Nonrandom sampling (1)
- 1. Systematic Convenience Purposive Population Genaralizability Ecological Genearalizability
- 2. Every nth individual in the population list is selected. eg: the principal of a school has 1000 students, she wants to know how students feel about the new menu at cafeteria, so the principal:- 1. get a list of the student’s name (in alphabetical order). 2. she select every 10th student until she has a sample of 100 students to be interviewed.
- 3. PERIODICITY-a marked bias sample caused by the arrangement pattern of individual on the list accidentally coincides with the sampling interval. eg: grouped by gpa(grade point average), high/low interval: only good/poor graded get chosen. Therefore, researchers should carefully examine the list and avoid bias.
- 4. CONVENIENCE SAMPLING A convenience sample is a group of individuals who conveniently available. samples: 1. first 50 people who walk in. 2. interview people at downtown. 3. two front rows students.
- 5. Bias- 1. not downtown = not interviewed 2. unwilling = not interviewed 3. willing = strong opinion 4. interview time = at work Ingeneral, convenience sample cannot be considered representative of any population and should be avoided.
- 6. PURPOSIVE SAMPLING Based on previous knowledge of a population and the specific purpose of the research,researcher use personal judgement to select a sample. eg:SUITABLE SAMPLE a. 2 good students,2 average & 2 weak b. sample from Retired Workers Association eg: sample know the target A. people in charge of school B. people with experience THUS, the only challenge with purposive sampling is the researcher’s previous knowledge must be thorough.
- 7. Sample should be as large as a researcher can obtain with a reasonable expenditure of time and energy. Ideals; 100 samples for DESCRIPTIVE STUDY 50 samples for CORRELATION STUDY 30 samples in each group for EXPERIMENTAL STUDY& CAUSAL-COMPARATIVE STUDY
- 8. External Validity = the result of a study can be generalised from sample to population. Ecological generalizability = result of a study can be generalised to other settings. Population generalizability = result of study can be generalised to the intended population -representativeness: relevant characteristics Overlooked “method”-random student=random teacher=random result. Lost subject effect representativeness, researcher who lost 10 % sample are advised to acknowledge this limitation.
- 9. 1. researcher should describe the sample thoroughly; reader judge the result validity. 2. Replication; repeat the study on different group or situation. If result is the same; generalise it.