When deciding to conduct research on chains and networks, a decision needs to be made about the unit of analysis that will be used for data collection. The present paper examines the validity of collecting data from one party in a chain or dyad. The research examines the differences in perception between dyadic pairs of buyers and sellers about the nature of their relationship and the inter-organisational information system. There were significant differences in perceptions about the nature of the inter-organisational information system in terms of some the types of information exchanged, the frequency of exchange of different types of information, whether information was exchanged as often as necessary, the direction information flowed and the richness of the communication media used to exchange information. Customers and suppliers did not have significant differences in perception of the informality of the information systems. There were differences in perceptions about the relationship in terms of responsiveness and changes in commitment over time. In addition, there were differences in perception about their importance and loyalty to each other, and the predictability of demand and volume of supply. The implication for data collection in chain and network research is that greater accuracy will be gained by collection of data from all parties in dyadic, chain or network relationships. In addition, it would seem that data should be collected from multiple informants with at least one from each department that has dealings with the other party. Where a high level of accuracy is needed, diary records or electronic tracking may be needed to keep track of activities.