The PRIME Theory of motivation and its application to smoking ...

Improving replicability in addiction research Robert West University College London 1 The problem Replicability is the extent to which reported study findings are repeated when what is believed to be essentially the same study is repeated It is hampered by

Poorly defined constructs (e.g., craving, stress, recovery) Poorly specified models (e.g., unclear causal pathways) Weak measurement (e.g., use of self-report scales with low validity) Inconsistent/incomplete reporting (e.g., regarding intervention content) Inappropriate analyses (e.g., cherry picking comparisons) Bias in reporting and dissemination (e.g., file drawer problem) 2 Clarifying constructs Develop an addiction ontology setting out key terms, definitions and relationships using an established model Small section of

putative addiction ontology 3 Specifying models Develop a common model building system that clearly sets out what constructs are included, how they are defined and how they relate to each other Transtheoretical Model expressed in standard form 4

Improving measurement Develop and use a database of measures with indices of reliability and validity Construct label Measure name Validity index Sources Cigarette craving QSU

7 Tiffany 1990 Cigarette craving SUTS 7 West 2010 State anxiety STAI

8 Spielberg 1980 5 Improving reporting Develop and use an online study reporting tool for study reporting to ensure that key information is presented in a consistent way Meta-data Research questions Study ID: ADD-17203 Design Participants

Select which type of study feature you want to enter, then you will taken through a step-bystep process to enter the information Setting Groups Measures Analyses Findings Interpretation 6

Improving analysis Promote the use of the Open Science Framework and other resources to pre-specify analysis plans and record analysis history 7 Reducing reporting bias Extend trial registration to all major study types and establish an efficient system for linking study reports to registered protocols, automatically recording protocol deviations Extend competing interest reporting to non-commercial types and automate the process of coding these to allow their assessment in meta-regressions 8 Next steps

Addiction ontology Model building system Measures database Study reporting tool Open Science Framework Study registration Competing interest reporting 9