An error that may result in the measured value being either above or below that would otherwise be measured (i.e. a change is possible in either direction). Causes include limitations of instruments, environmental factors and slight variations in procedures.
An error that biases the result in one direction either above or below what would otherwise be measured (i.e. a change in one direction only). Causes include observational error, measuring instruments that are incorrectly calibrated or used incorrectly, and a flawed experimental method.
Any variable that is not the independent variable that has directly and systematically affected the dependent variable in an unwanted way. Confounding variables may have been an extraneous variable that was not controlled for, or a variable that simply cannot be controlled for.
Mistakes or miscalculations by the researcher. Personal errors should not be included in reporting and analysis of data. Measurements must be repeated to detect and correct personal errors.
The closeness of the agreement between the results of successive measurements of the same quantity being measured, carried out under the same conditions of measurement. Precision affects repeatability.
The closeness of the agreement between the results of measurements of the same quantity being measured, carried out under different conditions of measurement. Repeatability affects reproducibility.
How close a measurement is to the true value of the quantity being measured. True value: the value that would be found if the quantity could be measured perfectly. Systematic errors affect accuracy.
The extent to which an investigation investigates what it sets out and/or claims to investigate. Precision and accuracy affect internal validity. The appropriateness of the investigation design, sampling and allocation techniques should be considered.
The extent to which the results of research can be applied to similar individuals in a different setting (individuals who are different from the research population and sample). Relates to generalisability. Sampling techniques and inclusion criteria should be considered when trying to apply the results to the broader human population. Internal validity affects external validity.