Chemical engineering students’ perceptions of peer and self-assessment

ABSTRACT

This study explores Finnish chemical engineering students’ peer and self-assessment perceptions. It is divided into two substudies. In Study I, we investigate students’ perceptions of peer and self-assessment based on their responses to open questions and Likert-items adopted from existing literature. We identify six perception dimensions from these open responses: Learning and reflection, Motivation and participation, Self-efficacy, Exhaustion and workload, Cynicism, and Self-criticism. In Study II, we develop a new questionnaire for engineering students’ peer and self-assessment perceptions, as pre-existing questionnaires fail to adequately capture these dimensions. Exploratory factor analysis is used to investigate how student perceptions form larger factors, and how they relate to theoretically connected pedagogical measures like study-related burnout, self-efficacy, and learning approaches. We find that negative perceptions correlate with study-related burnout and an unreflective approach to learning. Meanwhile, positive perceptions of self-efficacy and learning in peer and self-assessment are associated with the deep approach, underlining the importance of facilitating positive assessment experiences in students.

KEYWORDS:

• Peer assessment
• self-assessment
• peer assessment perceptions
• self-assessment perceptions
• peer learning/teaching

Introduction

The rapid pace of development in engineering has underlined the importance of training our graduates to become lifelong learners (Guest 2006). To be lifelong learners, students should be able to employ deep learning strategies and self-regulate their learning and motivation (Malan and Stegmann 2018). Consequently, engineering educators should adopt teaching and assessment strategies that facilitate both learning and developing such skills, including peer and self-assessment (Nieminen, Asikainen, and Rämö 2021; Wanner and Palmer 2018).

In general, any assessment decision includes two steps: identifying assessment criteria and determining how well the assessed work meets these criteria (Boud and Falchikov 1989). Self-assessment involves learners in judging their own learning, especially their learning outcomes or achievements (Boud and Falchikov 1989). Meanwhile, in peer assessment students make these evaluations of similar status peers (Topping 1998). These general definitions encapsulate assessment practices including peer- and self-review; peer and self-evaluation; peer and self-grading and peer feedback (Panadero and Alqassab 2019). They do not, however, delineate the respective roles the teachers and students play in determining the assessment criteria or processes.

Peer and self-assessment have been employed in engineering education to assess tasks including oral presentations (Liow 2008; Nikolic, Stirling, and Ros 2018), written reports (Andersson and Weurlander 2019; Rafiq and Fullerton 1996 Van Hattum-Janssen, Pacheco, and Vasconcelos 2004;), and mathematical problems (Boud and Holmes 1981; O’moore and Baldock 2007). While the validity and benefits of peer and self-assessment have largely been established by previous research (Ashenafi 2017; Falchikov and Boud 1989; Falchikov and Goldfinch 2000; Hadzhikoleva, Hadzhikolev, and Kasakliev 2019; Topping 1998; Wanner and Palmer 2018; Yan and Brown 2017), the literature on students’ peer and self-assessment perceptions is scarcer (Andrade and Du 2007; Asikainen et al. 2014; Li, Huang, and Cheng 2022; Planas Lladó et al. 2014; Wanner and Palmer 2018).

Students’ perceptions of peer and self-assessment depend on how the assessment is implemented (Planas Lladó et al. 2014) and might serve as important mediators for the benefits associated with these practices. Consequently, carefully constructed questionnaires of student perceptions are necessary to identify successful practices, as well as the specific advantages of a given practice. Unfortunately, engineering education researchers, and the peer and self-assessment research community in general, have typically utilised self-developed surveys to gauge student impressions with little statistical consideration for their reliability or the underlying conceptual structures (Andersson and Weurlander 2019; Boud and Holmes 1981; Chang and de Lemos Coutinho 2022; González De Sande and Godino-Llorente 2014; McGarr and Clifford 2013; Nikolic, Stirling, and Ros 2018; O’moore and Baldock 2007; Planas Lladó et al. 2014; Van Hattum-Janssen and Lourenço 2008; Willey and Gardner 2010). When such analyses have been conducted, the Cronbach’s α (Cronbach 1951) reliabilities of the emergent subscales have been as low as 0.28 (Moneypenny, Evans, and Kraha 2018), which raises concern over poor interrelatedness or an insufficient number of factor items (Taber 2018). Thus, in this study we examine chemical engineering students’ perceptions of peer and self-assessment to develop a carefully validated perception questionnaire. The article is divided into two substudies: Study I focuses on the qualitative investigation of students’ peer and self-assessment perceptions and Study II on the development of our Perceptions of peer and self-assessment (PPSA) questionnaire.

