Permanent sample plot (PSP) studies play a major role in ecological and management research. Bias and error may arise in such studies at each stage of the assessment procedure from plot establishment through to data interpretation, and can undermine confidence in the use of PSP data. This paper considers these problems and suggests how they might be identified and minimised. Examples are taken from Budongo Forest, Uganda, which provides some of the world's longest-term PSP data from both primary and treated tropical forest (1939-1993 and 1933-1992 respectively). Three distinct types of artefact in PSP studies are recognised: errors in methodology, impact of methodology, and inappropriate analysis. Deficiencies in each aspect of plot evaluation can have considerable influence on the evaluation of change with time and of differences between sites. Each technique can be controlled and/or improved. Data correction and 'harmonisation' procedures are discussed. Interpretation of PSP data must consider the many pitfalls of assessing such complex data. Some of the central considerations in data interpretation are reviewed and critical areas are defined e.g. time period between observations influences the evaluation of dynamic quantities. Some conclusions disagree with those of other authors. The most significant factors identified (many with implications for vegetation monitoring generally) are summarised and a set of recommendations is provided.
