construct validity - how good is the description/concept

frequency claim 1 number, 1 measured variable

• involve one measured variable
• express rate or degree
  • epidemiology studies of incidence and prevalence
  • public opinion surveys

association claims

• involve two (or more) variables
• assert that the value of one variable varies systematically with the value of another variable (correlation)
• you can say something about the overall relationships between two variable
  • texas is hotter than massachusetts (temp and state)

curvilinear graph makes a bell curve type shape, does not go in one direction

causal claims

• involve two or more variables
• assert that one variable is responsible for impacting another
  • "smoking causes lung cancer"

the four big validities

• construct validity
• external validity _how well the results represent the population or can be generalized to others
• statistic
• internal