ABSTRACT
Information may be obtained by direct observation or indirectly by enquiring of individuals concerned with the problem. Each of these sources is subject to various types of error: nevertheless it will often be possible, by being aware of this, to make at least some allowance for error. Experimental work typically involves the use of more than one independent observer, measures of an agreement being calculated to show the degree to which the information is subject to the individual characteristics of particular observers. if both observers independently show a high degree of agreement, this suggests that the information closely reflects what actually happened rather than such things as fluctuation in the mood or attentiveness of the observer. A low degree of agreement suggests that what actually happened was responsible for only part of the results, a good deal of the data reflecting idiosyncracies of the observers.
