SAMPLING TECHNIQUES IN RESEARCH - types & procedure

Samling methods in research

1. SAMPLING METHODS A.Jeevarathinam Assistant Professor of Home Sciecne V.V.Vanniaperumal College for Women Virudhunagar

2. Sample A sample is defined as a smaller set of data that a researcher chooses or selects from a larger population by using a pre-defined selection method. These elements are known as sample points, sampling units, or observations. Creating a sample is an efficient method of conducting research .

3. Sampling units A sampling unit is the building block of a data set; an individual member of the population, a cluster of members, or some other predefined unit. sampling unit is the individual element or set of elements considered for selection in the sampling process . The target population is the collection of sampling units.

4. Sampling frame   A sampling frame is a list of all target populations from which a sample is to be taken so that it is comprehensive and up-to-date. Examples of sampling frames are employee databases, electoral rolls, telephone books, specialist directories.

5. SAMPLING METHODS Sampling is a fundamental technique in research that involves selecting a smaller group, or sample, from a larger population to represent the whole. The sampling technique, or the sampling method, is a statistical approach used for selecting a representative sample from a population.

6. Types of Sampling Method

7. PROBABILITY VS NON PROBABILITY Probability sampling uses random sampling techniques to create a sample. In this method, every individual of the population has an equal chance of being selected.  The non-probability sampling method is a technique in which the researcher selects the sample based on subjective judgment rather than the random selection. In this method, all elements do not have an equal chance of being selected. 

8. It is an entirely random method of selecting the sample. A simple random sample takes a small, random portion of the entire population to represent the entire data set, where each member has an equal probability of being chosen. A sampling error can occur with a simple random sample if the sample does not end up accurately reflecting the population it is supposed to represent. Simple random sampling

9. Simple random sampling steps

10. Simple random sampling An example of Simple random sampling is when a market research firm is hired by a car manufacturer to gauge public opinion on their new electric vehicle model. The firm uses a list of 500,000 registered drivers, each with a unique number, and randomly selects 2,000 to survey using statistical software. This approach helps obtain an unbiased view of driver preferences and purchase likelihood.

11. Lottery method

12. Random number table

13. Physical methods Simple, early methods of random selection may use dice, flipping coins, or spinning wheels.

14. Cluster random sampling In cluster sampling, researchers divide a population into smaller groups known as clusters.  The division depends on the apparent or easily identifiable characteristics. The formation of clusters can be based on geographical proximity or common characteristics, which are correlated to the main variable of the study. They then randomly select among these clusters to form a sample. Cluster sampling is a method of probability sampling that is often used to study large populations, particularly those that are widely geographically dispersed.

15. Single-stage Cluster Sampling Example: An e-commerce company studying shopping behavior across the United States might randomly select a few states, like California, Texas, and New York, and collect data from all customers within those states.

16. Two-stage Cluster Sampling Example: A video streaming platform conducting a survey on user preferences across regions might first randomly select cities or metropolitan areas (clusters). Then, within each chosen city or metro area, they would randomly select a set number of subscribers. For instance, in the United States, they might randomly choose 15 major cities like New York and Los Angeles. Within each city, they could then select 500 subscribers to participate in the survey.

17. Multi-stage Cluster Sampling Example: A global social media platform wanted to study the impact of its ad targeting algorithms on user engagement across regions. They employed a multi-stage cluster sampling approach: Randomly selected 10 countries from global operations. Within each country, randomly chose 5 states/provinces/regions. From each state/province/region, randomly selected 20 cities/towns. Randomly sampled 100 active users from each selected city/town.

18. Examples of Cluster Random Sampling Performance of Student Studying in Schools Imagine assessing the reading skills of fifth-graders in a city. Divide the city into clusters: Group schools by district, geographic proximity, or similar demographics. Randomly select a few clusters: Choose a representative sample of schools from each group. Sample students within chosen clusters: Test a random sample of fifth-graders within each selected school. By analyzing the reading skills of students in the chosen clusters, we can draw inferences about the reading proficiency of fifth-graders across the entire city with improved efficiency and cost-effectiveness.

19. Examples of Cluster Random Sampling Evaluating Public health programs in Rural areas: Consider assessing the effectiveness of a new vaccination program in remote villages. Reaching every village for individual interviews might be challenging. Instead, we can opt for cluster sampling : Group villages into clusters: Based on geographic proximity, accessibility or population size, divide the villages into manageable groups. Randomly select a few clusters: Choose a representative sample of clusters from the whole set. Sample households within chosen clusters: Survey households within each selected cluster to gather data on vaccination coverage and program awareness.

