Chapter 3: 3 Claims and 4 Validities Flashcards
1
Q
What is a variable?
A
something that varies
2
Q
What are the variable(s) and level(s) to this statement: Most students don’t know when news is fake
A
- Variable: knowing if news is fake
- Levels: knowing/not knowing
3
Q
What are constants?
A
- Something that could vary- but only has 1 level in this study
- ex. “students don’t know when the news is fake”- students is the constant
4
Q
What is the constant and variable in this statement: 15% of Americans smoke
A
- Constant= Americans
- Variable= Smoking vs not smoking
5
Q
Manipulated vs measured variable
A
- Measured variable: dependent variable, something you count/observe. Ex. height and IQ
- Manipulated variable: independent variable, the research controls. Has different levels
6
Q
Constructs/Conceptual Variables vs Operational Definitions
A
- Constructs/Conceptual Variables- are more abstract and broadly explain your interests. Ex. coffee consumption, and happiness
- Operational variables: aka operational definitions, defining a variable in terms you are going to use to measure it
7
Q
How could you turn “coffee consumption” and “happiness” from constructs/conceptual variables to operational variables?
A
- Caffeine: mg/day
- Happiness: rank on 10 point scale
8
Q
What is a claim?
A
The argument you are trying to make
9
Q
What is a frequency claim?
A
- statement about # of occurences
- ex. 39% of teens admit to texting while driving. Screentime for children younger than 2 doubled last year. Most students/people do not know when they are reading misinformation
10
Q
What is an association claim?
A
- One level of one variable is related to/associated with a particular levfel of another variable
- States a relationship between at least 2 variables- a correlation coeeficient. helps make predictions
- ex. study links coffee consumption to lower levels of depression in women. There is no relationship between late dinner times and childhood obesity.
11
Q
What are the 3 types of association claims?
A
- Positive associations: as the value of one variable increases, the value of the 2nd variable also increases
- Negative associations: as the value of one variable increases, the value of the 2nd variable decreases
- Zero association: there really isn’t a relationship between variables
12
Q
What is a causal claim?
A
- A statement that says changes to one variable are responsible for changes to a second variable.
- Makes a statement about cause and effect. Ex. Pretending to be batman makes kids stay on task.
- Can also make a statement about lack of causal relationship. There is 0 evidence to support that vaccinations cause autism
13
Q
What are the 3 criteria for causal claims
A
- Two variables must be correlated- their relationship cannot be 0
- The causal variable comes first and the outcome variable comes second.
- No other explanatory variables are present- internal validity
14
Q
What is validity?
A
- Correctness/Accuracy: how well a study actually measures what it is intended to study.
- Allows inferences from study to be appropriate, meaningful, and useful
- Several kinds of validity: Face/Internal/External/Construct
15
Q
What is internal validity?
A
- All about cause-and-effect
- What causes the outcome: did changing the IV result in changes in the DV
- Lab experiments: high in internal validity, individual differences are controlled, and the independent variable is isolated from contamination
16
Q
Internal Validity
The 3 Musts of Causality: Co-Variation
A
- Easiest to demonstrate
- Says that changes in variable #1 predictably correspond to changes in variable #2
- Showing presence of correlation
17
Q
Internal Validity
The 3 Musts of Causality: Temporal precedence
A
- Says that changes in the IV always precede changes in the DV
- Your manipulation must come before the change in behavior
18
Q
Internal Validity
The 3 Musts of Causality: Third Variable Problem/Eliminating Confounds
A
- Third Variable Problem: creates a false appearance of a causal relationship. Ex. Ice cream sales and violent crime, Intelligence and shoe size
- So need to eliminate confounds: having a control group and doing the experiment double blind helps
19
Q
Internal Validity
Experimental Design
A
- Best suited to show causality
- Manipulate IV, Control confounds, carefully define DV
- Great for Internal validity, but not as great for external validity (as IV increases, EV decreases)
20
Q
What is external validity?
A
- Generalizability: extent to which study results apply to other settings, people, social settings, and cultures
- With respect to people: will results seen in one sample occur in another sample? Seen in different ages, genders, and groups?
