Chapter 5: Sampling Design PDF

Chapter 5 1 STAGE 2: SAMPLING DESIGN Copyright 2022 © McGraw Hill LLC. All rights reserved. No reproduction or distribution without the prior written consent of McGraw Hill LLC....

Chapter 5 1 STAGE 2: SAMPLING DESIGN Copyright 2022 © McGraw Hill LLC. All rights reserved. No reproduction or distribution without the prior written consent of McGraw Hill LLC. 5-1  Industry Thought Leadership ©McGraw Hill Access the text alternative for slide images. 5-2  Research Thought Leaders “When you sample something, you're using the crutch of borrowing chords and melodies from a song that's already great, that's already stood the test of time, that's already special.” Gerald Earl Gillum (G-Eazy), American rapper and record producer ©McGraw Hill 5-3  Exhibits ©McGraw Hill 5-4  Exhibit 5-1: Sampling Design in the Research Process ©McGraw Hill Access the text alternative for slide images. 5-5  Exhibit 5-2: Common Types of Target Populations in Business Research 1 ©McGraw Hill Access the text alternative for slide images. 5-6  Exhibit 5-3: Example Population Parameters in the Metro U Dining Study Population Parameter of Interest Data Level and Measurement Scale Frequency of eating on or near Ratio data (actual number of eating campus at a restaurant within the experiences). last 30 days Ordinal data (less than 5 times per month, greater than 5 but fewer than 10 times per month, greater than 10 times per month). Proportion of student/employees Nominal data (interested, not expressing interest in the dining interested). club Proportion of students/employees Interval data ($5–9.99, $10–14.99, spending money per person per $15–19.99, $20–24.99, $25–29.99) visit ©McGraw Hill 5-7  Exhibit 5-4: Data Types and Characteristics Data Type Data Characteristics Example Nominal Classification Respondent type (faculty, staff, student) Ordinal Classification and order Preferred doneness of steak (well done, medium well, medium rare, rare) Interval Classification, order, How rated last restaurant experience and distance (scale of 1–10; l = very poor, 10 = exceptional) Ratio Classification, order, Average amount spent per person for distance, and natural last dinner in restaurant. origin ©McGraw Hill 5-8  Exhibit 5-5: Sources of Error ©McGraw Hill Access the text alternative for slide images. 5-9  Exhibit 5-6: Types of Sampling Designs 1 Case Selection Representation Basis of Representation Basis Probability of Nonprobability Unrestricted Simple random Convenience Restricted Complex random Purposive Systematic. Judgment. Cluster. Quota. Stratified. Snowball. Double. ©McGraw Hill 5-10  Selecting a random sample is accomplished with the aid of computer software, a table of random numbers, or a calculator with a random number generator. Drawing slips out of a hat or Ping-Pong balls from a drum serves as an alternative if every Exhibit 5-7: case in the sample frame has an equal chance of selection. Mixing the slips (or balls) and returning them between every selection ensures that every case is just as How to likely to be selected as any other. A table of random numbers (such as Appendix D, Exhibit D-10) is a practical Choose a solution when no software program is available. Random number tables contain digits that have no systematic organization. Whether you look at rows, columns, or Random diagonals, you will find neither sequence nor order. Exhibit C-1 in Appendix C is arranged into 10 columns of five-digit strings, but this is solely for readability. Sample Assume the researchers want a sample of 10 from a population of 95 cases. How will the researcher begin? 1 Assign each case within the sample frame a unique number from 01 to 95. 2 Identify a random start from the random number table. Drop a pencil point-first onto the table with closed eyes. Let’s say the pencil dot lands on the eighth column from the left and 10 numbers down from the top of Exhibit C-1, marking the five digits 05067. 