Loglinear modelling has been carried out to ascertain if certain diagnostic variables (patient-spouse status, being psychiatric sufferer or not), had any interaction with social variables (whether first met in medical setting, whether this was the first marriage) and psychometric variables (psychoticism, neuroticism, extroversion, social naivism, GHQ caseness) in determining one’s willingness to choose the same partner. There were altogether twenty one viable models involving part or all of each groups of variables. The interaction varied and was beyond statistical modelling at this stage.