Chapter 3 - Interrogation Tools for Consumers of Research Flashcards
define a measured variable
a variable in a study whose levels (values( are observed and recorded
define a manipulated variable
a variable researchers control, usually by assigning study participants to the different levels of that variable. some variables cannot be manipulated due to ethical reasons, examples being childhood trauma or race.
define constructs/conceptual variables
a variable of interest, at an abstract level, defined as part of a formal statement of a psychological theory. cannot be directly observed so we need to operationalize them.
define operational variables/operationalizations
the specific way in which a concept of interest is measured or manipulated as a variable in a study. a definition of a concept in terms of precisely described operations, measures, or procedures.
define the frequency claim
a claim that describes a particular rate or degree of a single variable. (ex. is how frequent or common something is)
variable is always measured and not manipulated and only focuses on a single variable.
- Forty-one Percent of Children Worldwide Experience Moderate Food Insecurity
define association claim
a claim about two variables, in which the value of one variable is said to vary systemically with the value of another variable. Mostly found in correlational studies
variables are always measured and involve at least two variables.
- Study Links Coffee Consumption to Lower Depression in Women
define the difference between a positive and negative association with the association claim
positive: high levels of one go with high levels of another, and same for low (linking exercise to higher pay)
negative: high levels of one go with low levels of another (coffee drinking leading to less depression)
define causal claims
a claim arguing that a specific change in one variables is responsible for influencing the value of another variable. one variable is manipulated, and one is measured and the rest is held constant (an experiment)
- To Appear More Intimidating, Just Tilt Your Head Down, Study Suggests
define a variable
it is anything that varies. it has multiple levels (values).
ex. education can be broken down into different levels such as university, high school, and college.
if examining only a single value of a variable, then it is considered a constant
how do scatterplots work to describe a correlation?
they show the association between two quantitative variables.
they display form, direction and strength (close bounded) of the association.
how to know the difference between a causal and correlational relationship?
causal: X changes first and then causes Y to change too
- leads, results, affects, boost, could increase
correlational: X and Y change together
- related, associated, linked, worse, likely
correlational and causal can not be linked together!
define validity
the appropriateness of a conclusion or decision
define construct validity
an indication of how well a variable was measured or manipulated in a study. researchers must make sure that variables are measured reliably ( to get similar scores of repeated testing) and that the different levels accurately represent difference.
define generalizability
the extent to which subjects in a study represent the population; how well the setting repreesnet other settings or contexts
define external validity
an indication of how well the results generalize to, individuals or contexts beside those in the study itself
define statistical validity
the extent to which statistical conclusions derived from a study are accurate and reasonable. How well can the numbers support the claim?
how do we go about evaluating statistical validity?
- start at the point estimate (single estimate of some population based on the data)
- find the precision of that estimate with confidence interval or the marine of error of the estimate
- improve the validity with multiple estimates
define confidecnce interval
range to capture the population value for some point estimate
define margin of error of the estimate
inferential statistic providing a range of values that has a high probability of containing true population value
what are the type of validity used for frequency claims?
construct, external and statistical validity
how do we interrogate an assocation claim with validity?
see how well the correlational study behind the claim supports construct, external and statistical validity
using construct validity for association claims
want to look at how both variables are measured. the variables need to be studied well in order for the results to be reported and have more confidence in them.
define internal validity
in a relationship between A and B, the extent to which A, rather than some other variables C, is responsible for changes in B. ruling out that there is no thrid variables causing changes to B
what is the 3 criteria for causation?
covariance, temporal procedence and internal validity
define covariance
the degree to which 2 variables go together. the proposed causal variable must vary systemically with changes in the proposed outcome variable.
define temporal precedence
stating that the proposed causal variable comes first in time, before the proposed outcome variable. ie. cause and effect
define independent and dependent varible
independent is the one that is getting manipulated (must have more than 1 level) , and dependent is the one that is being measured to see if there is a cause and effect.
define random assignment
use of random method to assign participants into different experimental groups
how is validity shown in frequency claims?
construct: how did the research define the construct, measure the variable
external: how can we generalize to large population, how was sample selected
statistical: how accurate, margin of error (bigger the less confidence)
how is validity shown in association claims?
construct and external: same as frequency
statistical: how strong is the association, more scattered points means weaker association
how is validity shown in causal claims?
construct: is manipulation appropraite, how well is the outcome of the measured
external: same as other claims
statistical: does it minimize errors, looking at the average more than comparing
internal: can you rule out other variables that might have caused a change in the measured other than the manipulated
what validity do you prioritize for a frequency claim?
external, you want it to be able to generalize to a bigger population
what validity do you prioritize for a causal claim?
internal validity, you want to make sure there is no confounding variables involved with the experiment t change the results
