The findings of the present study imply that disagreement between PhD supervisors and students occurs and the nature of disagreements changes over time within the PhD education. When occurring, it can be condensed into five categories: important decisions, supervisors not being up-to-date, dubious advice, mediating between supervisors, and interpersonal relationships. Parts of this material can, with the informants’ written consent, be seen on the Internet .
Discussion about results
The nature of disagreements changes over time
As PhD education continues several years, it is not surprising that the nature of the disagreements change over time. In what particular ways it changes over time was not expressed by the informants but our implicit impression was that it was mainly due to the student’s increasing knowledge and ability to make own decisions. Early disagreements may indicate immaturity of the student while disagreements later may indicate that the student is maturing making their own decisions. Consequently, disagreements may need to be addressed differently depending on when they occur.
The severity of disagreements, whether the disagreements were resolved or what impact they may have had on the research project, were not addressed in this study. It is possible that inclusion of supervisors/PhD students with experiences of serious disagreements may have yielded quite different results. Our view is that the disagreements in the present study were mainly not of serious character.
Involvement in important decisions
Supervisors in the present study did not mind that students were involved in important decisions, but rather concerned for the consequences of such decisions. Since supervision is part of the supervisor’s career as well as the student’s, outcome is very important for the supervisor, and thereby, the department. However, Cullen et al.  noted that some supervisors left it up to the student to make major decisions. The students then, in retrospect, felt a general lack of moral support from the supervisor . The PhD students in the present study claimed autonomy and a wish to influence major decisions, but not until having matured sufficiently to feel confident enough to do so, implying that supervising seems to be a balancing act as to the degree of influence by the supervisor.
The model adopted at the university also influences degree of student involvement in important decisions. Massachusetts Institute of Technology (MIT) has developed a model for contract research with the industry. In this model, the PhD projects are short term, limited in scope, and with clearly identified milestones for delivery  leaving the student with few possibilities to be involved in important decisions. Work is directed by the supervisor and the industrial sponsor, and the student is often employed by the sponsor after finishing the PhD-education. This model often provides financial security. Another model is adopted in Cambridge. It is up to the student to form his or her own project with guidance from the supervisor. To be able to influence the research project offers a powerful motivating factor for the student . The freedom to choose research direction has also been highlighted in the model of the Vienna University of Technology, together with the requirements to have a PhD by publication, to let students work in shared offices towards joint deadlines, and to involve students in reviewing articles . The supervisors are supporting and helping, but most criticism is given by external sources . The freedom to define own research topic entails that the supervisor educate the student to take high-level decisions and become owner of the project . The shared offices may help to socialize the students. Theories of socialization have been connected to the issue of attrition in doctoral education where inappropriate socialization may be related to students departing the PhD program .
Supervisors not up-to-date
As supervisors often have more than one PhD student in progress, they cannot keep up to date the way students do. This creates disagreements as the student feels time is lost dwelling on issues already dealt with. As supervisor hours are scarce, this affects not only the student but the supervisor as well. Lauvås & Handal gives advice for optimal use of feedback in research supervision: “Do not try to conceal inadequate preparation” . PhD students need to be able to thoroughly rely on advice and comments from their supervisor. Unacceptable work ethics such as trying to conceal inadequate preparation or adopting an attitude of neglect is not usually accepted by students .
A source of disagreement between supervisors and PhD students arises when the student receives dubious advice. Inadequate knowledge and skills of the supervisor is a known criticism of PhD supervision .
Students, in the present study, expressed concerns that dubious advice was time wasting. However, students must also mature in their abilities to judge what is correct and what is dubious. A physicist interviewed by Gumport  articulated the following:
“I try to teach them a set of skills. The biggest one is to know when you’re right and when you’re wrong. It’s common for them to miss it when they’re wrong. After a while they can see it. It’s intuitive partially” .
Although the student is ultimately responsible for his or her own work , and may feel the supervisor should have all the right answers, a more nuanced picture of dubious advice is realizing that there is in fact no manual and both the supervisor and PhD student learn during the process.
Supervisors in the present study adopted a humble attitude saying they were not always sure of what was right or wrong. This is in accordance with Delamont et al.  who discussed the nature of academic supervision. They mean that the skills of academic judgement, evaluating research and assessing publications, must be learned over the course of a career and there is no manual with instructions. Confidence is of fundamental importance in the supervisory process . The supervisor needs to be confident in the student and the student needs to feel confident in the overall judgment of the supervisor, but the confidence has to be ‘informed’ not blind faith . It is then only natural that a PhD supervisor is not always precise when providing advice, but sometimes gives dubious advice.
