Stats

  • Population – All individuals

  • Sample – Subset of individuals

  • Census – Gathers information from every individual in a population

  • Inference – Drawing conclusions about a population from a sample

  • Simple Random Sample (SRS) – All groups of size n have an equal chance of being chosen

  • Describe how to randomize – Calculator, Table of random digits, Names from a hat

  • Stratified – Individuals in groups are alike, Select individuals from each group

  • Cluster – Individuals in groups represent entire population, Select entire group

  • Bias – Estimates are consistently too high or too low

  • Convenience – Individuals are easy to reach

  • Voluntary Response – Only some choose to respond

  • Response – People lie or poor wording of questions

  • Non-response – Individuals can not be reached or choose not to reply

  • Undercoverage – Sampling excludes part of population

  • Observational Study – Looks at relationship between two variables

  • Retrospective – Looking back

  • Prospective – Looking forward

  • Experiment – Treatments assigned, looking for cause and effect

  • Explanatory Variable - What we assign or study

  • Response Variable – What we measure

  • Confounding Variables – Other variables that may have caused the result

  • Why we randomize – Create roughly equivalent groups

  • Experimental Units – Who gets the treatment

  • Replication – Enough individuals to consider the results valid, or repeat the study with more individuals

  • More data, more better ☺

  • More data – Smaller Margin of Error

  • Control – keeping other variables constant

  • Control Group – Placebo or old treatment for comparison

  • Placebo Effect – Dummy response to fake treatment

  • Single Blind – One side (individuals or experimenter) does not know what treatment is given

  • Double Blind – Both sides don’t know

  • Statistically Significant – Results unlikely to occur by chance alone (< 5% generally)

  • Blocking – Stratifying, but for experiments

  • Matched Pairs – Blocking in groups of two individuals

  • Random selection allows us to make conclusions about the entire population

  • Random assignment allows us to make conclusions about cause and effect

  • Four principles of experimental design – Random, control, compare and replication