1.1 Background

As mentioned above, mentors require additional support in assessing underperforming students. According to Hunt et al. (2016), the identification of the type of support needed and how it should be provided is less clear. Furthermore, mentors needed to feel secure in failing an underperforming nursing student in a clinical assessment (Hunt et al., 2016). The authors identified that mentors required four types of support: emotional support (opportunities to attain reassurance and debrief), appraisal support (feedback and affirmation), instrumental support (time and resources), and informational support (advice, guidance, and navigation). Both Black (2011) and Hunt et al. (2016) recognized emotional support from the link lecturer, which was delivered face to face through human contact, as the primary supportive mechanism in these situations.

In a confirmatory factor analysis of a ‘Failing to fail’ questionnaire, Bachmann, Grønvik, Hauge, and Julnes (2019) confirmed a five-factor structure model in the questionnaire. The following factors were named: 1) Insufficient mentoring competence; 2) insufficient support in the working environment; 3) the emotional process dominates the assessment; 4) insufficient support from the university; and 5) Decision-making is dissociated from learning outcomes. The authors identified a significant and negative covariance between Factors 4) and 5), indicating that insufficient support from the University increased the dissociation of mentors' decision-making from learning outcomes. Based on this result, support from the link lecturer is important to prevent mentors from failing to fail underperforming students which is consistent with Hunt et al. (2016). However, a discrepancy between the findings reported by Hunt et al. (2016) and Bachmann et al. (2019) has been observed. The discrepancy appears to be the importance of emotional support. Hunt et al. (2016) emphasize the importance of emotional support, because mentors experience emotional stress during the assessment of an underperforming student. Bachmann et al. (2019) did not identify a significant correlation between the factor ‘emotional process dominates the assessment’ and any of the other factors. In a quantitative study, Hauge et al. (2019) identified factors contributing the phenomenon of failing to fail students. The factors strongly associated with failing to fail, were that the student did not put the patient's life at risk and the mentor gave the student benefit of the doubt. Hauge et al. (2019) also identified a lack of support as one of the reasons, but the mentors generally disagreed that passing failing students is associated with personal challenges and burdens. The studies by Bachmann et al. (2019) and Hauge et al. (2019) do not indicate emotional support as the primary supportive mechanism for assessing underperforming students.

Since studies are discordant in the current meanings of support, a further exploration of the type of support mentors need from the link lecturer when assessing underperforming students is necessary. The contribution of our study was to clarify and add new dimensions to the issue of support.

2.1 Aim

The aim of this study was to identify patterns of subjective viewpoints about the type of support nurse mentors need from link lecturer when assessing underperforming students. In this context, underperforming students were defined as students who are at risk of not passing the middle- or the final assessment in clinical studies.

2.2 Design

In this non-experimental exploratory study, we used Q-methodology. Q-methodology is a research approach that synthesizes qualitative and quantitative methods, which enables conversion of subjective human perspectives into objective outcomes (Dziopa & Ahern, 2011; Watts & Stenner, 2005). The methodology was developed for a scientific investigation of subjectivity, such as viewpoints, feelings, and beliefs regarding topic being investigated (Brown, 1980). This approach aims to cluster participants into groups based on their specific perspectives (Chen et al., 2016). Several researchers support the use of Q-methodology in nursing (Ha, 2015a, 2015b; Ho & Gross, 2015; Huang et al., 2019; Jueng, Huang, Li, Liang, & Huang, 2017; Simons, 2013).

2.3 Participants

We set experience in assessing underperforming nurse students as an inclusion criterion to ensure that the participants have viewpoints relevant to this study (Watts & Stenner, 2012). Brown (1980) suggested that a p-sample of 40–60 is sufficient for most studies, but a far smaller sample may be needed for some studies. However, Watts and Stenner (2012) argue that an analysis is possible with several participants less than the number of items in the Q-set. We recruited a convenience sample of 26 nurse mentors from two different nurse educational institutions.

several.

2.4 Data collection

2.4.2 Q-sorting

Since our Q-set contains less than 40 statements, we chose a nine-point (+4 ‘most agreed’ to −4 ‘most disagreed’) distribution (Brown, 1980). Our participants were likely to be experts and particularly knowledgeable of the topic and therefore we chose a flattened distribution with a 27-cell grid (Figure 1). This type of distribution provided a greater opportunity to examine fine-grained discriminations at the extremes of the distribution (Watts & Stenner, 2012). The data were collected from May–September 2018.

Before Q-sorting, participants completed the demographic data. They ranked their own subjective viewpoints of the 27 statements (#). All the Q-sort situations were audio-recorded to achieve a richer and more detailed understanding of each participants' Q-sort. After sorting, we asked them to explain the meaning of the statements placed at the extremes and whether they wanted to comment on some statements. Finally, we asked if any information was missing in the Q-set (Watts & Stenner, 2012).

2.5 Ethical considerations

The Norwegian Social Science Data Services (NSD; number 824598) approved the study. We received permission from the University College and the hospital trust to conduct the study. After approval, the participants were informed and invited to participate voluntarily. All participants signed the informed consent form.