3 Determine how the digits in the random number table will be assigned to the sample frame to choose the specified sample size (researchers agree to read the first two digits in this column downward until 10 are selected) 4 Select the sample cases from the sample frame (05, 27, 69, 94, 18, 61, 36, 85, 71, and 83) using the above process. (The digit 94 appeared twice and the second instance was omitted; 00 was omitted because the sample frame started with 01.) Other approaches to selecting digits are endless: horizontally right to left, bottom to top, diagonally across columns, and so forth. Computer selection of a simple random sample will be more efficient for larger projects ©McGraw Hill 5-11  Exhibit 5-8: Comparison of Stratified and Cluster Sampling Stratified Sampling Cluster Sampling 1. We divide the population into a few 1. We divide the population into many subgroups. subgroups. Each subgroup has many cases Each subgroup has few cases in it. in it. Subgroups are selected according to some criterion of ease or Subgroups are selected availability in data collection. according to some criterion that is related to the variables under study. 2. We try to secure homogeneity 2. We try to secure heterogeneity within within subgroups. subgroups 3. We try to secure heterogeneity 3. We try to secure homogeneity between subgroups. between subgroups. 4. We randomly choose cases from 4. We randomly choose several within each subgroup. subgroups that we then typically study in depth. ©McGraw Hill 5-12  Exhibit 5-9: Comparison of Probability Sampling Designs 1 Type Description Advantages Disadvantages Simple Random Each population case has Easy to implement with Requires a listing of an equal chance of being automatic dialing (random- population cases. Cost: High selected into the sample. digit dialing) and with computerized voice Takes more time to Use: Moderate Sample drawn using response systems. implement. random Number table/generator. Uses larger sample sizes. Produces larger errors. Systematic Using a random start, Simple to design. Periodicity within the selects a population case population may skew the Cost: Moderate and following the Easier to use than the sample and results. sampling skip interval simple random. Use: Moderate selects every kth case. If the population list has Easy to determine a monotonic trend, a sampling distribution of biased estimate will mean or proportion. result based on the start point. ©McGraw Hill 5-13  Exhibit 5-9: Comparison of Probability Sampling Designs 2 Type Description Advantages Disadvantages Stratified Divides population into Researcher controls sample Increased error will result if Cost: High subpopulations or strata and size within strata. subgroups are selected at Use: Moderate draws a simple random Increased statistical efficiency. different rates. sample from each stratum. Provides data to represent and Especially expensive if Results may be weighted and analyze subgroups. population strata must be combined. Enables use of different created. methods in strata. Cluster Population is divided into Provides an unbiased estimate Often lower statistical Cost: Moderate internally heterogeneous of population parameters if efficiency (more error) due to Use: High subgroups. Some subgroups properly done. subgroups being are randomly selected for Economically more efficient homogeneous rather than further study. than simple random. heterogeneous. Lowest cost per sample, especially with geographic clusters. Easy to do without a population list. Double Process includes collecting May reduce costs if first stage Increased costs if (Sequential or data from any type sample. results in enough data to stratify indiscriminately used. multiphase) Based on the information or cluster the population. Cost: Moderate found, a subsample is Use: Moderate selected for further study. ©McGraw Hill 5-14  Snapshots, CloseUps, & PicProfiles ©McGraw Hill 5-15  Snapshot: How to Avoid and Correct For Problem Participants ©McGraw Hill Access the text alternative for slide images. 5-16  Snapshot: How Nielsen Recruits its TV Families Letter of invitation Special website Comprehension quizzes Home visit as screen Opt-out process Opt-in process Follow-up letter and incentive ©McGraw Hill 5-17  PicProfile: Mixed Access Recruitment Multiple means for same study Reduces non-coverage error Reduces non-sampling error Phone recruiting difficult Computer recruiting viable Completion may differ from recruiting ©McGraw Hill 5-18  PicProfile: Humanizing Interactions “I think a lot about how best to connect with people, to create experiences that make us feel more—and not less—human.” Amelia Dunlop Deloitte Digital ©McGraw Hill 5-19  Text Images ©McGraw Hill 5-20  Text Images 1 ©McGraw Hill 5-21 5-21  Text Images 2 ©McGraw Hill 5-22 5-22  Text Images 3 ©McGraw Hill 5-23 5-23  Text Images 4 ©McGraw Hill 5-24 5-24  Text Images 5 ©McGraw Hill 5-25 5-25  Text Images 6 ©McGraw Hill 5-26 5-26  PulsePoint ©McGraw Hill 5-27  PulsePoint the average number of text messages sent and received per day by Americans aged 18 to 24. 