This paper takes the view that the 'level of analysis' problem, first posed by Singer (1961), has not been discussed to the extent that it deserves in the literature of International Relations. The concepts of the 'level of analysis' and the 'unit of analysis' have widely been used interchangeably. The author attempts to make a distinction between the 'level of analysis' and the 'unit of analysis', arguing that the 'level of analysis' is an issue of how to study (methodology and context) and the 'unit of analysis' is one of what to study (actor and object). It is suggested that, although a distinction is possible and necessary between those issues, they cannot completely be separated. They are actually linked to each other through the idea of 'the whole'.