2.6 Data analysis

A by-person factor analysis was conducted using PQ-Method 2.35 (Schmolck, 2014). Since most viewpoints of the mentors were our main concern, a principal component analysis with a varimax rotation was performed (Watts & Stenner, 2012). This rotation resulted in five discrete factors with eigenvalues greater than one, which accounted for 62% of the variance: 16%; 11%; 11%; 13%; and 11% respectively. The correlation between the factors was low (Table 1).

The factor loadings indicate the degree to which each Q-sort correlates with each of the five emerging factors (Table 2). An X marks a Q-sort loading significantly on one factor. The closer the value is to one, the more equivalent the Q-sort is to the factor.

According to Table 2, six of the 26 participants define Factor 1, three define Factor 2, four define Factor 3, five define Factor 4, and two define Factor 5. The resulting factor scores (z-scores) were converted back to the original values of the scale used in the Q-sort matrix (Table 3) to facilitate a visual inspection of the factors.

2.7 Validity

Validity in this Q-study includes content, face, and Q-sorting validity. To increase the content validity of statements, a research group of eight researchers reviewed them. Feedback during the Q-sorting stage indicated good face validity. Three mentors in a pilot study assessed the Q-sorting validity. The combination of a naturalistic and theoretical approach may strengthen the heterogeneity of the Q-set (Sæbjørnsen et al., 2016).

3.1 Characteristics of study participants

As shown in Table 4, one of the 26 participants was male. The mean age was 43 years, the mean years of experience was 15 years and each participant had mentored 15 students. Ten participants had educational credits in mentorship.

3.2 Results of the Q-sort factor analysis

The results of the factor analysis revealed five factors that accounted for 62% of the total variance. The factors were labelled as: 1) Confident in professional assessment and expects respect from link lecturer; 2) Confident about the limit but need guidance in documentation; 3) Confident in the assessment but need support in managing concerns; 4) Require knowledge and skills, not emotional support; and 5) Dialogue and collaboration rather than information. These factors are not absolute explanations of the nurse mentors' subjective viewpoints but rather our best plausible interpretation of the viewpoints included in the factor.

4.1 Subjective perspectives of support from the link lecturer

Consistent with the findings reported by Hunt et al. (2016), the central principle identified in this study is that mentors must feel secure in their roles. Based on our findings, the accessibility, approachability and willingness of link lecturer to participate in dialogue and collaboration is important. Furthermore, important elements in the dialogue are respect for the mentor's assessment and meeting their need for knowledge and skills. Hunt et al. (2016) claim that a link lecturer who fulfils their role effectively has four key attributes: accessibility, approachability, authoritative knowledge about the assessment and willingness to serve as an emotional anchor. Our findings about accessibility, approachability and meeting the mentors’ need for knowledge are consistent with the results reported by Hunt et al. (2016).

Five separate factors were identified from the perceptions of support among the surveyed nurse mentors. All mentors need some type of support from the link lecturer. A commonality of two of the perspectives, Factors 1 and 3 is that the mentors who reported these factors appear to be confident in the assessment. The mentors listing Factor 1 anticipate respect for their assessment. Strikingly different from Factor 1 and all the other perspectives, mentors who listed Factor 3 need support in managing their concerns. In contrast to Factors 1 and 3, the perspectives of Factors 2, 4, and 5 represent a need for different types of support to determine whether the assessments are acceptable, whether the student's progression/lack of progression is evident and strategies to facilitate and guide the student in the best possible manner.

The largest proportion of mentors with educational credits in mentorship listed Factors 1 and 3. Thus, mentors with educational credits in mentoring are more confident in their assessments, consistent with the findings by Bennett and McGowan (2014), Duffy (2003, 2006), Hauge (2015), and Mead, Hopkins, and Wilson (2011). Although the mentors are confident in the assessment, some require support in managing their concerns related to this complex and demanding situation. Consistent with the study by Duffy (2006), our mentors experienced challenges in communicating their concerns in a constructive and supportive manner and at the same time ensuring that the student clearly understands areas where he/she must improve. This challenge appears to persist, regardless of experience and education in mentoring. None of the mentors who defined Factors 4 and 5 had educational credits in mentorship and they expressed a need for knowledge, skills, dialogue and collaboration. Mentors listing Factor 2 included mentors with the longest experience as a nurse, have guided most students and one of three has educational credits in mentorship. Nevertheless, they need guidance to ensure that their documentation of an underperforming student is clear and accurate. The challenges associated with documentation are consistent with findings described by Bennett and McGowan (2014), Brown et al. (2012), Hauge (2015), and Jervis and Tilki (2011). Our findings emphasize that mentors express a need for different types of support, depending on their experience as a mentor and nurse, and educational credits in mentorship.

