Stats Ch. 4

design of a statistical study shows bias when…

consistently underestimate or overestimate value you wish to know

strata

group of similar individuals

stratified random sample

classify pop into strata —> separate SRS in each sample —> form sample

cluster

groups of people located near each other

cluster random sample

classify into clusters —> SRS of clusters

response bias

systematic patterns of incorrect responses

confounding

two variables are associated in a way that their effects on the response variable cannot be distinguished

a treatment is a combination of specific values of variables when

an experiment has several explanatory variables

comparison

compare two or more treatments

replication

use enough experimental units in each group so any differences in effects of treatments can be distinguished from chance differences between groups

control group (hint: 3 effects)

placebo effect, minimize variability, try to eliminate confounding effects

replication 3 effects

Effect of chance variation will average out, attempt to reduce variation but not completely eliminate, most important principle of experimental design

randomization incorporates chance so that

experiment is unaffected by external factors

block

group of experimental subjects known before the experiment to be similar in some way that is expected to affect responses of treatments

randomized block design

 random assignment of experimental units to treatments is carried out separately within each block

matched pairs design

randomized block design in which each block consists of a matching pair of similar experimental units. Each group gets both treatments.

statistically significant

An observed effect so large that it would rarely occur by chance