Lesson 1 3 Introduction to Experimental Design Basic

Lesson 1. 3 Introduction to Experimental Design

Basic guidelines for planning a statistical study 1) 2) First, identify the individuals or objects of interest. Specify the variables as well as protocols or procedures for taking measurements or making observations. 3) Determine if you will use an entire population or a representative sample. If using a sample, decide on a viable sampling method. 4) Collect the data. 5) Use appropriate descriptive statistics methods (chapters 2, 3, 4) and make decisions using appropriate inferential statistics methods (chapters 9 -12) 6) Finally, note any concerns you might have about your data collection methods and list any recommendations for future studies.

Ways to get data: Census: measurements or observations from the entire population Sample: measurements or observations on a few to obtain info on all Simulation: faking a real world phenomena…usually on a computer Observational study: getting info without influencing participants, doesn’t change the response or the variable Experiment: a treatment is imposed in order to observe a possible change in the response or variable

Experiments Control Group: no treatment or fake treatment (placebo) Treatment Group: treatment is given Single-blind Experiment: the participants do no know which treatment they have been given Double-blind Experiment: neither the participants nor the observers know which subjects are receiving the treatment Survey: asking people questions to get info Hidden bias: not measuring what you hoped to measure. Not random samples

Experiments, continued Confounding variable: possibly uncontrolled variables that influence experiments Ex…people who take a vitamin C pill everyday don’t get colds as often. The confounding variable might be that those people naturally eat more healthy, or the glass of water is cleaner Randomization: making the experiment as random as possible to help minimize confounding variables Replication: being able to reproduce an experiment, not just repeat it. Reproducing would likely give you identical results Blocking: putting groups together according to what they have in common. You then randomly assign tx within each block Matched Pairs Design: use either 2 matched people or the one person who receives each of 2 tx’s

Assignment p. 26: 1 -4