Sampling in Psychology (Psych 217) - Nov 2021 Notes PDF

Population, generalizability, and sampling: Nov.21 Population vs sample? Population = group of people interested in studying Sample = draw one sample from population generalizing ensure its representative of population (bigger sample better as it relates to population) **only draw one samp...

Population, generalizability, and sampling: Nov.21 Population vs sample? Population = group of people interested in studying Sample = draw one sample from population generalizing ensure its representative of population (bigger sample better as it relates to population) **only draw one sample for each study Random assignment: What? Every participant in your sample has equal opportunity to be assigned in study. Advantages: Disadvantages: Random selection: (random sampling) What? Everyone in population wish to generalize have equal chance to be in study. Advantages: Disadvantages: Random selection: Is random selection truly random? Sampling: What is a problem with sampling? Sampling bias. Sampling bias: The sample chosen may be biased. Impact on generalizability How sample obtained How can sampling go wrong? A. Sampling population B. Sampling frame = people you can invite (not guaranteed) C. Say yes or no D. Response rate = if participants complete the study (not just start) How can go wrong? Eventually the participants of the sample that complete the research may have common features that may not be threat to generalizability. Ex: sample of diabetic people. Only low severity diabetic participants completed. Why? Severe ps unable to complete due to medical reasons, death or other. So sample reduced and biased to only low severe level of diabetes. Basic and applied research: Basic = theory driven, general knowledge Random assignment and internal validity Applied = find a solution to problems (specific group of people targeting) Random selection more important When to use random selection and random assignment? True experiment: must have random assignment Not necessarily random sampling Random selection not always necessarily. When needed? Generalize a large population. (Ex. Conducting a survey, consumer research, election polling, teacher evaluation) Sampling methods: 1. Probability sampling (3): 2. Non-probability sampling: no randomness whoever comes first 1. Probability sampling: A. Simple random sampling: Everyone in population has equal chance to be selected for sample.  B. Stratified random sampling: Pick randomly but maintain the proportions. (Of total divisions) Sampling within each sub-group. Randomly sample from each sub-group equally. Each group part of sampling. C. Cluster sampling: When unable to do stratified or probability. (Dont have a list of people) Non-probability sampling: 1. Convenience sampling: (used a lot in UBC) Select whoever is available to participate. May be biased. Ex: putting a marketing in the nest (people who go in the nest may participate people who don’t go to the nest will not participate) 2. Purposive sampling: Selected for specific features of interest. Not generalizing. (Understand ones rational) Used mostly in qualitative research. (small #ps) 3. Quota sampling: Sampling of subgroups proportionate to population.  Ps based on availability. (Whoever wants to be selected) **a mix of purposive and convenience **not randomly selected (opposite to stratified) whoever comes in first Generalizability: Is sampling a concern for psychology? Large majority of samples are convenience samples. **convenience sample = particular participants accessible (ex: university students participating in HSP) The question arises, are these convenience samples representative of a larger population? Sub-disciplines of psychology must be concerned with generalizability. Basic research and generalizability: Should we use convenience samples? Controversial – favour of generalization Students more diverse Need a good reason why results shouldn’t generalize beyond these models **seek to replicate effects Controversial- against generalization Ethical issues : justice Students have knowledge of psychology Unrepresentative sample can’t be generalized Basic processes can be culture-specific Problem with sampling in western geography:  We are only 10% of the world’s population. Opinions on convenience sampling: What is our response to problems of convenience sampling? 1. Does not invalidate but needs replication 2. A lot of basic processes not need generalizability (ex:eyeball same for all) 3. College populations are diverse What is the solution? Collect samples outside of convenience sampling. Statistically analyse their differences. How? 1. Keep data of convenience sampling 2. Collect cross cultural data 3. Replicate and compare 4. Understand the differences. Is the theoretical justification? Generalizability and people: There are differences in sexes and genders. Solution: include difference sexes and genders. Look for interactions. Ex: many research on men OD but no women. Generalizing to other studies: Replication Can others get same effects as you? There are two types of replications. 1. Exact (direct) replication: use exact operationalization. 2. Conceptual replication: same construct but similar operationalization. Is effect specific to IV/DV or generalize to other similar techniques? May be more useful than direct replication. Why do direct replication? Direct replication does not give novel findings. But it adds more information on a finding. (Reliable? Disapprove? ….) Generalizing to other literature: Are some effects general across the literature/other studies? Literature review: What? Type of article in which don’t conduct study. Gather published evidence Summarize for others Meta-analysis: What? Combine studies from many studies in an area. How? Use statistical techniques. Ex: influence of video games on children’s aggression and adolescents Generalizability to real world: Does effect generalize across different situations or real setting? Purpose: study behavior in real world. Problem: Behavior studied in artificial settings. Solutions: Study in real world o Have high ecological validity (done in real world) o Very hard Study in an experimental realism o Setting induces a realistic experience o Artificial o Could have high ecological validity **ecological validity = high validity as done in real world