Chapter 4: Experiments & Surveys

population

every individual of the entire group

parameters

calculations of a population

sample

a select # of individuals from a population

statistics

calculations of a sample

census

something that counts the population every 10 years

bias

systemically inherent part of a survey

convenience samples

physically easy

voluntary response sample

anyone who wants to participate, usually someone with a strong opinion

simple random sample (SRS)

every group of individuals has an equal chance of being chosen

stratified random sample

a population is divided into homogeneous subgroups/strata

cluster sample

a population divided into pre-existing groups

undercoverage

when certain groups within the target population are excluded or have a significantly lower representation in the sample than they do in the overall population

nonresponse

individuals do not participate or provide data

observational study

observe individuals to measure an outcome; not influencing the outcome

experiment

impose a treatment that affects the outcome

explanatory

variable we think is affecting another variablere

response

what’s being measured or observed

control

variable that does not receive the treatments so there’s something to compare to

treatment

specific condition applied to the individuals

experimental units

smallest collection of individuals to which the treatments are appliedsub

subjects

humans who are receiving a treatment

factor

split explanatory variable into smaller categories

level

specific value for a factor

randomize

assign subjects in different groups to ensure equal chance of being placed in any group, without any pattern or predictability

replication

ensuring that there are an adequate # of experimental units in each treatment group

placebo effect

a beneficial effect produced by a "fake” treatment

single-blind

either the participants or the researchers are unaware of which treatment each person is receiving

double-blind

when both the participants and the researchers are unaware of which treatment each person is receiving

blocking

a technique used in experimental design to reduce variability and improve the accuracy of results by grouping similar experimental units

matched pairs

a specific type of block design where the blocks are size 2

inference

make generalizations about a large populations

systematic random sample

selects a subset of a population at regular fixed intervals after a random starting point is chosen