124 ©McGraw Hill 5-28 4-28  From the Headlines Discussions ©McGraw Hill 5-29  From the Headlines The milk market has been disrupted by upstarts like almond- and soy-based alternatives. Milk sales have fallen for 8 consecutive years. Some people have defected to avoid lactose; others, to avoid animal products. While milk can claim greater nutritional value and taste, that didn't keep the country's largest dairy, Dean Foods, from filing for bankruptcy protection. What research is needed to guide a turnaround for milk? Design the sampling plan for that research. ©McGraw Hill 5-30  Video Discussion ©McGraw Hill 5-31  Case Discussion ©McGraw Hill 5-32  Learning Objectives ©McGraw Hill 5-33  Learning Objectives Understand... The six tasks that comprise sampling design. The premises on which sampling theory is based. The characteristics of accuracy and precision for measuring sample validity. The two categories of sampling methods and the variety of sampling techniques within each category. The various sampling techniques within each category. ©McGraw Hill 5-34  Chapter Outline ©McGraw Hill 5-35  Sampling Design in the Research Process ©McGraw Hill Access the text alternative for slide images. 5-36  Sampling Design and Data Security ©McGraw Hill 5-37  Steps in Sampling Design Define Target Population & Case Define Population Parameters Define & Evaluate Sample Frames Define Number of Cases Define Sampling Method Define Selection & Recruiting Protocols ©McGraw Hill 5-38  Exhibit 5-2: Common Types of Target Populations in Business Research 2 ©McGraw Hill Access the text alternative for slide images. 5-39  Metro U Dining Study Who would patronize the club? How much would a they spend? What days would be most popular? What menu and service formats? How often would they use the club? Who would join the club? ©McGraw Hill 5-40  Step 2 in Sampling Design Define Target Population & Case Define Population Parameters Define & Evaluate Sample Frames Define Number of Cases Define Sampling Method Define Selection & Recruiting Protocols ©McGraw Hill 5-41  Data Types Review Data Type Data Characteristics Example Nominal Classification Respondent type (faculty, staff, student) Ordinal Classification and order Preferred doneness of steak (well done, medium well, medium rare, rare) Interval Classification, order, & How rated last restaurant experience distance (scale of 1-10; l = very poor, 10 = exceptional) Ratio Classification, order, Average $ amount spent per person for distance, & natural origin last dinner in restaurant. ©McGraw Hill 5-42  Metro U: Population Parameters 1 ©McGraw Hill 5-43 5-43  Metro U: Population Parameters 2 Population Parameter of Interest Data Level and Measurement Scale Frequency of eating on or near Ratio data (actual number of eating campus at a restaurant within the experiences). last 30 days Ordinal data (less than 5 times per month, greater than 5 but fewer than 10 times per month, greater than 10 times per month). Proportion of student/employees Nominal data (interested, not expressing interest in the dining club interested). Proportion of students/employees Interval data ($5–9.99, $10–14.99, spending money per person per visit $15–19.99, $20–24.99, $25–29.99). ©McGraw Hill 5-44  Step 3 in Sampling Design Define Target Population & Case Define Population Parameters Define & Evaluate Sample Frames Define Number of Cases Define Sampling Method Define Selection & Recruiting Protocols ©McGraw Hill 5-45  Sample Frame List of elements in population Complete and correct Error rate increases over time May include elements that must be screened out International frames most problematic ©McGraw Hill 5-46  Problems with Sample Frame Incomplete list Out-of-date list Too inclusive a list Inappropriate List ©McGraw Hill 5-47  Communities Engaged, loyal to company Digitally literate individuals Want to share ideas, concerns Two types: private and public Strength: fast, cheap answers Emerging: Use in innovation ©McGraw Hill 5-48  Metro U Sample Frame University directory? Modify directory with student enrollment additions? Modify directory with student enrollment deletions? Registrar’s list? Craft a list to include students, faculty, administration & alumni in area? ©McGraw Hill 5-49  Step 4 in Sampling Design Define Target Population & Case Define Population Parameters Define & Evaluate Sample Frames Define Number of Cases Define Sampling Method Define Selection & Recruiting Protocols ©McGraw Hill 5-50  The Basic Idea Behind a Sample ©McGraw Hill 5-51  Census versus Sample Census Sample ©McGraw Hill 5-52  When Is a Census Appropriate? ©McGraw Hill 5-53  Why Use a Sample? ©McGraw Hill 5-54  What Is a Valid Sample? ©McGraw Hill 5-55  Sources of Error ©McGraw Hill Access the text alternative for slide images. 