# Chapter 5 Sampling Chapter Objectives Definition of sampling

Chapter 5 – Sampling

Chapter Objectives Definition of sampling Understand the difference between probability and nonprobability sampling ○ Sample frame ○ Strata sampling vs. Quota sampling Sampling in qualitative research

What is sampling? Definition ○ The process by which researchers select a representative subset or part of the total population that could be studied for their topic so that they will be able to draw conclusions about the entire population Advantage ○ Smaller number of elements (e. g. , people, organizations) makes the research more manageable, time efficient, less costly, and more accurate ○ To cover China, you need 1. 3 billion questionnaires

Sample? Sample ○ Should be as much as representative of the entire population Large Sample ○ Can reduce sampling error ○ Not always represents entire population • Example – Sample of 2 million failed to predict Landon’s victory over Roosevelt

Things to consider… Consideration Target Population • • Accessible Population • General Characteristics Particular Characteristics Geographic / Time constraints Example • • Chinese business travelers Male, aged between 25 -45 • Business travellers flying from Shanghai Pudong International Airport Travellers flying in May and June 2008 Using Business Class lounges • Inclusion Criteria • Exclusion Criteria • • Specific criteria by which participants will be selected for the study Factors that will affect data collection Ethical issues • • Frustrated delayed travellers. Non-English speaking Non consenting in tape recording the interview

Sampling Frame ○ It is the source material or device from which a sample is drawn ○ It is a list of all those within a population who can be sampled ○ Includes individuals, households, institutions, etc. Example of Sampling Frame ○ Student lists in a high school ○ Loyalty program participants in a hotel ○ Lists of high schools in U. K.

Different Types of Sampling Frame Yes Probability Sampling No Non-probability sampling

I. Probability Sampling q Need to have information about population

Probability Sampling Selection bias ○ Certain members of the population under study are underrepresented or overrepresented To minimize selection bias ○ Target population has an equal opportunity for selection ○ Probability sampling

1. Simple Random Sampling #1 All members of the population under study have the same chance (probability) of being selected You will need ○ Random number table Example ○ You have a sample frame (e. g. , detailed contact information) of 1000 customers of a hotel ○ You are sampling 100 customers to interview

1. Simple Random Sampling #2 Utilization of random number table 39634 62349 74088 65564 16379 19713 39153 69459 17986 24537 14595 35050 40469 27478 44526 67331 93365 54526 22356 93208. . . 65358 70469 87149 89509 72176 18103 55169 79954 72002 20582 ○ Since you are collecting 100 customers, you should use last three digits (red color) ○ For instance, 39634 suggests you to use 634 th customer on the list, 62349 suggests to use 349 th customer on the list ○ Repeat this procedure until you reach 100 customers to interview

2. Systematic Sampling Select nth member of the targeted population ○ Giving out a questionnaire to each 8 th passenger to check in for a specific flight How to determine n (sampling fraction)? ○ If the passengers in the specific flight is 400, and the airline wishes to collect 50 samples, then the sampling fraction is 8 ○ A researcher should choose a number n between 1 and 8 ○ Give out surveys to nth customer until he collects 50 samples

3. Stratified Sampling #1 Population is divided into homogeneous and mutually exclusive groups called “strata” ○ Age, gender, market segment Procedure ○ Divide population into different strata ○ For each strata, collect samples by simple random sampling or systematic sampling

3. Stratified Sampling #2 Population Strata #1 Strata #2 Data collection by Simple random sampling Systematic sampling Strata #3 Strata #4 Data collection by Simple random sampling Systematic sampling

4. Cluster Sampling #1 Frequently used when population is geographically diverse ○ Divide population into mutually exclusive subsets Two types of cluster sampling ○ One stage cluster • Randomly select subsets • Sample entire participants in the selected subset ○ Two stage cluster • Randomly select subsets • Conduct simple random sampling for participants in the selected subset

4. Cluster Sampling #2 Graphical Illustration of Cluster Sampling ○ One-stage cluster sampling Randomly select subsets (Red Rectangles) Sample entire participants in the red rectangles