21
Q
External validity- generalizability
A
- Will results seen in one setting occur in another setting?
- Natural vs contrived
- Home vs school
22
Q
“Can’t be delicious and nutritious”
A
- Can’t have high external and high internal validity in one study
- Passive observation or active manipulation
23
Q
What is the construct in construct validity?
A
- Theoretical, intangible quality, trait, mental process, on which individuals differ
- Examples include: leadership ability, depression, intelligence, and love
24
Q
What is construct/conceptual validity?
A
- Construct validity is the idea that the IVs and DVs truly represent the constructs being studied or measured.
- Based on quality of operational definitions
25
Q
Construct Validity: “Makes research fun and challenging”
A
- Coming up with good ways to measure elusive things
- Example self-esteem: positive attitude of self, competence/likeability, stable/variable trait
26
Q
What is conceptual validity?
A
- How well the hypothesis fits the theory- How well it follows logically from the theory and appropriately “maps onto” the theory
- Also about if interest in this hypothesis or research question warrranted?
27
Q
What is reliability in research?
A
- Consistency of measurement over time
- Is what you’re studying driven by stable attributes, traits, and individual differences
- How good the study/measure is at producing similar results w/ repeated exposure to the same persons under similar circumstances
28
Q
Is reliability the same as validity
A
NO
29
Q
What are the kinds of reliability
A
- inter-rater/inter-observed, internal consistency, and test-retest
30
Q
What is test-retest reliability?
A
- Same measure/manipulation, twice
- How consistent study results are when study is executed multiple times?
- You expose the same group to the same study several times, the 2nd outcome should be predictable based on the first
31
Q
Test-Retest Reliability: Beware of maturation
A
- Expect lower test-retest reliability as more time passes between 2 studies. Because people change over time- raoidly changing traits and way of life/frame of mind
32
Q
When is the ideal re-test window?
A
2 to 4 weeks
33
Q
reliability
Internal consistency
A
- consistency inside the measure
- How well all parts of a study measure contribute meaningfully to the measure
- To fix the problem: give 2 alternate forms versions of same measure to same group (how much someone loves and hates the same thing) or “split-half” the test and correlate the scores on each half
34
Q
What is inter-rater reliability?
A
- The degree of agreement
- Multiple independent researchers judge/quantify each participant’s outcome score without confering with each other. (like judges in the olympics)
- Need time to train judges- they must understand and follow the rules for coding the outcome
35
Q
inter-rater reliability
When is degree of agreement really useful?
A
- When you can’t use self-rating to measure the construct
- when there are highly subjective measures of construct
ex. break dancing in olympics
36
Q
More is better
A
- Better representativeness
- Results are more reliable with more observations, measures, and observers
- More data=less error, and is a better reflection of the construct you are studying
- Results are also more valid- closer approximation of the actual/true relationship or phenomenon. Our sample comes closer to representing the population
37
Q
Discrete variables
A
- separate/nondivisible categories
- no values between adjacent categories
- whole countable values
- Qualitative categories
- Ex. # of kids or cars you have
- Good for frequency statements
38
Q
Continuous Variables
A
- Has infinite number of values
- Any number with fractional parts
- no fixed categories
- divided in many ways
- Ex. height, weight, velocity, minutes
39
Q
Scales of Measurement
Nominal
A
- set of categories with different names, only count frequency
- don’t describe direction or size of difference
- Examples: What car you own, academic major
40
Q
Scales of Measurement
Ordinal
A
- categories w/ different magnitude
- describe rank/sequential direction/order
- do not describe the size of the interval
- Ex. gold/silver/bronze, small/med/large
41
Q
Scales of Measurement
What do interval and ratio scales have in common?
A
- both have intervals of exactly the same size (ordered categories w sequential intervals)
- Both tell us size and direction of change
42
Q
Scales of Measurement
What is different between ratio and interval scales?
A
- Ratio: has absolute zero point. 0= complete absence of construct. Allows for ratio comparisons (2x as many beers, 5x as much pizza)
- Interval: has arbitrary Zero point. 0 for convenience/references. Cannot do ratio comparisons. (ex. temperature and golf score)