5-56  What Is a Sufficiently Large Sample? “In recent Gallup ‘Poll on polls,’... When asked about the scientific sampling foundation on which polls are based... most said that a survey of 1,500 – 2,000 respondents—a larger than average sample size for national polls—cannot represent the views of all Americans.” Frank Newport The Gallup Poll editor in chief The Gallup Organization ©McGraw Hill 5-57  When to Use a Larger Sample? ©McGraw Hill Access the text alternative for slide images. 5-58  Step 5 in Sampling Design Define Target Population & Case Define Population Parameters Define & Evaluate Sample Frames Define Number of Cases Define Sampling Method Define Selection & Recruiting Protocols ©McGraw Hill 5-59  Exhibit 5-6: Types of Sampling Designs 2 Case Selection Representation Basis Representation Basis of Probability of Nonprobability Unrestricted Simple random Convenience Restricted Complex random Purposive Systematic. Judgment. Cluster. Quota. Stratified. Snowball. Double. ©McGraw Hill 5-60  Simple Random Advantages Disadvantages Easy to implement with Requires list of random dialing. population elements. Time consuming. Larger sample needed. Produces larger errors. High cost. ©McGraw Hill 5-61 4-61  Systematic Advantages Disadvantages Simple to design. Periodicity within Easier than simple population may skew random. sample and results. Easy to determine Trends in list may bias sampling distribution of results. mean or proportion. Moderate cost. ©McGraw Hill 5-62 4-62  Stratified Advantages Disadvantages Control of sample size in Increased error if strata. subgroups are selected at Increased statistical different rates. efficiency. Especially expensive if Provides data to strata on population must represent and analyze be created. subgroups. High cost. Enables use of different methods in strata. ©McGraw Hill 5-63 4-63  Cluster Advantages Disadvantages Provides an unbiased Often lower statistical estimate of population efficiency due to parameters if properly subgroups being done. homogeneous rather than Economically more heterogeneous. efficient than simple Moderate cost. random. Lowest cost per sample. Easy to do without list. ©McGraw Hill 5-64 4-64  Stratified and Cluster Sampling 1 Stratified Cluster Population divided into Population divided into few subgroups. many subgroups. Homogeneity within Heterogeneity within subgroups. subgroups. Heterogeneity between Homogeneity between subgroups. subgroups. Random choice of cases Random choice of from within each subgroups. subgroup. ©McGraw Hill 5-65 4-65  Stratified and Cluster Sampling 2 Stratified Sampling Cluster Sampling 1. We divide the population into a few 1. We divide the population into many subgroups. subgroups. Each subgroup has many cases Each subgroup has few cases in it. in it. Subgroups are selected according to some criterion of ease or availability Subgroups are selected in data collection. according to some criterion that is related to the variables under study. 2. We try to secure homogeneity within 2. We try to secure heterogeneity within subgroups. subgroups 3. We try to secure heterogeneity 3. We try to secure homogeneity between between subgroups. subgroups. 4. We randomly choose cases from 4. We randomly choose several within each subgroup. subgroups that we then typically study in depth. ©McGraw Hill 5-66 4-66  Area Sampling Well defined political or geographical boundaries Low cost Frequently used ©McGraw Hill 5-67  Double Sampling Advantages Disadvantages May reduce costs if first Increased costs if stage results in enough discriminately used. data to stratify or cluster the population. ©McGraw Hill 5-68 4-68  Nonprobability Samples ©McGraw Hill 5-69  Nonprobability Sampling Methods Convenience Judgment Quota Snowball ©McGraw Hill 5-70  Step 6 in Sampling Design Define Target Population & Case Define Population Parameters Define & Evaluate Sample Frames Define Number of Cases Define Sampling Method Define Selection & Recruiting Protocols ©McGraw Hill 5-71  How to Choose a Random Sample Selecting