Study I – exploring students’ peer and self-assessment perceptions

In a recent review, Andrade (2019) called for more research on the cognitive and affective mechanisms of self-assessment. To answer this call and to provide groundwork for a carefully validated measure of student perceptions on peer and self-assessment, we investigated these perceptions in a chemical engineering course. Our key research questions in Study I are ‘What kinds of perceptions do students describe after participation in peer and self-assessment?’, and ‘Do these perceptions form broader categories with underlying communalities?’

In general, research suggests that students wish to be involved in the assessment process and that their perceptions of peer assessment skew towards the positive (Andersson and Weurlander 2019; Bazvand and Rasooli 2022; Carvalho 2013; Mulder et al. 2014; Nikolic, Stirling, and Ros 2018; Vickerman 2009; Wen and Tsai 2006). For example, students maintain that participation in peer assessment increases their understanding of the assessment criteria (Boud and Holmes 1981; Ndoye 2017; Wanner and Palmer 2018), helps them learn better (Asikainen et al. 2014; Moneypenny, Evans, and Kraha 2018; Mulder et al. 2014; Mulder, Pearce, and Baik 2014; Partanen 2020), develops critical thinking skills (Andersson and Weurlander 2019), improves motivation (Hanrahan and Isaacs 2001; Planas Lladó et al. 2014; Van Hattum-Janssen and Lourenço 2008), and enhances communication skills and autonomy (Moneypenny, Evans, and Kraha 2018; Shen, Bai, and Xue 2020). However, students can also find peer assessment time consuming and intimidating (Hanrahan and Isaacs 2001; Moneypenny, Evans, and Kraha 2018; Sluijsmans et al. 2001) and concern over the validity and fairness of peer assessment is commonplace (Andersson and Weurlander 2019; Carvalho 2013; Kaufman and Schunn 2011; Patton 2012; Van Zundert, Sluijsmans, and van Merriënboer 2010; Wanner and Palmer 2018). Some further worry about the impartiality and competence of their peer-assessors (Carvalho 2013; Kaufman and Schunn 2011; Mulder, Pearce, and Baik 2014; Willey and Gardner 2010) or their own competence as assessors (Van Hattum-Janssen and Lourenço 2008). Accordingly, students can be more supportive of peer assessment as a formative rather than a summative exercise (Patton 2012), although some feel that they should not be forced to do the teacher’s job (Van Hattum-Janssen and Lourenço 2008; Willey and Gardner 2010).

Perceptions of peer assessment are impacted by its implementation (Planas Lladó et al. 2014), the field of study (McGarr and Clifford 2013), and the assessor’s gender (Moneypenny, Evans, and Kraha 2018; Wen and Tsai 2006). Participation in peer assessment seems to improve students’ attitudes towards it (Mulder et al. 2014; Van Zundert, Sluijsmans, and van Merriënboer 2010), plausibly due to improvements in assessment efficacy. However, some studies have obtained conflicting results, with students’ attitudes decreasing or becoming more polarised after peer assessment (Mulder, Pearce, and Baik 2014; Planas Lladó et al. 2014). Kaufman and Schunn (2011) suggest that students’ perceptions of fairness are chiefly determined by whether peer assessment was perceived as useful and positive. Overall, reservations over fairness and validity may be mitigated with measures such as anonymity, individualised assessment of group work, inclusion of instructor assessment, and sufficient training and guidelines (Hanrahan and Isaacs 2001; Kaufman and Schunn 2011; Planas Lladó et al. 2014; Sridharan, Muttakin, and Mihret 2018; Vickerman 2009).

Though scarcer, the self-assessment literature similarly indicates that students tend to possess positive perceptions towards self-assessment (Andrade and Du 2007; Hanrahan and Isaacs 2001; Micán and Medina 2017; Wanner and Palmer 2018), especially if they are actively engaged in formative self-assessment (Andrade 2019). University students also feel that self-assessment helps them take responsibility for learning (Bourke 2014; Lopez and Kossack 2007; Ndoye 2017) and apply newfound skills (Murakami, Valvona, and Broudy 2012). According to students, it additionally guides evaluation and revision of work (Micán and Medina 2017), fosters self-regulated learning and reflection (Wang 2017), and facilitates critical thinking (Siow 2015; van Helvoort 2012). Furthermore, self-assessment practices can increase students’ intrinsic motivation by supporting autonomy (Gholami 2016). On the negative side, some experience self-assessment as frustrating and difficult when the evaluation criteria are not clear (Andrade and Du 2007; Hanrahan and Isaacs 2001). However, as in peer assessment, students’ attitudes tend to improve as they gain more self-assessment experience (Andrade and Du 2007).

