STAT 2000 Lecture Notes - Lecture 2: Randomized Experiment
January 13, 2016
Chapter 4 Continued
• Generalizing from sample to population requires random sampling
• Matched-Pairs Designs
o Randomly selected 20 students; have them all take both tests: MATCHED
PAIRS; LINKED
o Randomly selected 40 students; split into random groups of 20 where each group
would take a different test: NOT MATCHED PAIRS
• Causation
o To conclude that smoking causes cancer, you must be able to rule out the effect of
possible confounders (i.e. second hand smoke)
o If a randomized experiment finds an effect, we can conclude the difference in
treatments caused the effect
o Association is evidence of possible causation
Chapter 2: Exploring and Summarizing Data
• Variable: any characteristic we’re studying
• Characteristics will have variability; stat methods provide ways to measure and
understand characteristics
2.1: Different Types of Data
• Categorical: each observation belongs to one set of categories (gender, political
affiliations, etc.)
• Quantitative; observations are numerical values (age, SAT scores, etc.)
o Discrete: countable number of values
o Continuous: uncountable number of values; take on all values in an interval
• Think about finding the average to determine if data is categorical or quantitative
o Quantitative: makes sense to find average
• Examples
o Proportion of customers using Verizon is 43%: Categorical
▪ Use stat. p-hat; p-hat=0.43
▪ 43 have Verizon; 57 do not
▪ Question asked: do you use Verizon?
o Average high temp in March in Athens is 66 degrees Fahrenheit: Quantitative
▪ Use stat. x-bar; x-bar=66
▪ Question asked: what is the high temp for a day in March
• Answers: 55, 60, 70, 74, etc.
find more resources at oneclass.com
find more resources at oneclass.com