a random sample is accomplished with the aid of computer software, a table of random numbers, or a calculator with a random number generator. Drawing slips out of a hat or Ping-Pong balls from a drum serves as an alternative if every case in the sample frame has an equal chance of selection. Mixing the slips (or balls) and returning them between every selection ensures that every case is just as likely to be selected as any other. A table of random numbers (such as Appendix D, Exhibit D-10) is a practical solution when no software program is available. Random number tables contain digits that have no systematic organization. Whether you look at rows, columns, or diagonals, you will find neither sequence nor order. Exhibit C-1 in Appendix C is arranged into 10 columns of five-digit strings, but this is solely for readability. Assume the researchers want a sample of 10 from a population of 95 cases. How will the researcher begin? 1 Assign each case within the sample frame a unique number from 01 to 95. 2 Identify a random start from the random number table. Drop a pencil point-first onto the table with closed eyes. Let’s say the pencil dot lands on the eighth column from the left and 10 numbers down from the top of Exhibit C-1, marking the five digits 05067. 3 Determine how the digits in the random number table will be assigned to the sample frame to choose the specified sample size (researchers agree to read the first two digits in this column downward until 10 are selected) 4 Select the sample cases from the sample frame (05, 27, 69, 94, 18, 61, 36, 85, 71, and 83) using the above process. (The digit 94 appeared twice and the second instance was omitted; 00 was omitted because the sample frame started with 01.) Other approaches to selecting digits are endless: horizontally right to left, bottom to top, diagonally across columns, and so forth. Computer selection of a simple random sample will be more efficient for larger projects ©McGraw Hill 5-72  Ethical Issues Deception Incentives Quality ©McGraw Hill 5-73  Key Terms 1 Area sampling. Nonsampling error. Case. Population parameters. Census. Population proportion of incidence. Cluster sampling. Probability sampling. Community. Proportionate stratified sampling. Convenience sample. Quota sampling. Disproportionate stratified sampling. Sample. Double sampling. Sample frame. Judgment sampling. Sample statistics. Nonprobability sampling. Sampling. Nonresponse error. ©McGraw Hill 5-74  Key Terms 2 Simple random sample. Skip interval. Snowball sampling. Stratified random sampling. Sampling error. Systematic sampling. Systematic variance. Target population. ©McGraw Hill 5-75  Photo Attributions Slide Source Slide Source 17 ©Vasyl Shulga/Shutterstock 46 © Hero/Corbis/Glow Images 18 ©Shutterstock / Rawpixel.com 47 © one photo/Shutterstock 19 ©Shutterstock/Mangostar 48 © one photo/Shutterstock 21 ©senkaya/123RF 49 ©senkaya/123RF 22 ©Ollyy/Shutterstock 50 ©Getty Images/iStockphoto 23 ©Vasyl Shulga/Shutterstock 51 ©Design Pics / Leah Warkentin 24 ©Shutterstock / Rawpixel.com 52 ©Sjoerd van der Wal/Getty Images 25 ©Shutterstock / Rawpixel.com 55 ©Getty Images/iStockphoto 26 ©Shutterstock/Mangostar 58 ©?Didecs/Shutterstock.com 37 ©Shutterstock / Gorodenkoff 59 ©Shutterstock / Dmitry Kalinovsky 38 ©ERproductions Ltd/Blend Images L LC 67 ©Fotoa/Tetra Images/Alamy 40 ©senkaya/123RF 69 © Shutterstock / Matej Kastelic 41 ©Bill Oxford/Getty Images 70 ©rawpixel/123RF 43 ©senkaya/123RF 71 ©Shutterstock / Syda Productions 45 ©murat sarica/Getty Images 73 ©Wavebreakmedia/Shutterstock ©McGraw Hill 5-76  Chapter 5 2 STAGE 2: SAMPLING DESIGN Copyright 2022 © McGraw Hill LLC. All rights reserved. No reproduction or distribution without the prior written consent of McGraw Hill LLC. 5-77  Appendix 1 CALCULATE THE SAMPLE SIZE Copyright 2022 © McGraw Hill LLC. All rights reserved. No reproduction or distribution without the prior written consent of McGraw Hill LLC. 5-78  Metro U Random Samples of Preferred Lunch Times ©McGraw Hill Access the text alternative for slide images. 5-79  Effects on Standard Error of Mean of Increasing Precision Note: A 400 percent increase in sample size (from 25 to 100) would yield only a 200 percent increase in precision (from 0.16 to 0.08). Researchers are often asked to increase precision, but the question should be, at what cost? Each of those additional sample elements adds both time and cost to the study. ©McGraw Hill 5-80  Confidence Levels and the Normal Curve ©McGraw Hill Access the text alternative for slide images. 