Table 2.

Summary of the modes of influence.

Mode of influence	Nature of specific influence reporteda	Citations
Direct	Preventing or delaying dissemination of findings Changing information availability Prioritizing most desirable work Displaying preferences for output(s) above scientific rigour Framing of findings Changing staff involved in the project Diverting focus away from the project Controlling access to data, resources, funding, contracts, etc.	Yazahmeidi and Holman (2007), Gornall (2014), Katikireddi et al. (2014), The LSE GV314 Group (2014), Sedley (2016), Miller et al. (2017), Gordon et al. (2018), Ries and Kypri (2018), Williamson et al. (2019), McCrabb et al. (2021), Warin and Moore (2021)
Indirect	Using bribery Threatening job/contract loss Promoting the concept of research impact Creating fear of reporting undesired findings Encouraging research to reach wider policy audiences Encouraging the use of specific methodologies, approaches, language, dissemination channels, etc. Using IP as a tool to influence others Using the credibility of researchers as a divisive tool Creating relationships with the researchers to allow more influence on the research	Yazahmeidi and Holman (2007), Smith (2010, 2014), Haynes et al. (2011), Gornall (2014), The LSE GV314 Group (2014), Kypri (2015), Miller et al. (2017), Gordon et al. (2018), Ries and Kypri (2018), Storeng and Palmer (2019), Williamson et al. (2019), Newson et al. (2021), Warin and Moore (2021)
Subtle	Promoting societal impact through idealism Creating an environment where government ideology is shared Creating job/career insecurity Promoting a research environment where fear is present	Haynes et al. (2011), Smith (2010, 2014), Newson et al. (2021)

^aBased on the research process definition, the modes of influence have been summarized into broader categories related to their point in the research cycle using descriptive analysis to condense the modes into categories.