The longitudinal followup strategy in high-risk research is being increasingly complemented by the use of psychosis-proneness scales to select subjects for study who might be vulnerable to schizophrenia and who show differences on laboratory measures that could act as endophenotypic markers for use in genetic investigations. Three types of experimental paradigm have been adopted, drawn from cognitive psychology, psychophysiology, and the neuropsychology of hemisphere function. Results adopting each of these approaches are examined, alongside recent factor-analytic evidence that psychosis-proneness scales currently in use tap up to four different components that map onto the clinical heterogeneity of schizophrenia (and possibly other forms of psychosis). No one of these components clearly emerges as, or points to, a single indicator of risk, though some aspect of neurocognitive functioning seems a likely candidate. Even so, it is argued, the clinical expression of vulnerability must be due to a convergence in an individual of several components of risk since individually (and notably so for "susceptibility to positive symptoms") they are very common in the healthy population. In evaluating the evidence, attention is drawn to two crucially different ways that investigators in schizophrenia research have construed the notion of continuity (1) as subclinical defect (or forme fruste of disease) having varying expression or (2) as biologically based personality dimensions that simultaneously describe the dispositions to aberrations of function leading to degree of illness. It is noted that the model of continuity chosen can significantly shape the way the results of risk research are interpreted and the theories of psychosis to which they give rise.