 82Pact of forced relocation on nei To test the iny ghborliness, the researchers interview a random sample of six individuals about their neighbors both before and afier they were forced to move. Interviews yield the following scores of neighborliness (higher scores from 1104 greater neighborliness) [. r i 2 Respondent | Belore Move | After move Difference Difference (x) (K,) (D) (O°) >———_—_| A 2 1 1 1 8 1 2 4 1 c 3 1 2 4 o 3 1 2 4 | E 1 2 A 1 | J+ | 4 1 3 9 | Total 3X,=8 ED’=20 | | ‘The following Steps Step 1. Find the mean for each point in time. 4 xy, cy 8 . | = Ste M353, oe | Res 7238 e133 JA Step 2. Find the standard deviation for the difference between time | and time 2 } Where S,=_ standard deviation of the distribution of before- After difference scores D = after-move raw score subtracted from before- Move raw score | n= number of cases or respondents ofthe study Step 3. Substitute the value to the formula: /20~ (2.331.337 % = [5 333-1 1233 = 153ME Step a Find the standard error af the mean dillerence f + Ik yak sD \ Vos WN} vot NS " Step. Tanshate the sample mean dliflerence inte, units of standard error of the mean ditlerence of das AS i La os Step 5. Find the number of degrees of freedom df= nts 6-125 Step 6. Compare the obtained (ratio with the appropriate (ratio in the table 1, Appendiy A Obtained t= LAr Table 1 = 2571 with dt= 5) at =05 Step 7. fon Rule. Because the obtained t (1.47) is less than the 5% level of signiticane research hypothesis, Ceritical (2.571) at pothesis and reject the AUS 5, then we accept the null hyp Lesson 4: Test of Difference Between Proportions Example 1. A social psychologist is interested in how personality charac expressed in the car someone drives. He wonders whether men expres control than women by driving big ears. He takes a sample af 2000 over 18 and determines whether oF not they drive a full-size ¢ sample sizes for analysis were as follows; 180 for men and 15 hypotheses were formulated: Fistics are 8 greater need for ales and 200 females. at. Consequently, the final for women, The followin Nall Hypothesis: H,: ‘The proportions of men and women who drive big cars are equal Research Hypothesis: H, ‘The proportions of men and we omen who drive big cars are not equal. Male Female Overall Par 180 180 330 Own big cars a 48 129 Proportion with big cars AS 32 39Step 1. Compute the two sample proportions and the combined sample proportion 8 Bae ac 32 N, 180 pre PLANS (us0y(45)+ (1501632) _ 81448 129 NIN, 180+ 330 330 Step 2. Compute the standard error of the difference. N, +N s,=Pa-ps |[N - : anelaiered ys pa = PLP vf 7 } = (3901-39) | Taoyrs0 | 3 (39)(.61) ina = .0539 Step 3. Translate the difference between proportions into units of the standard error of the difference. P, -P, 45 - 32 z =241 0539 ey Step 4. Compare the obtained z = (2.41) with the critical value of z = 1.96. Because the obtained value (z = 2.41) is greater than the critical value of z = (1.96), then we reject the null hypothesis. The difference between sample proportions was statistically significant, the social psychologist was able to conclude that men and women generally tend to drive different sized cars. zvalue 68% 1.00 95%. 1.96 99% 258 Example 2. Testing about a Proportion. 1e morning beverage market shows that the primary breakfast beverage for A ino children is milk. A milk producer believes the figure is higher for the city. To test this idea, the researcher contacts a random sample of 550 residents of the place and asks which primary beverage they consume for breakfast that day: Suppose 115 Feplied that milk was their primary beverage, using 5% level, test the idea thatthe milk figure is higher for the particular City. 85Solution: Step 1. State the Null (H.) and Alternate (Ha) hypotheses Hy: p=.l7) Hep>.i7 Step 2. Compute the sample proportion (f) using the following formulas: x i n Where x = proportion who said that they use milk as breakfast beverage n= sample size 209 Step 3. Compute for Z using the following formula: pop (p)(@) 7 sample proportion population proportion = 1-p Substitute the values to the formula Lu Step 4. Decision Rule Reject the Null Hypothesis because the observed Z test (2.44) is Critical value of Z = 1.96. The clelated ts saitic is ote ered een, ae the value. Step 5. Business Implication. To make managerial decision, the Pe evidence to reject the null hypothesis thatthe breakioct beveragent cat enough in the City is milk. The researcher can conclude that the average breakfa of children of the children is more than 17%, eakfast beverageEXERCISE 9 Testing Hypotheses