Chapter 3 Three Claims, Four Validities: Interrogation Tools For Consumers Of Research Flashcards
What are the three claims??
- frequency
- association
- causal
What do claims do?
- make a statement about variable or about relationships between variables
- argument someone is trying to make
What is a variable?
- something that varies, it must have at least two levels (values)
- *core unit of psychological research (DV)
What is a constant variable?
- something that could potentially vary but that has only one level in the study in question (IV)
What is a manipulated variable?
- we control its levels by assigning participants to the varying levels of the variable
ex: P1 gets 20mg, P2 40mg etc
Measured variable?
- researchers record an observation, statement or value
ex: IQ, BP, Height
also abstract variables like depression
Variables can be described in what two ways?
- Conceptual definition
2. Operational definition
What is a conceptual definition of a variable?
- when researchers discuss theory or journalists write about research….(abstract concepts)
What is an operational definition of a variable?
- when one turns a concept of interest not a measured or manipulated variable….needed for testing hypotheses via empirical research
Give a run through example of both variable definitions.
- conceptual definition: weight gain
2. operational definition: WEIGH THEM
How are variables commonly stated as definition wise?
- conceptually
- to find out the operationalized variable look at how they were measured!
ex How did they measure “X”.
Describe what a frequency claim is!
- describe a rate or level of something
ex: More than 2 million US teens Depressed - two million= frequency/count
- they claim how frequent something is
- claims that mention the % of a variable…the # of people who fit the description or some group lvl
- do not show an associating between them and it does not claim one caused the other
Frequency Claims only focus on how many variables? Measured or Manipulated?
- only one variable
- ONLY MEASURED
Are Anecdotal claims frequency claims?
NO
- they report a problem but do not say anything about the frequency or rate….
- no report of the results of a social science study…they just show an illustrative story (no empirical back up)
What is an Association claim?
**also word indicators??*
- argues that one level of a variable is LIKELY to be ASSOCIATED with a particular lvl of another variable
- also called a correlate
- linked, association, correlated, pedicured, tied to, is at risk for
Association claims involve how many variables?Measured , manipulated?
at least two only MEASURED (this is what separates it from causal claims)
What are used to see if two variables are related after measuring variables for an association claim?
- descriptive statistics
What are the four types of Association Claims?
- positive associations
- negative associations
- zero associations
- curvilinear associations
Explain what a Positive Association is!
- also called?
- ex?
- reped by?
- aka + correlation
- high scores go with high scores; low scores go with low scores
Ex: High scores of abnormal fat go with more dementia symptoms
Ex: Low scores of abnormal fat go with less dementia symptoms - scatterplots
Explain what a Negative Association is!
- also called?
- ex?
- reped by?
- what does the negative refer to?
- aka - correlation or inverse association
- high scores go with low scores, and low scores go with high scores.
- ex: High rates of cell phone usage tied to low sperm quality
- scatterplot
- negative only refers to the slope it does not mean the relationship is bad!
Explain Zero association claims!
- ex?
- reped by?
- no association btwn the variables
- cloud of spots that has no slope and therefore a line draw through it would be nearly horizontal which has a slope of zero.
Ex: ADHD is NOT linked to future drug abuse - scatterplot
Explain Curvilinear Association claims!
- ex
- reped by
- the level of one variable changes its pattern as the other variable increases.
-ex: Relationships btwn Age and frequency of health care visits
( its a u relationship……when your young you visit a lot…as you get older not so much…then it increases once you are elderly) - scatterplot
What is a prediction in regards to association claims.
- mathematically looking to the future…aka using an association to make our estimates more accurate.
Predictions and + /- associations?
- if we know the level of one variable we can more accurately guess or predict the level of the other variable.
- ex: Heavy cell phone usage tied to poor sperm quality. If we known the amount of cell usage we can predict sperm quality.
Are predictions from association claims perfect?
- no usually they are off via certain margin
When it comes to association claims and predictions, what makes them stronger/more accurate?