Halse  has investigated the question of learning generated through the practice of doctoral supervision. She described that supervisors, like those in our study, learn to become a supervisor by doing it and thus represent the participatory and practice-based learning theory . In Sweden as well as in many universities across Europe, Australia and New Zeeland, professional development programs for doctoral supervisors have become mandatory. Even experienced professors should get a “driving licence” for PhD supervision. This formal training is based on the idea of the transmission model of learning. This model implies that good supervision is accomplished by attending a course, thus presuming that deficits in supervisor’s expertise can be remedied through formal, structured transmission of knowledge from instructor to learner/professor .
Previously, some supervisors tended to perceive the students as causing difficulties, making the assumption that the same supervisory resources and structures were adequate for all students; thus the structural and systemic problems that may exist were made invisible . Now supervisors’ and universities’ role have been penetrated in many studies. The idea of a learning alliance has been proposed, meaning that the goal of doctoral supervision is praxis and involves a learning alliance between each student, supervisor and university grounded in mutual respect to ensure a high quality PhD education .
Mediating between supervisors
As a rule, PhD students have one principal supervisor and one or more co-supervisors . The persons involved in PhD studies can therefore be seen as a team, with different roles. The findings from the present study suggest that it is most often the student who must take on the role of mediator between supervisors when needed. One supervisor even thought that managing different personalities could be considered a part of PhD education. This view was, however, not shared by all supervisors, whereby one supervisor expressed that mediating could be hard on the student. In any case, students are ultimately responsible for their work, and thereby responsible for making progress . By nature, the mobility of researchers is fairly high, creating geographically distributed teams and further difficulties. Disagreements in widespread teams are common , confirmed in the present study by students expressing difficulties in managing cooperation and mediating their in-house supervisor with other supervisors located at another university.
Both supervisors and students in the present study acknowledge that their relationships affect the process of the PhD education. Different personalities seem to require different behaviour. The relationship between the supervisor and student changes as the PhD education progresses. Handal & Lauvås  noted that the level of competence may shift from the supervisor to the PhD student, as the student in time acquires knowledge superior to that of the supervisor in his/her narrow field. For some supervisors, this is an affirmative event, but for some it may be threatening.
Disruptions in the relationship between student and supervisor can, according to Delamont et al. , be intellectual, personal or structural. In the case of intellectual or personal disruption, a change of supervisor is advisable as soon as possible. A structural disruption, such as the death of a supervisor, or more commonly, the transfer of a supervisor to another location, may be detrimental to the process in PhD education. Some of the supervisors in our study were careful to establish intellectual and personal boundaries between them and their students, whereas others did not mind to be personally involved with their students and eventually became friends. The duality in the supervising situation: to support and demand at the same time may generate tensions and strain within the tutoring relation . Constructive criticism is necessary if good work is to be produced, since this assists students in thinking analytically and moving forward in their development. Supervisors’ educational development throughout the Western world is located largely within an administrative framework that emphasises supervisors’ and students’ mutual roles and responsibilities . Yet some of these programs focus solely on the administrative roles and responsibilities of supervisors, attempting to provide technical “fixes” that deny the genuine difficulties and complexities involved in supervision relationships .
Discussion of methods
Limitations of the methodology used in the present study include the notion of the researcher’s knowledge of the phenomenon and openness. Human science research builds upon intersubjectivity . The researcher must acknowledge the fact that he or she interacts with the informant and thereby also has the possibility to influence the thoughts and expressions of the informant. It is by acknowledging and reflecting on past knowledge of the phenomenon that the researcher can keep an open mind, thus refraining from dominating the interview. It has been the aim of the researchers to adopt an open mind throughout data collection and analysis of the present study. Recognizing that RG, GJ and AB are supervisors in higher education, we reflected on our past knowledge and throughout the analysis striving to set it aside to benefit openness. If we have been successful, these limitations turn to strengths, as we thereby convey the experiences of the informants in a nuanced manner.
Another methodological concern is validity of the results . It has been the aim of the researchers to include informants able to convey experiences relative to disagreement between supervisors and students of higher education at university. However, as their experiences are contextual and from individual perspectives, it has also been the aim of the researchers to form a broader understanding of the phenomena, bearing in mind that no truths are presented, merely the experiences of the informants, sensitive to the receiver of the message.
Furthermore, we strived for a high level of trustworthiness of the results in the present study by reflecting on the concepts of credibility, dependability and transferability . Credibility was attained through carefulness in selection of the context, the informants and data collection. Furthermore, videotaping the interviews enabled preservation of silent expressions such as body language. Dependability was strived for by truthfulness when collecting the data in a predetermined, condensed time frame. Describing the context, informants, process of gathering data, and giving the methodological description and presentation of data as complete as possible facilitated transferability.
The researchers had an intention to achieve a strategic sample with variation in age, gender, and experience in PhD education. One female and eight male supervisors agreed to participate. It is possible that a more even spread in gender among the supervisors, would have influenced the results.
Four students were interviewed as a group due to practical circumstances. It is possible that interviewing them one by one may have changed the type of data received. However, the interviewer focused on an open attitude, leaving room for everyone to comment and freely speak about their experiences.