Systematic Error / Bias (Non-random error):
A systematic error or bias occurs when the results differ in a systematic way from the true value. It is the difference between the study results and reality. All studies contain some random error that is determined by chance. If the errors are not determined by chance alone and that the depart from the true values in a systematic way bias. It is a major problem to be considered in the designed studies.
Selection or Sampling Bias:
Occurs when those subjects who participate in a study are not representative of the population that they are supposed to represent. Most common is when subjects select themselves for a study. Best dealt with by a random selection of subjects. Response bias is similar in that the respondents to a survey differ systematically from non-respondents. Reduced external validity
Occurs when subjects are asked question about past events – same may remember more than others.
Control and t/t group act differently.
A systematic difference in the measurements between groups.
Within observer variation:
Inconsistency between measures done by a single individual – largely random in that it is unpredictable in direction
Between observer variation:
Systematic differences between observers – this may be large in relation to the real difference between groups.