PCE - Planning and Conducting Experiments Overview

Math_APStatBanner.png Planning and Conducting Experiments Overview

Introduction

This unit will introduce the basic principles of sound experimental design.   Three of these principles apply to EVERY experiment while the fourth is only necessary in certain circumstances. You will understand the purpose of randomization of treatment groups, learn how to the use a control group to minimize variation in the results, and understand the importance of selecting a sample size large enough to provide adequate information for forming a conclusion.   A quick way to recall these three principles is to think:    Randomization , Replication  or Repetition , and Control .   The fourth principle is Blocking  and is only necessary when mixing all the participants together could potentially cause HUGE variation in results.   A simple everyday example of blocking is the act of separating laundry into fabric type and/or color since the groups react differently to different water temperatures.   Other common blocks would be by gender into males and females, or by age with adolescents, middle-age, and seniors, or by ethnicity.   It is best to group subjects with similarities into a block and keep them separate throughout the experiment.
It is always a good idea, although not a requirement, to perform a "blind" experiment, one in which the participants are unaware of which treatment they are receiving.   Even better would be a "double blind" experiment in which neither the participants nor the data collectors are aware of the treatments.   In your investigative task you will design an experiment and demonstrate its design through the use of flow charts and paragraph form description.  Of course, you will use all statistical methods learned so far, including graphs, to analyze your results.

Essential Questions

  • How is an experiment different from a study?  
  • What are the characteristics of a well-designed experiment?
  • What different designs are available and how do we decide which to use?  
  • How can biased responses be minimized? 
  • What is a flow chart?

Key Terms

The following key terms will help you understand the content in this module.

Experimental units- individuals on whom an experiment is performed

Subjects- experimental units are called SUBJECTS or participants when they are human

Treatment-process, intervention, or other circumstance applied to the randomly assigned experimental units

Factor- variable whose level is controlled by the researcher

Level -specific values that the experimenter chooses for a factor

Placebo- a null treatment known to have no effect, administered for comparison to insure the observed effect is a result of treatment and is not due to the placebo effect

Placebo effect- a non-treatment administered for comparison because many subjects do respond to this null treatment (called placebo effect),

Control group- the experimental units that are assigned to a baseline treatment level, typically the placebo, providing a basis for comparison

Randomization- assigns subjects to treatment groups randomly in an effort to reduce bias

Completely randomized- is the ideal simple experimental design just as a simple random sample is the ideal simple sample

Randomized block design- groups of experimental units are similar and gathered together into blocks to isolate variability due to a specific trait and make the differences due to treatment more clear

Block design- gather similar experimental units into groups to reduce variability and see differences more clearly

Matched pairs- subjects are matched in pairs and each treatment is given to one subject in each pair - pre and post testing can be considered matched pair results on the same subject

Statistically significant- a difference observed is too large to believe that it is likely to have occurred naturally or by chance

Replication- refers to using a large enough sample of individuals OR repeating the entire experiment with new samples

Double-blind- when every individual from either of the main groups (subjects and evaluators) is blinded

Single-blind- when the subjects of the experiment are unaware of the treatment

Math_OverviewBottomBanner.png IMAGES CREATED BY GAVS