STAT 100 EXAM 2

Causal inference

an inference about which factors may be responsible for causing a particular effect on something

causal factor

the something we will do that may have an effect

outcome factor

the thing that will be affected

causal graphs

a graphic depiction of the relationship between a causal and outcome factor

observed outcome

what actually happened in the real world world

counterfactual question

what would’ve happened in a parallel universe. Presented as what if questions

treatment observations

observations that were exposed to some causal factor

control observations

observations that are very similar to treatments observations but weren't exposed to the causal factor

Baseline characteristics

attributes that we think might be related to the outcome we’re studying

controlling

act of taking steps to ensure the baseline characteristics for some factors are the same between the control & treatment observations

matching

the process of finding suitable control observations to compare treatment observations. Considered an explicit control

treatment effect

the best guess as to how much an effect a causal factor has on an outcome factor

internal validity

the degree to which the study’s design supports making a causal inference. Depends on the extent to which the control observations are like the treatment observations

causal bias

a study has poor internal validity. Depends on the extent to which the control and treatment observations are alike

case control study

individual control observations are matched to each individual “treatment” observations

average treatment effect

treatment observations - control observations. Can be used to calculate treatment effect size with ATE/SD.

cohort control

an entire group of control observations are matched to the group of “treatment” observations based on aggregate characteristics

confounding

a difference in baseline characteristics between the treatment & control observations

confounding bias

prevent us from making valid inferences based on the study’s result

frequency & relative frequency data relationships

use for categorical data

Calculate — for percents and frequencies

absolute difference & relative difference

find the — for two quantitative attributes

correlation

relationship between casual and confounder factors

signifies a difference in baseline characteristics

relationships between outcome and confounder factors

signifies a relationships a relationship with the outcome factor

moderator

considers whether the ATE will be different for different groups of observations. ATE between groups will help detect if something is a moderating factor

random assignment

uses a random process to determine which observation receives a treatment and serves an implicit control b/c it doesn’t guarantee similarity between groups

matched pairs

two observations are matched based on the important baseline characteristics, one observation is randomly assigned to the treatment and the other is assigned to the control

block randomization

an entire group of similar people are matched and blocked together within each block

randomized controlled study

when a study does not block based on baseline characteristics but still randomly assigns observations a treatment or control group

sampling

the act of selecting observations from which to collect data. Selected observations = the sample

population of interest

all the observations that we’re interested in studying. Sample of observations is selected form this

parameters

characteristics about an entire population of interest

sample statistic

computes averages for a sample

generalization inference

using a sample statistic to determine the value of a parameter. Used to describe an entire POI but only have a sample of the POI

estimate

the value we use as a stand-in for the parameter

external validity

the degree to which a sampling strategy supports making a generalization inference. Depends on how similar the observations in the sample are to the POI

representativeness

the extent to which the observations in the sample are similar to the observations in the POI

sample mean = value of parameter

sample is representative of the entire population

sample bias

prevent us from making a general inference

inclusion criteria

the way in which statisticians define the POI by specifying this

volunteer/convenience sampling

sample strategy in which you make a survey/study available to some or all respondents who meet the inclusion criteria

probability sampling

all eligible observations are assigned a non-zero probability to be included in the sample & then some observations are selected using a random process

strata

similar groups in the POI

quota sampling

strategy in which you have specific limits for a specific strata. Also called volunteer sampling with strata

stratified random sample

probability sampling w/ strata

multi-stage sample

the sample is selected in stages with random sampling used at each stage

survey response

the study of how & why individuals choose to respond to surveys

social exchange theory

in exchange for giving their data, respondents should get something in return

survey incentive

the act of offering something in return for a respondent providing their data. Reduces non-response, constitute an ethical return

non-response

when a respondent selected for the sample chooses not to respond to the survey either to specific questions or the entirety of the survey. Interferes with the ability to make generalizations

stratification and random assignment have the — external validity

highest

blocking and random assignment have the — internal validity

highest