Abstract The scientific value of the outcome of an ex- periment is closely related to its design and analysis. This article deals with the design issues of pseudorepli- cation (whether the experimental design has the statisti- cal features needed to answer the question as posed) and execution errors (problems arising from how the experi- ment was conducted). Three issues of analysis are also dealt with: the number and type of response measures to record; how measures should, and should not, be com- bined into a single response measure; and how to inter- pret an apparent lack of response. Interactive playback is considered separately because it raises its own specific design and analysis issues. Although the examples gen- erally refer to video playback, these issues are common to all experiments in behaviour. Key words Pseudoreplication · Experimental error · Response measures Introduction An experimental approach to scientific understanding re- lies critically on the design and analysis of experiments. Such a statement is so self-evidently true that perhaps it is worth emphasising that all experiments are, to a great- er or lesser extent, limited in their explanatory power by aspects of their design and analysis. Playback is an ex- perimental technique in which natural or synthetic sig- nals are broadcast and the response of animals noted. Playback experiments in any sensory modality are sim- ply a subset of all possible experiments; therefore the de- sign and analysis considerations discussed in this article are most definitely not specific to video playback. Such considerations also apply to various alternative ap- proaches to investigating visual stimuli (e.g. live stimuli, dummies, still photographic images) as well as to many other sorts of experiment. However, since the Lisbon Video Playback Workshop dealt with video playback, I shall use examples drawn from this method to explain general design and analysis considerations. In much the same way that the consensus paper in this issue (Oliveira et al. 2000) is not an exhaustive list of features to be con- sidered, neither is this article a recipe for that unattain- able goal – the perfect experiment. Rather it should be regarded as a starting point. It is based on general discus- sions of experimental design and analysis in biology (e.g. Sokal and Rolf 1981; Barnard et al. 1993) as well as more specific issues related to behaviour (Martin and Bateson 1993; Milinski 1997). It has been influenced strongly by experience with playback of acoustic stimuli. Many biologists find experimental design and analy- sis to be considerably less exciting topics than the vari- ous stimulus representations and manipulations that can be performed on a personal computer. This does not alter the fact that experimental design and analysis cannot be ignored. The starting point for this article is that adequate vid- eo stimuli are available, as is knowledge of the appropri- ate environment and context in which to play them back. It should be obvious from the rest of the articles in this issue that it takes considerable thought and expertise to get to such a starting point. Just in case it is not, let me emphasise that I consider stimulus design, context, and delivery every bit as critical to a meaningful experiment as the issues I shall discuss. Issues of experimental design In this section I shall deal with two issues: (1) pseudo- replication, that is, whether the experimental design has the statistical features needed to answer the question as posed, and (2) experiment execution, that is, possible problems arising from how the playback experiment is Communicated by R.F. Oliveira P.K. McGregor ( ✉ ) Department of Animal Behaviour, Zoological Institute, University of Copenhagen, Tagensvej 16, 2200 Copenhagen N, Denmark e-mail: pkmcgregor@zi.ku.dk acta ethol (2000) 3:3–8 © Springer-Verlag and ISPA 2000 REVIEW Peter K. McGregor Playback experiments: design and analysis Received: 23 September 1999 / Received in revised form: 24 February 2000 / Accepted: 25 February 2000