To summarise, there appear to be many parallels between student perceptions of peer and self-assessment. Indeed, in some learning contexts, such as peer and self-assessment of contributions to a group task, it might not even make sense to separate the two. Consequently, past studies have often studied peer and self-assessment simultaneously (Hanrahan and Isaacs 2001; Ndoye 2017; Nikolic, Stirling, and Ros 2018; Van Hattum-Janssen and Lourenço 2008; Wanner and Palmer 2018; Willey and Gardner 2010). In Study I, we seek to understand how these literature trends are reflected in chemical engineering students’ perceptions of peer and self-assessment and whether these perceptions form broader underlying categories.

Methods

Course background

The peer and self-assessment perception data were collected from a bachelor’s level Chemical thermodynamics course at Aalto University in Finland in 2019. The course was an obligatory part of the chemical engineering curriculum. Although some courses at the university utilise peer and self-assessment, it is still a relatively novel and infrequently encountered teaching tool for the chemical engineering students.

Six sets of course problems were graded through peer and self-assessment, which constituted 10% of the course grade. The peer and self-assessment scheme shared elements with the Peer Assessment Learning Sessions proposed by O'moore and Baldock (2007): The problem deadlines were spread throughout the course and relayed to the students from the start. The actual problems were published at least a week before the deadline. During this time, students could participate in up to three walk-in study halls where they could work on the problems collaboratively, with teaching staff at hand to provide aid. Each problem set typically consisted of 2–3 problems with subtasks. These subtasks spanned the second and third knowledge categories and third through fifth cognitive processing categories of the Revised Bloom's taxonomy (Anderson and Krathwohl 2001).

The students submitted their solutions digitally to the course’s online learning platform where they also performed the peer and self-assessment. Each student assessed their own solutions and those of two peers. All peer assessments were conducted anonymously. Clear assessment instructions, an assessment rubric, and model solutions were available to guide the assessment process. These were created by the course instructor, so the students did not participate in the shaping of the assessment criteria or practices. The peer and self-assessments combined grading with a formative component: In addition to numerical marks, students were obligated to provide written feedback and justifications for their evaluation in an open response field. For example, when students deducted marks, they were instructed to include an explanation of these deductions that was based on the assessment rubric. The open response fields were only automatically checked for whether they contained text or not. The content or quality of student feedback was not examined.

The final mark was calculated as an average of the three assessments students received. If the minimum and maximum marks differed substantially, a course instructor provided the ultimate verdict. Students received credit based both on the number and quality of their numerical assessments. The teacher also provided general feedback based on trends observed in the solutions.

Data collection

An outline of Study I is provided in Figure 1. The data was gathered from an online survey that the students responded to at the end of the course. It included two open questions and the first version of the Perceptions of peer and self-assessment (PPSA1) questionnaire. The questionnaire contained 26 items on a five-point Likert scale with higher values indicating more agreement. Participants received a small amount of credit, irrespective of whether they provided research permission. They were also informed that they could renege research permission at any time with no detrimental effects on their course performance. The course homepage contained detailed information on the research. Based on national guidelines from the Finnish Advisory Board on Research Integrity, the research did not need a prior ethical review statement from the ethical board as it did not utilise sensitive personal information, participants were legally competent adults, and the research did not risk the physical integrity or cause mental harm or threats of safety to the participants beyond everyday life.

Figure 1. Outline of the development of the PPSA questionnaire through Studies I and II.

The PPSA1 items are shown in Figures 2 and 3. They were modified from existing peer assessment questionnaires (Planas Lladó et al. 2014; Wen and Tsai 2006), although we added one additional item regarding autonomy to the self-assessment part. The open questions included ‘What kind of thoughts and emotions did peer assessment and self-assessment evoke in you?’ and ‘How have peer and self-assessment affected your studying and learning?’

Figure 2. Percentages of different Likert-responses for perceptions of peer assessment (PA) in the PPSA1 and the mean values for each question. The response range at the bottom of the Figure is from Strongly disagree (= 1) to Strongly agree (= 5). Meanwhile, the superscripts after the items on the left indicate their source with 1 = Planas Lladó et al. (2014) and 2 = Wen and Tsai (2006).

