British Journal of Clinical Psychology (2013), 52, 380–393 © 2013 The British Psychological Society www.wileyonlinelibrary.com ‘Jumping to conclusions’ in first-episode psychosis: A longitudinal study Robert Dudley 1,2 *, Kate Daley 3 , Marsha Nicholson 2 , Debra Shaftoe 2 , Helen Spencer 2 , Kate Cavanagh 4 and Mark Freeston 1 1 Clinical Psychology, Institute of Neuroscience, Newcastle University, Newcastle-upon-Tyne, UK 2 Early Intervention in Psychosis Service, Northumberland, Tyne and Wear Foundation Trust, Sunderland, UK 3 Clinical Psychology, School of Psychology, Newcastle University, Newcastle-upon-Tyne, UK 4 School of Psychology, University of Sussex, Falmer, UK Objectives. People with psychotic symptoms are reported to have a characteristic reasoning style in which they jump to conclusions. To date, little research has been conducted to investigate if this style changes over time and is associated with improvements or worsening of symptoms. This study considered these questions. Methods. Thirty-one service users were recruited from a first-episode service and completed measures of reasoning, psychotic, and non-psychotic symptomatology at two time points over 2 years. Results. Over time, people with psychosis generally became less hasty in their decision-making. Those who became less hasty in their reasoning were less symptomatic. For those who remained very hasty in their reasoning, this was associated with a worsening specifically of the delusional beliefs. Conclusions. This work supports the notion that there is a critical time in the first few years of psychosis during which symptoms and reasoning can change. However, where reasoning style does not change, this may be associated with greater difficulties associated with delusional beliefs. Practitioner points  Reasoning biases are common in people with psychosis.  The persistence of hasty reasoning may be a marker for poor outcome.  Reasoning biases can be addressed by specific treatments.  The results are based on a small sample of clinical participants.  This study did not consider if treatments for reasoning would have improved outcome of people with psychosis.  This study may have benefitted from comparison with a control group of people without psychosis. *Correspondence should be addressed to Robert Dudley, Clinical Psychology, Newcastle University, Ridley Building, Newcastle-upon-Tyne NE1 7RU, UK (email: r.e.j.dudley@ncl.ac.uk). DOI:10.1111/bjc.12023 380