Luke Douglas A, Baumann Ana A, Carothers Bobbi J, Landsverk John, Proctor Enola K
George Warren Brown School of Social Work, Washington University in St. Louis, Campus Box 1196, One Brookings Drive, St. Louis, MO, 63130, USA.
Oregon Social Learning Center, 10 Shelton McMurphey Blvd., Eugene, OR, 97401, USA.
Implement Sci. 2016 Oct 13;11(1):137. doi: 10.1186/s13012-016-0499-y.
Training investigators for the rapidly developing field of implementation science requires both mentoring and scientific collaboration. Using social network descriptive analyses, visualization, and modeling, this paper presents results of an evaluation of the mentoring and collaborations fostered over time through the National Institute of Mental Health (NIMH) supported by Implementation Research Institute (IRI).
Data were comprised of IRI participant self-reported collaborations and mentoring relationships, measured in three annual surveys from 2012 to 2014. Network descriptive statistics, visualizations, and network statistical modeling were conducted to examine patterns of mentoring and collaboration among IRI participants and to model the relationship between mentoring and subsequent collaboration.
Findings suggest that IRI is successful in forming mentoring relationships among its participants, and that these mentoring relationships are related to future scientific collaborations. Exponential random graph network models demonstrated that mentoring received in 2012 was positively and significantly related to the likelihood of having a scientific collaboration 2 years later in 2014 (p = 0.001). More specifically, mentoring was significantly related to future collaborations focusing on new research (p = 0.009), grant submissions (p = 0.003), and publications (p = 0.017). Predictions based on the network model suggest that for every additional mentoring relationships established in 2012, the likelihood of a scientific collaboration 2 years later is increased by almost 7 %.
These results support the importance of mentoring in implementation science specifically and team science more generally. Mentoring relationships were established quickly and early by the IRI core faculty. IRI fellows reported increasing scientific collaboration of all types over time, including starting new research, submitting new grants, presenting research results, and publishing peer-reviewed papers. Statistical network models demonstrated that mentoring was strongly and significantly related to subsequent scientific collaboration, which supported a core design principle of the IRI. Future work should establish the link between mentoring and scientific productivity. These results may be of interest to team science, as they suggest the importance of mentoring for future team collaborations, as well as illustrate the utility of network analysis for studying team characteristics and activities.
为快速发展的实施科学领域培养研究人员既需要指导，也需要科学合作。本文运用社会网络描述性分析、可视化和建模方法，展示了对由美国国立精神卫生研究所（NIMH）资助、实施研究机构（IRI）推动的长期指导与合作所进行评估的结果。
数据包括IRI参与者自我报告的合作与指导关系，这些数据来自2012年至2014年的三次年度调查。进行了网络描述性统计、可视化以及网络统计建模，以研究IRI参与者之间的指导与合作模式，并对指导与后续合作之间的关系进行建模。
研究结果表明，IRI成功地在其参与者之间建立了指导关系，并且这些指导关系与未来的科学合作相关。指数随机图网络模型显示，2012年接受的指导与2014年两年后进行科学合作的可能性呈正相关且具有显著性（p = 0.001）。更具体地说，指导与专注于新研究（p = 0.009）、资助申请（p = 0.003）和发表论文（p = 0.017）的未来合作显著相关。基于网络模型的预测表明，2012年每新增一个指导关系，两年后进行科学合作的可能性就会增加近7%。
这些结果支持了指导在实施科学领域以及更广泛的团队科学中的重要性。IRI的核心教员迅速且早期就建立了指导关系。IRI的研究员报告称，随着时间的推移，各类科学合作不断增加，包括开展新研究、提交新资助申请、展示研究成果以及发表同行评审论文。统计网络模型表明，指导与后续的科学合作密切且显著相关，这支持了IRI的一项核心设计原则。未来的工作应确立指导与科研生产力之间的联系。这些结果可能会引起团队科学领域的兴趣，因为它们表明了指导对未来团队合作的重要性，同时也展示了网络分析在研究团队特征和活动方面的效用。
