Ostracism has been shown to have a variety of negative effects on the target, though the reported nature of these has varied due to differing manipulations of the phenomenon. The experiments reported attempt to characterise the consequences of computer-mediated (CM) ostracism using a range of approaches. Investigation of on-line social behaviour in Internet chat rooms revealed similar patterns of group behaviour, including ostracism, to those reported in face-to-face communication. Fourier analyses revealed periodic structure to the comments made in CM group communication qualitatively similar to that previously reported in dyadic face-to-face communication. Identity changes and ostracism occurrences were also found to be periodic. The power spectra produced by these analyses revealed that comments, identity changes and ostracism typically showed a gradient of 1/f - 11/, a characteristic common to many physical and biological systems but not previously reported in social communication. Further experiments revealed that level of anonymity did not modulate a reduction in comments made by CM ostracised participants. Thus the partial anonymity conferred by CM communication cannot account for the differential effect of ostracism in CM and face-to-face contexts. Experiments on the effect of ostracism upon cognition revealed that ostracised participants reacted significantly slower on a Strop task, but made fewer errors. Performance on a Remote Associates Task was bifurcated whereby those who accurately estimated the extent of their exclusion performed significantly worse than non-ostracised participants. Ostracised participants who under-estimated their exclusion performed similarly to non-ostracised participants. Many of these findings are consistent with a hypothesis that ostracism may be considered a form of (social) pain whose consequences may be mediated by neural substrates that partially overlap with those implicated in responses to physical pain. The findings suggest that a full characterisation of individual and situational differences in ostracism effects may require a combination of techniques, from neuron-imaging to traditional social psychological methodologies.