20. Steps to Cluster Random Sampling To perform cluster random sampling, we ca use the following process: Step 1:  Clearly identify the entire group of interest. Step 2:  Divide the population into distinct and mutually exclusive clusters. Step 3:  Use a random sampling technique to choose specific clusters. Step 4:  Include all individuals within the selected clusters in the sample. Step 5:  Apply the study's methodology to collect data from all elements within selected clusters. Step 6:   Analyze collected data considering the cluster structure. Step 7:  Communicate study findings including details about the cluster random sampling method used.

21. Stratified random sampling Stratified sampling is a method of selecting a sample in which the population is first divided into homogeneous subgroups, or strata, based on certain characteristics that are relevant to the study.  These characteristics could be age, gender, income level, geographic location, or any other factor deemed significant. Samples are then randomly selected from each subgroup to ensure proportional representativeness.

22. Formula of Stratified Sampling The formula for calculating the sample size in stratified sampling is given by: s=(S/N)*n Where: s = sample size of stratrum S = population size of stratrum N = total population size n = total sample size

23. Stratified Sampling Example Let's assume we have a total population of 10,000 internet users, and we want to survey 1,000 of them. Here's how we can calculate the sample size for each age group: N: 10,000 internet users , n: 1,000 internet users , S: Teenagers (2,000 users), Young adults (3,000 users), Adults (3,500 users), Seniors (1,500 users) For teenagers: 2,000/10,000x1,000=200 For young adults: 3,000/10,000x1,000=300 For adults: 3,500/10,000x1,000=350 For Seniors: 1,500/10,000x1,000=150 So, based on the formula, we will select randomly and survey 200 teenagers, 300 young adults, 350 adults, and 150 seniors.

24. Types of stratified random sampling

25. Proportional Stratified Sampling: In proportional stratified sampling, the size of the sample taken from each group (or stratum) is directly proportional to the size of that group in the overall population. This means that larger groups contribute more to the sample, while smaller groups contribute less. Suppose we're conducting an online survey to understand music preferences among different age groups. If our population is 1,000, which consists of 60% teenagers and 40% adults. Then in proportional stratified sampling, we wanted a sample of 500 participants, we would survey 300 teenagers and 200 adults.

26. Disproportional Stratified Sampling Disproportional stratified sampling involves intentionally sampling different proportions from each group, regardless of their size in the population. This method is useful when certain subgroups are of particular interest and need more representation in the sample. Let's say we're researching online shopping habits, specifically focusing on luxury purchases. While only 5% of the population makes luxury purchases, we want to ensure a larger representation of this group in our sample. So, in disproportional stratified sampling, we might choose to survey 20% of luxury shoppers, 30% of moderate spenders, and 50% of budget-conscious shoppers to gather more insights into luxury purchasing behavior .

27. Example: Step 1:  You segment individuals based on income brackets (e.g., low-income, middle-income, high-income). Step 2:  If you're conducting a study on consumer spending habits across income brackets and the population comprises 40% low-income earners, 30% middle-income earners, and 30% high-income earners, allocate sample sizes accordingly to mirror these proportions. Step 3:  After stratifying the population into categories, apply simple random sampling to randomly select participants from each group. Alternatively, employ systematic sampling by selecting every nth individual from a sorted list of population members within each age category. Step 4:  Researchers conduct online surveys to collect spending habits data from each income group. Step 5:   Analyze spending patterns across income brackets to identify trends. This step helps in interpreting findings and recognizing patterns within each stratum and across the entire population.

28. Systematic sampling Systematic sampling stands as a cornerstone probability sampling method in statistics and research, facilitating the selection of random samples from larger populations with a fixed interval.  It involves selecting every nth member from the population after establishing a random starting point, ensuring an equitable chance for each member to be included in the sample.

29. Systematic Random Sampling Example: Suppose a market research company wants to survey its customers about their satisfaction with the website's user experience. They have a customer database of 10,000 email addresses. Using systematic random sampling, they decide to select every 50th customer from the list, starting from a randomly chosen starting point. They send the online survey to the selected customers.

30.  Linear Systematic Sampling

31. Example: Imagine a researcher conducting an online survey to understand the preferences of users on a particular e-commerce website. They want to ensure that their sample represents the entire user population systematically. Organize Users:  The researcher arranges all registered users systematically, perhaps by registration date or user ID. Determine Sample Size:  Let's say they want to survey 150 out of 1500 users. Calculate Sampling Interval:  Using the formula k=N/n, where N=1500 and n=150, the sampling interval (k) is 10 (In cases where k isn’t an integer, choose the closest integer to N/n.) Select Initial Starting Point:  The researcher randomly chooses a number between 1 and 10, let's say they pick 5. Choose Sample Members:  Starting from the 5th user, they select every 10th user thereafter. Repeat Process to select the remaining individuals to the sample.