5-81  Standard Errors Associated with Areas Under the Normal Curve Approximate Degree Standard Error (Z) Percent of Area* of Confidence 1.00 68.27 68% 1.65 90.10 90 1.96 95.00 95 3.00 99.73 99 *Includes both tails in a normal distribution. ©McGraw Hill 5-82  Metro U: Comparison of Population Distribution, Sample Distribution, & Distribution of Sample Means ©McGraw Hill Access the text alternative for slide images. 5-83  Metro U: Estimates Associated with Various Confidence Levels Approximate Degree Interval Range of Dining Visits per of Confidence Month 68% µ is between 9.48 and 10.52 visits 90% μ is between 9.14 and 10.86 visits 95% μ is between 8.98 and 11.02 visits 99% μ is between 8.44 and 11.56 visits ©McGraw Hill 5-84  Metro U: Sampling Design Decision of Meal Frequency and Joining Constructs Metro U Decisions “Meal Frequency” “Joining” (nominal, Sampling Issues (interval, ratio data) ordinal data) 1. The precision desired and how to quantify it: 95% confidence (Z = 1.96) 95% confidence (Z = 1.96) The confidence researcher wants in the estimate ±0.5 meal per month ±0.10 (10 percent) (selected based on risk). The size of the interval estimate the researcher will accept (based on risk). 2. The expected range in the population for the question used 0 to 30 meals 0 to 100% to measure precision: Measure of Central Tendency 10 30% Sample mean. Sample proportion of population with the given attribute being measured. 4.1 pq = 0.30(0.70) = 0.21 Measure of Dispersion Standard deviation. Measure of sample dispersion. 3. Whether a finite population adjustment should be used. No No 4. Estimate of standard deviation of population: 0.5/1.96 = 0.255 0.10/1.96 = 0.051 Standard error of mean. Standard error of the proportion. 5. Sample size calculation. See formula (p. 120) See formula (p. 121) 6. Calculated sample size. n = 259* n = 81 ©McGraw Hill 5-85  Appendix 2 CALCULATE THE SAMPLE SIZE Copyright 2022 © McGraw Hill LLC. All rights reserved. No reproduction or distribution without the prior written consent of McGraw Hill LLC. 5-86 Accessibility Content: Text Alternatives for Images 5-87  Industry Thought Leadership – Text Alternative Return to parent-slide containing images. An image shows cover page of chapter 5 for Stage 2: Sampling Design. The content reads, learning objectives. After reading this chapter, you should understand: LO5-1, the six tasks that comprise sampling design. LO5-2, the premises on which sampling theory is based. LO5-3, the characteristics of accuracy and precision for measuring sample validity. LO5-4, the two categories of sampling methods and the variety of sampling techniques within each category. LO5-5, the various sampling techniques and when each is used. LO5-6, the ethical issues related to sampling design. Return to parent-slide containing images. ©McGraw Hill 5-88  Exhibit 5-1: Sampling Design in the Research Process – Text Alternative Return to parent-slide containing images. The process starts with Investigative Questions, and flows to Sampling Design, which encompasses the remainder of the flow. The inner flow starts with Define Target Population & Case, then to Define Population Parameters, then to Define Number of Cases. From there a choice is made to Choose Sample or Choose Census, then Define & Evaluate Sample Frames. From there a side flow leads to Identify Existing Sample Frames, then to Evaluate Existing Sample Frames, with the choice to either Accept and flow to Select Sample Frames or Reject and flow to Modify or Construct Sample Frames, which then flows to Select Sample Frames, which leads back to the main flow. Define Sampling Method has two categories: Nonprobability and Probability - both flow to Define Selection & Recruiting Protocols, and finally to Draw Cases. Return to parent-slide containing images. ©McGraw Hill 5-89  Exhibit 5-2: Common Types of Target Populations in Business Research – Text Alternative 1 Return to parent-slide containing images. The diagram is made up of a circle surrounded by six smaller circles. The large circle in the middle is labeled Target Population. The surrounding circles are: People, Organizations, Events, Objects, Settings, and Texts & Records. Return to parent-slide containing images. ©McGraw Hill 5-90  Exhibit 5-5: Sources of Error – Text Alternative Return to parent-slide containing images. The sources of error are sampling error and sample