Final Exam Flashcards
Population sample
Entire population is used in the research
Sample of subset of population
A smaller amount of the population that is representative of the population as a whole
Sampling frame
Samples chosen from a specific source like a university
Census
A systemic collection of data from all members of a population
Probability sample
A randomly selected sample of the population
Non-probability sample
A sample that isnt selected randomly
Representative sample
A sample that accurately reflects the population
Facts
Observations about the world around us (the sky is blue today)
Hypothesis
A proposed explanation for a phenomenon. A starting point for further investigation and testing.
Theory
A well-substantiated explanation made through repeated testing, and has passed all tests and experiments
Law
A statement based on repeated experimental observations that describes some phenomenon of nature
Unstructured interview
The researcher has a brief set of topics they’ll cover
Semi-structured interview
The researcher uses an interview guide with a list of topics and questions
Ethnography
Includes participant observation, immersion in a social setting, field notes, long periods of time
Overt role
When the people you’re observing know you’re conducting research
Covert role
When you’re doing undercover research
Open setting
The researcher has open access to social settings. Like in public areas
Closed setting
The researcher does not have free access, no public access, like researching a gang
Grounded theory
The theory is created as the research is conducted and info and data are collected. “Grounded” in observations and data
Normal distribution
A bell shaped curve of a probability sample. If the sample is large enough, it should be around 95% accurate. With one side of popularion -1.95 and other side +1.95 away from the population mean
Nominal variable
Categorical variable - where they have no relationship to each other like belonging to a religion
Ordinal variable
The different variables are rank-ordered, great than > and less than
Interval ratio
The distance between the variables is the same (like minutes as a marker for time spent doing cardio)
Experimental designs need…
Validity, reliability, and replicability
Experimental designs need an environment that js
Controlled, to control for all confounding variables
Problems in experimental designs…
Confounding variables: history, maturation, testing, and losing participants over time
External validity
When confounding variables are controlled for
Hawthorne effect
When people know they’re being experimented on, it changes the way they behave thus the outcome
Quasi-experimental design
Includes non-random assignment to groups
Ex-post facto design
Useful to test circumstances that have already occurred, that they are unable to test (like child abuse)
Mixed-methods research
Uses triangulation, where one method is used to test the findings of the other
Grounded theory
When the theory is built as the research is conducted
Disadvantages of covert roles
Although it might be easier to gain access to closed settings, doing research without free prior informed consent can be an ethical issue. Research ethics board needs to approve of your covert role
Interview advantages
Less intrusive than ethnography. Might get info that ethnography missed. Interviewees can retell events that happened. Can be more focused on certain aspects rhat ethnkgeaoher mjssed
Ethnography advantages
Complete immersion in a social setting can allow the researcher to pick up on many aspects, that interviews may miss. The social setting will have a context from being in it all the time
Grounded theory
Uses inductive reasoning. “Grounded” in observations and data
Structured observation
Formulate explicit rules outlining what behaviours they might observe and how they will record (and code) them
Quantitative sampling
When the proper sampling technique is used, it can vastly improve validity, as well as improve time and resources needed to complete the research withojut comprosimign tbr findjng
Three sources of bias
Not using random selectjon. If it excludes some cases, then its not representative. Or some people refuse to participate or cant be contacted
Problems w structured interviews
Respondents consistently agree or disagree. Some people lie about their answer to please the interviewer. Prompting suggestions to the interviewee. The way the questions are ordered influence the answer
Probability
Ranges from 0 to 1, wjth 1 being 100% certainty.
Bivariate hypothesis testing
Used to answer the causal question “are X and Y related?” Bivariate means “two variables” - doesnt help control for confounding variables.
P-value
Aka probability value thag ranges from 0 to 1. Rates the probability that we would see the observed relationship between 2 variables in our sample data if there were truly no relatjonship between them
Null-hypothesis
What wr wojld observe if our hypothesis is incorrect.
Simpsons paradox
Most statistical percentages have a different story behind then that is nkt being tokd. Hides a conditional variable aka lurking variable