4.2 Emotional support

In contrast to previous qualitative studies highlighting the importance of emotional support (Duffy, 2006; Hauge, 2015; Hughes et al., 2016; Jervis & Tilki, 2011; Luhanga et al., 2008a), the mentors in our study did not express a need for emotional support. This finding is consistent with the studies by Bachmann et al. (2019) and Hauge et al. (2019) but differs substantially from the study by Hunt et al. (2016) who stated that emotional support is the most important type of support mechanism. The mentors admit that the situations are challenging. However, they highlight these challenges as an important part of being a mentor. They also note that experience obtained from challenging situations, as a nurse is transferable to mentoring. Although they claim that challenging situations are part of being a mentor, they emphasize the importance of patient safety and working with competent colleagues in the future, consistent by Luhanga, Larocque, et al. (2014) and what Black et al. (2014) mentioned as moral integrity.

Several explanations for the clear discrepancy between the findings reported by Bachmann et al. (2019) and Hauge et al. (2019) from the data presented by Hunt et al. (2016) regarding emotional support are plausible. Hunt et al. (2016) performed a qualitative study of a clinical assessment among nurse mentors in England. Bachmann et al. (2019) and Hauge et al. (2019) present results from quantitative studies; the current study uses Q-methodology and all of the studies explore the experiences of Norwegian nurse mentors. The difference may be explained by cultural differences or differences in the nurse education programme. Methods used to define the content of emotional support will also be important. One of our participants commented during the sorting that being taken seriously was also a form of emotional support. Hunt et al. (2016) claim that emotional support include opportunities to attain reassurance and debrief. None of our statements contained these terms and none of the participants' described these elements during the Q-sorting.

4.3 Implications for the link lecturer and the education programme

The national leadership in nursing education in Norway states that education in mentorship for the mentor is desirable and each educational institution is required to offer this type of education. In addition, national learning outcomes have been prepared for a further education of 10 credits (The Norwegian Association of Higher Education Institutions, 2018). An important question related to the assessment of underperforming students is whether the education focuses on this area. Tuomikoski et al. (2020) found that mentoring education increased across all mentoring competence areas. However, they do not clearly express assessment of weak students. Duffy (2003, 2006) argue that issues related to failing in clinical studies must be incorporated in mentor education. She claims that training in this area will allow mentors to become confident in their assessment role, identify problems early and understand the importance of documentation and continuous feedback to the student. However, the Norwegian national guidelines have not formulated learning outcomes associated with the assessment of underperforming students. Consistent with other studies (Bennett & McGowan, 2014; Duffy, 2003; Luhanga et al., 2008a), our findings confirm that the assessment of underperforming students is challenging. The methods the mentor uses to meet with, guide and assess an underperforming student are very important and may have an impact on patient safety and the professional's reputation (Luhanga, Larocque, et al., 2014). The education programme should therefore explicitly focus on the weak student. We propose that education in mentorship must implement a learning outcome related to strategies to identify, guide and assess underperforming students.

Mentors appear to be more confident in their role when support processes are formalized (Hunt et al., 2016). Our findings highlight the necessity of a link lecturer who is accessible and meets the mentors' need for knowledge and skills in assessment and mentoring. Both the organization and the reputation of clinical studies are important to ensure that a link lecturer is available. The education programme must prioritize resources to ensure that the link lecturer has the opportunity to follow up with the student and the mentor, which is a challenge in Norway. In the academic process and due to the increasing proportion of theory in nurse education, the tendency has been to cut link lecturers' resources for follow-up in clinical studies. A link lecturer typically participates in three meetings during the practice period: an expectation meeting and middle and final assessments. Because some educational institutions have cut resources for the practice follow-up, they have chosen to pull the link lecturer from the middle or final assessment. If one must choose between these two assessments, our findings emphasize the importance of prioritizing the middle assessment. Mentors are generally most uncertain whether they have a correct perspective of the student according to the learning outcomes at the middle assessment.

Our findings underscore the importance of link lecturer's knowledge and skills related to the assessment of students in clinical studies. In this context, a natural question is whether link lecturer in nursing education has this competence. Link lecturer is in demand in pedagogical education. However, our experience is that assessments in clinical studies are not included in the general pedagogical education. We therefore highlight the importance of the responsibility of nursing education for providing link lecturer with the necessary education in guiding and assessing students in clinical studies. This strategy would provide the link lecturer better conditions to meet the expectations of support from the nurse mentors, which is an important contribution to increase the quality of clinical studies for students.

4.4 Limitations and strengths

The limitations of this study are associated with the aim of a Q-study, which is to explore patterns of subjectivity. The Q-method does not intend to develop general knowledge about a population. Nevertheless, our study provides clarity and adds new dimensions to the issue of support from link lecturer in assessing underperforming students that has been explored using other methods. This study may include biases caused by the authors’ pre-conceptions. We discussed the results several times and involved our research group in the discussion related to the interpretation of our data to reduce these biases.

This study has several strengths. We assess the content validity as good. No participants missed any statement in the Q-sort. This finding supports the hypothesis that the concourse is representative of the universal perspectives of support from the link lecturer in assessing underperforming students. The participants' comments confirm the validity of our factor interpretation. The low correlation between our five factors indicates the presence of five different perspectives (Watts & Stenner, 2012).