- if there is a strong relationship btwn two variables = more accurate prediction
- even if there is a slight association it helps us make predictions vs if we did not know about the association.
With predicting association claims (+/-) if we know absolutely nothing about an association what will our predictions be based on?
- average values
Associations help us make predictions by reducing what?
the size of our prediction error
Can Zero association claims help us make predictions?
- NO
- we cannot predict the lvl of one variable from the level of the other therefore our best bet is to guess the mean or avg.
Explain what a causal claim is!
- the variables ____
- can be what three classifications
- words that indicate causal ?
- how are these held in comparison to other claims
- tentative language that still means causal?
- argue that one of the variables is responsible for CHANGING the either
- the variables covary
- can be +, - or curvilinear
- cause, enhance, curb
- hold these to higher standards
- could, may seem, suggest, possible,potential
What are the three steps to go from an association claim to a causal one?
- establish two variables are correlated (cannot have a zero relationship…..the cause variable and the outcome v)
- must show that the causal variable came first and the outcome v came last
- establish no other explanation exists for the relationship
What are the four big validities?
- Construct
- External
- Statistical
- Internal
Valid claim??
- reasonable, accurate and justifiable
* never just say tis valid…we must specify the type
What two validates apply to interrogating frequency claims?
- Construct and External validity
Frequency Claim:
- Construct Validity ?
- how well did they measure their variables
L> how well did the study measure or manipulate a variable - must establish that each variable has been measured reliably (similar scores on repeated testings) ad that the different lvld of a variable accurately correspond to the true differences.
Frequency Claim:
- External Validity?
- how well the results of the study generalize to or represent people and contexts besides those in the study itself!
- ensure that participants rep the population they are suppose to
What are the three types of validity that apply to interrogating Association Claims?
- Construct
- External
- Statistical
Association Claim:
- Construct Validity?
-assess each variable
- how well were they measured
L> if well we can trust the conclusions
Association Claim:
- External Validity?
- can the association generalize to other populations? other contexts, times, places?
Association Claim:
- Statistical Validity?
- statistics are sued to describe data and estimate the probability that results can or cannot be attributed to chance
- extent to which stat conclusions are accurate and reasonable
Association Claim:
- Statistical Validity?
L> What two types of errors does good statistical validity minimize?
- Type I Errors: based on results concluding there is an association btwn 2 variables when there is none. (false alarm)
- Type II Errors: based on results conclude that there is one association when there is one. (Misses)
Covariance??
- As “A” changes, “B” changes
What are the three rules for Causation?
- Covariance
- Temporal precedence: one variable came first in time before another variable
EX: Music lessons enhance IQ…Music lessons must be proven to come first before IQ gains - Internal Validity (third variable rule): a study should be able to rule out alternative explanations for the association.
How do researchers attempt to support a causal claim?
- experiment!
L> gold standard of psychological research
L> manipulate the variable they think is the cause and measure the variable they think is the effect. - Manipulated variable: IV
-measured variable : DV
How do experiments provide temporal precedence and internal validity for causal claims?
- by manipulating one variable and measuring the other one can ensure that by manipulating the causal variable it came first.
-can control for alt explanations (ensure internal v)
L> ex: Random assignment: ensure participants are as similar as possible.
Causal Claims:
- Does debt stress really cause health problems?
- Do family meals really curb eating disorders?
* * are these right or wrong ?
1. A)Covariance: established! (positive association) B)Temporal Precedence: NO C) Internal Validity? NO...alt explanations are possible. 2. A) Covariance: YES B) Temporal Precedence: NO C) Internal Validity: noooooo
Causal Claims:
- External Validity
- Construct Validity
- Statistical Validity
- do the results generalize?
- interrogate both the manipulated and measured variable (how well they were measured)
L>operationalizing manipulated variables one needs to create a specific task or situation that will represent each lvl. - evaluate how well the design of the study allowed researchers to minimize the probability of making the relevant conclusion mistake a false alarm. How strong is the association?Is the difference statistically significant