Figure 3. Percentages of different Likert-responses for perceptions of self-assessment (SA) in the PPSA1 and the mean values for each question. The response range at the bottom of the Figure is from Strongly disagree (= 1) to Strongly agree (= 5). Meanwhile, the superscripts after the items on the left indicate their source with 1 = Planas Lladó et al. (2014) and 2 = Wen and Tsai (2006).

Of the 149 enrolled 134 (90%) provided research permission. Meanwhile, 114 (96%) of the 119 students who did not drop out during the course responded to the end-of-course survey. Of these, 103 provided open responses. Half (52%) of the participants were females. Their mean age was 21.4, with a range of 19–35 years. Most (78%) were doing their second year.

Analysis

The data were analysed through abductive content analysis which implies a shift in analysis between inductive and deductive approaches (Graneheim, Lindgren, and Lundman 2017). In our case, the categories emerged from the data but theory was also used in the categorisation process. All responses were independently read through by two of the authors to identify common characteristics. Whenever a characteristic was recognised, that text segment was coded, and the segments showing similar characteristics were grouped. One response could potentially include multiple characteristics.

Results

Students’ perceptions of peer and self-assessment in the open responses

Initially, students’ perceptions of peer and self-assessment were categorised as positive (35%), negative (34%), both positive and negative (27%), or neutral (4%). Neutral responses were typically short and noncommittal, such as ‘It was ok’ or ‘I have no opinion.’ Next, three subcategories of positive and negative experiences were identified: The first positive category was Learning and reflection, which included participants’ expressions of peer or self-assessment improving learning. The second positive category, Motivation and participation, consisted of expressions showing motivation, inspiration, and increased involvement. The third positive category, Self-efficacy, described expectations of capability and success in assessment. Meanwhile, the first negative category, Exhaustion and workload, contained perceptions of assessment taking too much time while being exhausting. The second negative category, Cynicism, comprised of expressions of not seeing value in peer and self-assessment. The third negative category, Self-criticism, covered descriptions of assessment being difficult. Students in this category also doubted their abilities to assess or felt uncomfortable exposing their solutions to peers. Table 1 displays examples of both peer and self-assessment perceptions in all categories.

Table 1. Examples of characteristic responses in different open response categories for peer and self-assessment in the questionnaires.

Category	Example (Peer assessment)	Example (Self-assessment)
Learning and reflection	When you familiarise yourself with the assessment criteria and model solutions, learning is kind of inevitable	Self-assessment was good when you had to go through again the assignments you had done, and you saw if something had gone wrong
Motivation and interest	Peer assessment was interesting and useful, the course felt different from other courses, because teachers’ evaluation was not the only assessment method	Self -assessment was nice change to traditional assessment
Self-efficacy	I believe that the peer assessment will go well	I believe I am fair and just in my assessment, also towards myself
Exhaustion and workload	Workload was too big	I experienced it … solely as increasing workload
Cynicism	I doubt that many students can assess accurately	Is this worth it, people may just get frustrated
Self-critisism	It’s hard to assess other work when you did not succeed in it yourself	It’s hard to be objective when conducting self- assessment

Questionnaire results

Student responses to the 26 questions of the PPSA1 are summarised in Figures 2 and 3 which show the percentages of the five Likert categories from Strongly Disagree (= 1) to Strongly Agree (= 5) and the means for each of these categories as black dots. In accordance with the qualitative results, student experiences of peer and self-assessment varied widely with all the means in the middle quintile. Meanwhile, the Strongly Disagree and Strongly Agree categories typically constituted less than 15% of the responses. The Neither Agree nor Disagree category was particularly prevalent for the two questions related to the subjectivity and demandingness of the assessor for both peer and self-assessment. An explorative factor analysis of the PPSA1 results failed to produce a reasonable factor structure for the questionnaire.

Discussion

In Study I, we investigated chemical engineering students’ perceptions of peer and self-assessment. These perceptions were initially categorised as either positive, negative, or neutral. Then, three subcategories were identified for both positive and negative perceptions.

