science
makes claims that are:
Testable and consistent with well established scientific facts
Confronts data rather than ignoring it
No vague language
Committed to an active, ongoing program of research
pseudoscience
- Unscientific thinking masquerading as scientific thinking
- Uses precise, scientific sounding language
- No evidence of continued research over time or new knowledge
- Reliance of anecdotes as evidence
- Reliance on authority endorsements (or false authorities)
- Extraordinary claims without supporting evidence
- Evidence relies on confirmation rather than refutation
- In true science, we look to disconfirm rather than confirmation
Intuition-
it feels true
A gut feeling
Subjective
We are biased
Authority
-the scientist says it is true
Could be a good starting point
Different experts have different views
They view things with their own lens of expertise
They may not always be right
- Rationalism-
- it makes sense logically
- People will not knowingly ingest poisons (yet, people smoke, so this is not the case)
- Our own logic is based on our psychology, which differs from person to person
- Empiricism-
- I observed it to be true
- Might make a generalization that cannot be taken out of context
- We all observe things differently and come to different observational conclusions; for example, the duck rabbit picture
Tenacity-
using what we have known to be true in the past; method of knowing based largely on habit or superstition
Theory
- Explanation of behavior that can be tested through research
- Hypothesis: prediction regarding the results of a research study
- Must be stated in declarative form
- Be brief
- Testable
- Reflect a theory or literature upon which it is based
- Posit an expected relationship between variables
Construct
When researchers use a theory, they typically work with a conceptual definition of a variable. The term constructs represents these abstract concepts that we aim to measure (e.g. depression, intellectual abilities, substance abuse)
Measurement
refers to the process of assigning arbitrary symbols (usually numbers), according to a predetermined set of rules, to different events or objects.
Operational definition:
A definition of a variable in terms of the actual procedures used by the researcher to measure and/or manipulate it
Variables can be described in different ways:
- Continuous vs. discrete (# levels?)
- Scales of Measurement (NOIR)
- Independent and dependent variables
- Manipulated vs. non-manipulated
- Extraneous and Confounding
Extraneous:
Unplanned and uncontrolled factor(s) that can arise in a study and affect the outcome. Extraneous variables are typically randomly distributed influences that detract from the researcher’s efforts to measure what was intended to be measured.
Confounding:
An unwanted factor that affect groups differently and make it difficult to know what caused changes in the dv. With the presence of a confound, it is not possible to determine which variable is at work (the IV or the counfounding variable).
quasi-independent variable, subject variable or classification variable.
If the researcher is unable to manipulate the variable or it’s based on characteristics of the individual that cannot be manipulated it is a non-manipulated variable. Also referred to as a quasi-independent variable, subject variable or classification variable.
Why is anecdotal evidence insufficient?
It’s a sloppy way to collect data
Goal of scientific research-identify and collect data from samples of participants that are representative of the whole population
The plural of anecdote is not data
- Probability sampling types
- Probability:
- Simple random
- Systematic sampling
- Stratified random
- Cluster sampling
simple random sampling
- The entire population is known to begin with
- i.e. all Carolina students
- You randomly select a thousand students to participate and each student has an equal probability of being selected
Systematic sampling
I’ve got the entire population but every fourth person will end up in my sample
Stratified random
All students at carolina, and you know the population values already
Ie. 40% male and 60% female
So you select a sample that is also specifically 40% male and 60% female
cluster sampling
I’m selecting everyone in a dorm out of a list of dorms. So the sampling is in clusters of individuals rather than just individuals; BUT just for it to be groups doesn’t make it cluster, it has to be ALL individuals in a certain group. i.e. if you have participants from every group than its not cluster, you have to have ALL participants from SOME groups, not SOME participants from ALL groups.
One advantage-they’re already in groups, and its easy in terms of resources because they’re generally all in one location
Sometimes the people within the cluster are more similar ot one another than they are across the clusters so it could be a disadvantage
Because its random, its possible that you leave relevant samples out (i.e. you accidentally omit entire south campus leading to a disproportionately upperclassmen sample)
Nonprobability:
Convenience/haphazard/volunteer
- You ask for volunteers or use students in a particular class)
Quota:
- sample of nonrandom students (trying to get a 40 60 male female split, but you have ot rely on volunteers, so its not random)
Snowball sampling:
- Referral based method: tell all your friends
- Extraneous variables can result in 3 possibilities
- possibilities:
- Ev has no efffect
- Ev affects all of the groups/conditions in the same manner
- EV AFFECTS THE GORUPS/CONDITIONS DIFFERENTIALLY
- *only this one is problematic and is thus confounding
- Random assignment vs. random selection
- Random assignment
* Randomly assigning participants you’ve already recruited to the levels of your IV
* Impacts internal validity, which is more important than external validity in the case of an experiment- Random selection
- Actually recruiting individuals (and doing so randomly)
- Randomly: you have equal probability of being selected form the population
- Random selection
- Random assignment
