Research Methodology 3 (Sampling) PDF

Research Methodology 3 (Sampling) By Dr. Ahmed Ghandour  Sampling Definition: It is the process of selecting a number of study units from a defined study population. Aim of sampling: To reduce: cost Manpower Time...

Research Methodology 3 (Sampling) By Dr. Ahmed Ghandour  Sampling Definition: It is the process of selecting a number of study units from a defined study population. Aim of sampling: To reduce: cost Manpower Time Sample size depends on: Disease prevalence Availability of cost, man power and time Aim of the study. SAMPLE SIZE Calculated by special equations for each study design. Can e automatically calculated by using statistical programs (e.g. Epi info). Types of sampling: Purposive (selective) Random  Samples chosen according to person's own judgment Each unit of population has an equal chance  Can't be considered as representative samples or probability of being chosen  Common used in pilot study Types: Types: 1. Simple random sample 1. Convenience sampling 2. Systematic random sample 2. Quota sampling 3. Stratified random sample 3. Snow ball sampling 4. Multistage random sample 4. Self selection sampling 5. Cluster (simple or multistage) random 5. Purpusive sampling sample SAMPLING TECHNIQUES A. Purposive (selective) / probability sampling: Samples chosen according to person's own judgment Can't be considered as representative samples 3 types: 1. Convenience sampling 2. Quota sampling 3. Snow ball sampling 4. Self selection sampling 5. Purposive sampling 1. Convenience sampling Usually used when you take a number of participants less than the calculated sample size because of small budget, limited time or resources, …etc. 2. Quota sampling: e.g.: if we like to make a survey for student satisfaction about medical education in a faculty of medicine. Total number of students = 2000 students divided as follows: Grade 1: 800 students (40 %) Grade 2: 400 students (20 %) Grade 3: 400 students (20 %) Grade 4: 200 students (10 %) Grade 5: 100 students (5 %) Grade 6: 100 students (5 %) So, if the sample size is 200 students (10 % of all students) → the sample must include 40 students from grade 1, 20 students from grade 2 and so on. You keep inviting the participants to the research until each of these quotas are filled. Snow ball sampling: It is used when the health problem under studying is hard to reach. e.g.: HIV/AIDS among commercial sex workers or IV drug addicts, homeless individuals,…etc. SAMPLING TECHNIQUES B. Random (non- probability sampling): Each unit of population has an equal chance or probability of being chosen 5 types: 1. Simple random sample 2. Systematic random sample 3. Stratified random sample 4. Multistage random sample 5. Cluster random sample 1. Simple random sample - In this type → the population should be homogenous and has a definite sampling frame (list of identification of the individuals in the population to be sampled) - After deciding the size of required sample → we can use the tables of random sampling numbers or the lottery method to chose the sampling units 2. Systematic random sampling - After a population are arranged in list or form → sample taken as the following examples: e.g. if sample 14 from 140 → initial selection→ 4, 14, 24, 34 …etc. 3. Stratified random sampling - Used in non homogenous population especially in variables known to affect the subject we study e.g. age on height. - In this condition → we divide the population into subgroup (strata) and from each stratum → we take a sample randomly after determining its size. e.g., nj = n x (n`/N) nj = sample size of each stratum n= sample size n`=number of population of that stratum N = total population. 4. Multistage random sampling - In this type → sampling frame is divided into a population of "1st stage sampling units" of which a 1st stage sample is taken by random sampling → each of the selected 1st stage unit is subdivided into → "2nd stage sampling units" of which a sample is taken … → 3rd stage … may be done if required. 5. Cluster sampling - In this type → population are clustered in form of villages or clusters of house holds. - This cluster sample → liable to bias if the disease under study is clustered in the population.