4. Cluster Sampling #3 Graphical Illustration of Cluster Sampling ○ Two-stage cluster sampling Randomly select subsets (Red Rectangles) Conduct simple random sampling for the participants in each red rectangles

II. Non-Probability Sampling q You don’t need information about population

Non-Probability Sampling Definition ○ Sampling where it is not possible to specify the probability that any person or other unit on which the survey is based will be included in the sample (Smith, 1983) Advantages ○ Select samples purposively ○ Enable researchers to reach difficult-to-identify members of the population Disadvantages ○ Difficult to make valid inference about the entire population because the sample selected is not representative

1. Convenience Sampling Definition ○ Members of the population are chosen based on their relative ease of access ○ Interviews can be determined by what was convenient, not by random ○ Also called as haphazard or accidental sampling

2. Judgemental Sampling Definition ○ Another form of convenience sampling where participants are handpicked from the accessible population ○ Researcher select participants that are representative of the entire population ○ Very subjective sampling method (can be biased)

3. Quota Sampling Definition ○ Similar to stratified sampling, population is divided into mutually exclusive subsets ○ Then judgement is used to select the participants from each stratum based on specified proportion Stratified vs. Quota ○ Stratified sampling requires random sampling for each strata ○ Quota sampling does not require random sampling for each quota

4. Snowball Sampling Definition ○ A non-probability sampling technique where existing study subjects recruit future subjects from among their acquaintances The sample group grows like a rolling snowball ○ The first respondent refers a friend ○ The friend also refers a friend, and so on

5. Self-Selection Sampling Definition ○ A non-probability sampling in which individuals identify their wish to participate in the study Procedure ○ Recruit participants via advertisement or posting on articles, magazines, journals, e-mails, newsgroups, etc. ○ People who have interest (volunteers) in the survey will participate

III. Other Things to Consider q Sample size q Non-response q Sampling in qualitative research

1. Sample Size #1 Arbitrary in nature ○ The larger the sample you study, the more likely your findings will be representative of population and the more correct your inferences about data Sampling Error (Level of Precision) ○ Range in which the true value of the population is estimated to be ○ If you find that 60% of hotel guests in a sample agree with the proposed new restaurant concept with a precision rate of ± 5%, then you can conclude that between 55% and 65% of hotel guests agree with the new concept

1. Sample Size #2 Level of Confidence (Risk Level) ○ Based on the notion that the larger the size of a sample from a population, the closer the average value of the answers will be to the true population value ○ Researcher prefer 95% confidence level • If the sample was selected 100 times, at least 95 of these times it would represent the entire population

2. Non-Response #1 Response rate ○ Ratio of number of participants who have actually taken part in your study divided by the number of people in your sample ○ Low response rate may have negative effect on the credibility of findings, because the sample will be less likely to present the overall population ○ 15% ~ 20% will be considered acceptable response rate

2. Non-Response #2 Two Different Response Rate ○ Total Response Rate • Completes: # of participants who completed survey • Not Qualified: # of participants who do not qualify ○ Active Response Rate • Not Contacted: # of participants included in sample but cannot be contacted

3. Sampling in Qualitative Research Goal of Qualitative Research ○ Gaining in-depth understanding about the topic or phenomenon under study and extrapolate the findings beyond the material in hand Sampling in Qualitative Research ○ Must serve the purpose of in-depth understanding ○ Samples must be selected for the information-rich data ○ Similar to the judgemental or purpose sampling

3. 1. Theoretical Sampling Based on grounded theory ○ Researcher simultaneously collects, codes, and analyses the date in order to decide what data to collect next • Researcher starts with purposive sampling by selecting the sample seem more appropriate to describe the phenomenon under study • Researcher moves into theoretical sampling, which involves selecting samples that are more appropriate to understand theories found in earlier stage ○ Sample size is not defined from the outset of investigation ○ Sampling terminates when the researcher feels that enough data was collected

3. 2. Maximum Variation Sampling Type of Purposeful Sampling ○ Central themes that cut across a great deal of participant variation ○ Trying to achieve representativeness of the total population, not by equal probabilities but by including a wide range of extremes

- Slides: 32