estimate is not the population parameter. Sampling error: non-sampling error. Sample estimate is not the population parameter: inappropriate sampling method, errors in measurement instrument, and behavioral effects. Return to parent-slide containing images. ©McGraw Hill 5-91  Snapshot: How to Avoid and Correct For Problem Participants – Text Alternative Return to parent-slide containing images. This diagram shows a few ways by which Research Participants can Avoid problems like Noncompliers, to Avoid Frequent, Repeaters, Rush throughs, Fakes, Liars, and Straightliners. Return to parent-slide containing images. ©McGraw Hill 5-92  Sampling Design in the Research Process – Text Alternative Return to parent-slide containing images. The process starts with Investigative Questions, and flows to Sampling Design, which encompasses the remainder of the flow. The inner flow starts with Define Target Population & Case, then to Define Population Parameters, then to Define Number of Cases. From there a choice is made to Choose Sample or Choose Census, then Define & Evaluate Sample Frames. From there a side flow leads to Identify Existing Sample Frames, then to Evaluate Existing Sample Frames, with the choice to either Accept and flow to Select Sample Frames or Reject and flow to Modify or Construct Sample Frames, which then flows to Select Sample Frames, which leads back to the main flow. Define Sampling Method has two categories: Nonprobability and Probability - both flow to Define Selection & Recruiting Protocols, and finally to Draw Cases. Return to parent-slide containing images. ©McGraw Hill 5-93  Exhibit 5-2: Common Types of Target Populations in Business Research – Text Alternative 2 Return to parent-slide containing images. The diagram is made up of a circle surrounded by six smaller circles. The large circle in the middle is labeled Target Population. The surrounding circles are: People, Organizations, Events, Objects, Settings, and Texts & Records. Return to parent-slide containing images. ©McGraw Hill 5-94  Sources of Error – Text Alternative Return to parent-slide containing images. The diagram is made up of two rows of rectangles. The first row has "What" written outside the first rectangle, which is labeled Sampling Error. The next rectangle is three times as wide and is labeled Non-Sampling Error. The second row has "Why" written outside the first rectangle, which is labeled Sample estimate is not the population parameter. This is followed by three rectangles: Inappropriate Sampling Method, Errors in Measurement Instrument, and Behavioral Effects. Return to parent-slide containing images. ©McGraw Hill 5-95  When to Use a Larger Sample? – Text Alternative Return to parent-slide containing images. The larger sample can be used for population variance, desired precision, small error range, confidence level, and number of subgroups. Return to parent-slide containing images. ©McGraw Hill 5-96  Metro U Random Samples of Preferred Lunch Times – Text Alternative Return to parent-slide containing images. This image displays Random Samples of Preferred Lunch Times. Population of preferred lunch times includes samples like n1, n2, n3, and n4. Return to parent-slide containing images. ©McGraw Hill 5-97  Confidence Levels and the Normal Curve – Text Alternative Return to parent-slide containing images. Shown is a Normal bell curve with standard deviation divisions marked. The mean is 0, and one standard deviation above and below are +/− 1 sigma-sub-xbar, which consists of 68% of the data. Two standard deviations above and below are +/− 1.96 sigma-sub-x bar, which consists of 95% of the data. Return to parent-slide containing images. ©McGraw Hill 5-98  Metro U: Comparison of Population Distribution, Sample Distribution, & Distribution of Sample Means – Text Alternative Return to parent-slide containing images. Part A, the Population distribution shows a right-skewed curve that appears to start two standard deviations below the mean, peaking in the range of one standard deviation below, then descending through to three standard deviations above the mean. Part B shows the Distribution of means from repeated samples of a fixed size, n=64. 68% of the area is shaded. The distribution of sample means is a Normal curve with no skew. Part C is a sample distribution, which consists of a series of rectangles that make up a slightly right-skewed bell curve. Return to parent-slide containing images. ©McGraw Hill 5-99