About Single Population avalue of f= 465. Use =.05 to test this hypothesis 87 Parameter Name Rati : Course & Curriculum Year Date 1, Suppose you are testing H, = 45 versus H.:> 45 random sample of 310 people producesJe schedule. Suppose a rhe researcher randomly Jetermine that 303 of these have eter Mjoes the test show enough accounting firms offer | ayes Hexibl nies offer employer ge. T 2. A study showed that 79% of com researcher believes that in accounting firms this ah interviews eleterrn jevel of significances tly lower proportion of selected 415 accounting firms and thr Hlexible schedule scheduling. With 1! evidence to conclude that a significa employees flexible scheduling? 3. Suppose you are testing H,; p = .29 versus H,: p #.29, A random sample of 740 items shows that 207 have these characteristic. With a .05 probability of committing a Type 1 error, test the hypothesis. For the p-value method, what is the probability of the observed z value for the problem? If you have used the critical value method, what would be the two critical values be? How do the sample results compare with the critical value?Lesson 5. One Tailed and Two-Tailed Tests On led Test. A one-tailed test rejects the null hypothesis at only one tail sampling distribution or when the rejection region is located at only extreme of the ra of values for the test statistic Two-tailed test. Two-t ed tests distribution or when the rejection region is located at both extremes of the distribution. Lesson 5. One-Tailed Test for: A, Statistical Inferences for two Related Populations eject the null hypothesis at both tails of the sampling Example. Suppose that an educational researcher wishes to test whether a particular remedial math program significantly improves math skills. Student Before (X,) | _Aller(X,) | Difference (0) | Difference? (D*) 1 58 66 8 64 2 63 68 45 25 3 66 n 4 36 4 70 76 4 36 5 63 78 45 225 6 51 56 4 25 7 44 69 25 625 8 58 55 3 9 50 55 4 32623 | _ Y= 595 One Tailed Test (Sample) Null Hypothesis : Math ability does not improve after remediation. (H,2 4) Alternative Hypothesis: Math ability improves afier remediation u,) school graduates than among public high school graduates > Hy Steps in Computing the Data to make the final decision, Step1 Determine the sample mean xy, =r X°F 1Step 2 Compute for the sample standard deviation 526 5.8 Step 3 Calculate the standard error of the difference between means. Standard Error = 1.95 Step 4 Translate the sample means difference into units of the standard error of the ditference. Step5 Determine the degrees of freedom. df = N+N,-2 = 22+50-2 df = 70 Step 6 Compare the obtained ¢ ratio (1.55) with the critical t ratio = 1.671 obtainedt = 1.55 tabler = 1.671 df = 70 a = 05 Step7. Decision Rule: Do not reject the mull hypothes is ess than the critical t (1.671) "*Pothesis because the obtained t (1.55) means . expectations (85 and 82), the difference war "#5 COMsistent with t fe was 7 he professors obtained was not statistically significant," “8hificant. In short, the result Five-Step Hypothesis Testing Procedure Used: Step 1: State the Null and Alternative Hypotheses Step 2: Select the Level of Significance Step 3: Give the test statistic and compute Step 4: Formulate the Decision Rule Step: Interpret the result and make an implicationEXERCIS| Testing the Difference Between Means Name Rating —____ Course & Curriculum Year Date Instructions: Five-Step Hypothesis Testing Procedure Used: 1, A criminologist was interested wl the race of the defend: the prison terms giv hhether there was any disparity in sentencing based on lant, She selected at random 18 burglary convictions and compared fen to the 10 whites and 8 blacks sampled. ‘The sentenced length s (in years) are shown for the white and blacks offenders. Using the data, test the null hypothesis that white and blacks convicted burglary in this jurisdiction do not differ with respect prison sentence length. Back [4 8 7 3 5 4 5 4 White ]3°5 4 7 5 5 6 93esidents on one floor of a ! 2. In a field experiment on the effects of perceived conteal, a ieieome, ; ursing home were given opportunities for increased Fo ine, choose anid keane arrange their own furniture, decide how to spend free Bea eal The the | of plant), whereas the residents on another floor were te 2 21-point scale) fallow tall took care of these details. The feelings of well-being (om hése data, test the null the conditions of increased and no increased control. Using ~ LAR hypothesis that this minimal manipulation of perception of cont residents fecling of well-beingA social psychologist was interested in sex diflerences in the sociability of tee Using the number of good friends as a measure, he compared the sociability of eight females anu seven male teenagers, Test the null hypothesis of no difference with respect to sociability between females and males, What dev your results indicate?" 