The quality of the relationship between academic research supervisors and their students has been recognized as the most important determinant of successful and timely postgraduate degree completion. Many functions have been deemed important in facilitating this alliance. However, due to resource constraints and personal capabilities, supervisors must prioritize certain functions over others. It is, therefore, important to know the relative value of supervisory functions. The current study employed choice-based conjoint methodology to investigate the importance of different supervisory functions to Australian postgraduate students. These methods are notable for overcoming many of the drawbacks associated with conventional survey methods. Participants included 570 postgraduate students who completed 10 choice tasks comprising 16 attributes, each 3 levels, using a partial profile design. Preferences were ascertained using importance and utility scores. Latent class analysis was used to probe the presence of distinct sample segments whereas randomized first choice simulations were carried out to estimate the sample’s endorsement of hypothetical supervisors. Results revealed three key findings: (a) students valued academic integrity, constructive feedback, open communication, and bonding as the most preferred supervisory attributes; (b) student preferences were similar regardless of background differences and the sample was best classified as a single group; and (c) students preferred supervisors who fostered caring/supportive relationships over those who focused more strictly on instrumental functions. These results suggest that supervisors and training programs should focus on the interpersonal aspects of supervision and supporting the psychosocial needs of students. (PsycINFO Database Record (c) 2019 APA, all rights reserved)

中文翻译：

---

研究监督的基本要素：离散选择实验。

学术研究主管与其学生之间关系的质量已被公认为成功和及时完成研究生学位的最重要决定因素。许多职能被认为对促进这种联盟很重要。然而，由于资源限制和个人能力，主管必须优先考虑某些功能。因此，了解监管职能的相对价值非常重要。目前的研究采用基于选择的联合方法来调查不同监督职能对澳大利亚研究生的重要性。这些方法以克服与传统调查方法相关的许多缺点而著称。参与者包括 570 名研究生，他们完成了包含 16 个属性的 10 个选择任务，每 3 层，采用局部轮廓设计。使用重要性和效用分数确定偏好。潜在类别分析用于探测不同样本片段的存在，而随机首选模拟则用于估计样本对假设主管的认可。结果揭示了三个主要发现：(a) 学生将学术诚信、建设性反馈、开放式沟通和凝聚力作为最喜欢的监督属性；(b) 无论背景差异如何，学生的偏好都是相似的，最好将样本归类为一个群体；(c) 学生更喜欢培养关怀/支持关系的导师，而不是那些更专注于工具功能的导师。这些结果表明，督导者和培训项目应该关注督导和支持学生的社会心理需求的人际关系方面。（PsycINFO 数据库记录 (c) 2019 APA，保留所有权利）