The Scientific Method and Experimental Design

Document Sample
The Scientific Method and Experimental Design Powered By Docstoc
					                        The Scientific Method and Experimental Design


A. Scientific Method” a comprehensive (non-linear), logical problem solving process
       a. Steps of the scientific method
                 i. Observation (qualitative and quantitative): done throughout the scientific process.
                ii. Identify the problem or question
              iii. Hypothesis(es): Educated guess as to the solution of the problem. Must show cause
                    and effect (e.g., “If…Then…Because…). Must include both independent and
                    dependent variables.
               iv. Prediction: Expected results (typically quantitative: graph)
                v. Experimentation
                         1. Includes materials and methods
                         2. Types of: Lab experiments or controlled experiments, field experiments or
                             natural experiments, hybrid experiments (field + lab), models that predict
                             reality (e.g., climate models that predict long-term forecasts using many
                             variables)
                                 a. Blind Studies
                                           i. Single-blind: subjects receiving the treatment (experimental
                                               group) don’t know they are receiving it.
                                          ii. Double-blind: neither the treatment group nor the researcher
                                               knows who is receiving the treatment. The researcher only
                                               finds out who received it after the study is over.
               vi. Results (data): Usually in tables
              vii. Analysis: What you DO to the data
                         1. examples: graphs, %, averages
             viii. Conclusions: revisit hypothesis (es): Hypothesis supported by data or not?
                    Hypotheses are NOT right or wrong as conclusions are as only as good as your data
                    (i.e., experimental error)
               ix. Theory: Hypothesis widely supported by multiple experiments from different
                    researchers. Theories can be wrong (e.g., technology changes)
                x. Law: Fact! The theory fails to be proven faulty.

B. Experimental Design
      a. All good experiments must have
               i. 1 or more hypotheses (see above) and predictions (i.e., expected results)
              ii. control group
            iii. treatment group (manipulate independent variables)
             iv. Identify control or confounding variables
              v. Replication (independent, repeated trials of the experiment)
             vi. A large sample size that is randomly selected (no bias) using a random number
                  generator or table, or randomly drawing from a hat like a lottery.
            vii. Measurements of the dependent variable (what are you observing? It must be
                  measurable!!!)
           viii. Mathematical or statistical comparison between experimental (treatment) and
                  control groups
             ix. Description of results: Graph (labeled) and/or a data table that is consistent with the
                  experimental design.