4. “Are intellectual more likely to wear a beard?” wondered a social psychologist. Ay Partial test, he asked samples of bearded and clean-shaven men about their level of educational attainment. Using these data on years of schooling, test for the significance of the difference between means Beard No beard 18 Na 23 it 12 2 14 16 18 4 16 12 15 125. Twelve participant underwent an assertiveness training program. Before and alter completing the program, they were asked to complete a checklist af adjectives that describes their persona compiled in the y: The number of adjectives describing positive attributes were relore” and “after” condition, Using the data in the table, compute “t Remember to include in your answer (a) Null hypothesis; (b) the research hypothestss (c) (a) degrees of freedom (e) the critical value; and ({) a brief statement about whether or not is significant Number of Positive Adjectives Participants Before Alter i 4 a 2 24 22 3 24 a 4 16 19 5 18 18 6 20 25 7 15 W 8 20 18 9 W7 a 10 16 24 "1 12 15 12 19 24 97&. You are interested in determining whether an experimental birth sss pill tas side effect of changing blood pressure. You give five of them a placebo for a month ang then measure their diastolic blood pressure, Then you switch them to the birth contro} pill for a month and again measure their blood pressure. ‘The other five women receive the same treatment except they were given the birth control pill first for @ month, followed by the placebo for a month. ‘The blood pressure readings are shown here, Diastolic Blood Pressure k Subject Number Birth Control Pill Placebo 1 108 102 a 76 68 i 3 69 66 : 4 78 n i 5 m4 76 5 85 80 7 79 82 8 78 79 2 80 78 10 at 85 (a) State Ho and Hi; (b) calculate “t’; (c) calculate the degrees of freedom: (A) critical value; (e)write a brief interpretation of the result.7. Use the given data and 1% level of significance to test the following hypothesis. Assume the differences are normally distributed in the population. Pair Sample 1 Sample 2 1 38 22 2 27 28 3) 30 a 4 27 38 5 30 38 6 a 26 7 36 19 8 38 3 9 33 35 35 44. iG eeUse the data given to test the following hypothesis (a = .05). Assume the differences ay, normally distributed in the population. Individual Before Alter 1 107 102 ‘ 2 99 98 ! 3 110 190 ‘ 4 113 108 i 5 6 89 i 6 98 101 i 7 100 99 { 8 102 102 i 9 107 105 10 109 110 " 104 102 12 99 96 13 tot 100 - 1009. Construct a 93% confidence interval to estimate D from the following sample information. Assume the differences are normally distributed in the population Client Before After | 1 32 40 | 2 28 23 3 35 6 | 4 32 32 5 26 29 6 25 a 7 a7 39 | 8 16 30 9 35 n 101 = ili eeers that mogt | ention of sales manag st | 10. The vice president of marketing brought to the attention A maintained client of the company’s manufacturer representatives eocee managers bro ight the reps in, relationships in a disorganized, haphazard way ae se an organizer to schedule on how to Use @ ig session OF at more effectively. Sales reps were = their effort. Sales managers on a randomly selected day ther significantly differences in the for a three-day seminar and trai visits and recall pertinent information about each clier taught how to schedule visits most efficiently to maxim were given data on the number of site visits by sales reps 0” °° both before and after the seminar use the following data 10 test ¥ more site visits were made after the seminar (a = .05). Assume © ‘number of site visits are normally distributed. Client [Before after 1 2 4 2 4 5 3 1 3 4 3 3 5 4 3 6 2 5 i s 2 3 ‘ a 3 4 i if 9 1 102.11. Eleven employees were put under the care of the company nurse of high cholesterol readings. The nurse lectured them on the dangers of this condition and put them on a new diet. Shown are the cholesterol readings of the 11 employees both before the new diet and one month after use the diet began. Construct a 98% confidence interval to estimate the population mean difference of cholesterol reading for people who are involved in this program, Assume differences in cholesterol readings are normally distributed in the population. Employee Betore Alter 1 255 197 2 230 225 3 290 215 4 242 216 5 300 240 6 250 235 7 215 190 8 230 240 9 225 200 10 219 203 1 236 223 103 iteme