Principles of Experiments

Experiments are based on:

Careful planning and design - e.g.

  • Observation – seedlings bend as they grow.
  • Hypothesis – seedling are growing towards light coming from one side.
  • Design – shoebox with interior compartments – one with a hole on top and the other with a hole to the side.
  • One variable here – light. A variable is a factor that may change in an experiment. Other variables may include wind (draught) or temperature – but these are kept constant in the above experiment (draught –free room at constant temperature).

Safe procedures

  • Wear lab coat and safety glasses.
  • Be aware of the safety information for chemicals.
  • Tie back long hair.
  • Never place fingers in eyes/mouth unless washed.
  • Report all accidents to teacher/supervisor
  • Avoid contact between electrical equipment and water.

Control experiment – a comparison

  • used to provide a standard against which the actual experiment can be judged. The control or copy experiment differs in one variable from the real experiment.
  • Example 1 - In seedlings experiment have a third compartment with no hole.
  • Example 2 – testing the effect of an arthritic drug. A number of patients are x-rayed and deformities noted. The drug is administered for 6 months and they are x-rayed again and found to have improved.

Fair procedures

  • Large sample size - Produces more reliable results.  In seedling experiment use a large number of seedlings. Reduces the risk that results are individual differences. In arthritic drug test if one person treated and recovered it may have been due to person recovering of their own accord whereas if hundreds of people treated and improve then the result is more likely to be due to the drug given.
  • Random selection - prevents bias/unfairness. It would not be fair to test seeds of one type in above experiment. Want to prove that results apply to all seeds. In arthritic drug test it would not be fair to test only females or those above 50 years of age – the results may be influenced by the factor selected (i.e. females over 50 years of age).
  • Others must be able to replicate experiment - Reporting experiments publicly so that they can be replicated/repeated.
  • Double blind testing. Neither the person being tested or the tester should know who is receiving the real treatment or who is receiving the placebo.

previousPrevious - Scientific Method
Next - Scientific Method Limitationsnext