32. Circular Systematic Sampling

33. Example: We have a population of 14 individuals numbered from 1 to 14. We want to select a sample of 4 individuals using circular systematic sampling. Calculate the sampling interval: k = 14/4 = 3 (choose the closest integer to N/n) Start randomly between 1 to 14: Let's say we randomly start at individual number 4. Create samples by skipping through k units: We select individuals 4, 7, and 11. Repeat until you select members of the entire population: Since we have only two individuals in our sample, the process ends here. However, we would continue until all 14 individuals are sampled, resulting in 14 samples, if we wanted to sample the entire population.

34. Convenience Sampling Convenience sampling is a type of non-probability sampling method in research where the sample is drawn from the part of the population that is readily available and easiest for the researcher to access.  This selection is often influenced by factors such as geographical proximity, availability at a specific time, or willingness to participate. Convenience sampling is also known as grab sampling, accidental sampling, or opportunity sampling.

35. How a Tech Startup Uses Convenience Sampling Online Identify the Target Population:  The tech startup targets tech-savvy users of similar mobile apps to gather feedback on its beta version. Increase Sample Size:  They promote a feedback link within the app and on social media for a week to maximize participant numbers. Include Diverse Questions:  The online survey features both qualitative (e.g., feedback on user interface) and quantitative questions (e.g., satisfaction ratings). Validate Results:  They regularly monitors survey responses to ensure data reliability and consistency. Complement with Probability Sampling:  The startup plans to use stratified sampling, randomly selecting a subset of users from their app’s user base.

36. Quota sampling  Quota sampling is a non-probability method where researchers divide the population into subgroups (quotas) and select participants from each subgroup to ensure representation based on characteristics like age, gender, or income. Unlike probability sampling methods, where individuals are randomly selected, quota sampling involves setting predefined quotas for specific characteristics and selecting individuals who meet those criteria until the quotas are filled.

37. 1. Define the population and research objectives  Potential smartphone buyers in a large city.  2 : Identify Strata   Identify important characteristics based on the research objectives, such as age, gender, and income level.    Age, Gender, and Income (e.g., 18-24, Male, Low Income).  3 : Set Quotas   Based on the population distribution or research needs, assign quotas (target numbers) for each subgroup. The quotas should reflect the proportion or importance of each group within the overall population.    50 participants in the 18-24, Male, Low Income group, 60 in the 25-34, Female, etc.  4 : Recruit Participants   Use various recruitment methods (online surveys, in-person interviews, social media) to gather participants for each quota.  Social media ads for 18-24 males, in-store surveys for females 25-34.  5 : Collect Data   Ensure that the number of participants meets the quotas defined for each stratum.    Record responses until quotas are met for each stratum.  6 : Monitor & Adjust   As data is being collected, monitor the progress to ensure each quota is being filled. If one quota is underrepresented, adjust recruitment methods.    Adjust methods if some quotas are not being met.  7 : Analyze Data   Once quotas are met, analyze the data for differences across the strata.    Compare preferences across the age, gender, and income strata.  8 : Report Findings   Present the findings based on the quotas.  Report that younger, low-income males prefer budget models, while older males prefer premium models. 

38. Example:  A brand wants to test the popularity of a new product across different income levels. Steps: Identify the income groups: low, middle, and high income. Decide on the number of participants for each group: 50 from low-income, 50 from middle-income, and 50 from high-income. Ensure each income group is represented equally in the sample, even if the population proportions differ. Gather opinions from each income group to understand the product's appeal across different economic levels.

39. Judgmental Sampling Judgmental sampling, also called purposive sampling or authoritative sampling, is a non-probability sampling technique in which the sample members are chosen only on the basis of the researcher’s knowledge and judgment. 

40. Snow ball sampling

41. Snow ball sampling Also known as  chain sampling  or  network sampling , snowball sampling begins with one or more study participants. It then continues on the basis of referrals from those participants. This process continues until you reach the desired sample, or a saturation point.

42. Linear snowball sampling  relies on one referral per participant. In  exponential non-discriminative snowball sampling , the first participant provides multiple referrals. In other words, the researcher recruits the first participant, and this participant in turn recruits several others.  In exponential discriminative snow ball sampling , participants give multiple referrals. However, the researcher screens those referrals, choosing only those who meet specific criteria to participate in the sample.

43. THANK YOU