The positive perception subcategories were Learning and reflection, Motivation and participation, and Self-efficacy. Student responses in these categories highlighted how peer and self-assessment increased motivation and understanding of the assessment criteria. Many also described how peer and self-assesment improved learning by facilitating reflection, critical thinking, and review. These positive perceptions closely echoed the ones found in the literature review at the beginning of this study. For example, civil engineering students in Van Hattum-Janssen and Lourenço’s (2008) study also felt that peer assessment increased their motivation to study. Meanwhile, electrical engineering students in Andersson and Weurlander’s (2019) study described how seeing others’ work benefitted their learning, similarly to the students in this study.

The negative perception subcategories were Exhaustion, Cynicism, and Self-criticism. Perceptions in these categories included concerns over increased workload, their own or peers’ ability to assess, and the impartiality of both peer and self-assessment. These were also in close agreement with the existing literature on student peer and self-assessment perceptions. For instance, akin to the engineering students in Willey and Gardner’s study (2010), several students questioned the competence of their peer-assessors. In the meantime, others suspected their own capability to assess, in line with Van Hattum-Janssen and Lourenço’s (2008) results.

Even though there was evidence of all six identified perception subcategories in the literature, existing measures of student’s peer and self-assessment perceptions miss some of them. For example, the questionnaire by Wen and Tsai (2006) seemingly fails to tap into the self-efficacy and self-criticism aspects of student perceptions. On the other hand, both the interdisciplinary questionnaire by Planas Lladó et al. (2014) and our PPSA1 lack questions in the Exhaustion and workload category. There might also be some structural issues with these questionnaires, as a recent study (Moneypenny, Evans, and Kraha 2018) showed Cronbach’s $\alpha$ reliabilities below 0.4 for two of the four subscales in the questionnaire by Wen and Tsai (2006). Even though Cronbach’s $\alpha$ reliabilities are sample-specific (Taber 2018) and Wen and Tsai themselves report values above 0.6 for all subscales, this raises concerns over poor interrelatedness or an insufficient number of factor items in the questionnaire (Taber 2018). These shortcomings in existing questionnaires combined with the failure of PPSA1 to produce a reasonable factor structure necessitated an overhaul of the PPSA1 based on the perception subcategories.

Study II – developing the perceptions of peer and self-assessment -questionnaire

In Study I, we found that current instruments failed to capture key aspects of chemical engineering students’ peer and self-assessment perceptions. These results underscored the need for a carefully constructed and validated student perception questionnaire that includes all six of the student perception subcategories identified in Study I. In Study II, we aim to develop such a questionnaire. We validate our questionnaire by looking at the connections between peer and self-assessment perceptions, learning approaches, self-efficacy, and study-related burnout.

Students’ predispositions towards a particular learning approach colour their perceptions of assessment (Struyven, Dochy, and Janssens 2005). Two learning approaches have predominated the literature: the deep approach and the unreflective approach (Entwistle 2009; Entwistle and McCune 2004; Lindblom-Ylänne, Parpala, and Postareff 2019). Students high in the deep approach tend to be intrinsically motivated to understand the topic, focus on critical thinking, and seek to connect new material with pre-existing knowledge. Meanwhile, students displaying the unreflective approach tend to utilise repetitive learning strategies to memorise key aspects of the course, which results in a fragmented knowledge base.

The connection between peer and self-assessment and the deep approach appears to be bidirectional in the sense that the mere participation in peer and self-assessment fosters a deeper approach (Lynch, McNamara, and Seery 2012; Nieminen, Asikainen, and Rämö 2021). Students applying the deep approach also tend to perceive the teaching-learning environment more positively, whereas students applying the unreflective approach hold more negative views (Parpala et al. 2010). This implies that positive perceptions of peer and self-assessment are more likely associated with the deep approach and negative perceptions with the unreflective approach. Students applying the deep approach also seem to benefit more from peer and self-assessment (Yang and Tsai 2010) and are likely better equipped to recognise the benefits associated with peer and self-assessment, which may further catalyse positive attitudes toward these activities.

In addition to impacting students’ learning approaches (Scouller 1998; Struyven, Dochy, and Janssens 2005), assessment practices like peer and self-assessment can impact students’ self-efficacy (Bandura 1997; Van Dinther, Dochy, and Segers 2011). Self-efficacy is defined as a person’s perception of their capability to organise and execute required actions to achieve a desired outcome (Bandura 1997). It is one of the most central non-intellective determinants of academic performance (e.g. Richardson, Abraham, and Bond 2012; Schneider and Preckel 2017). Nieminen, Asikainen, and Rämö (2021) demonstrated that summative self-assessment can promote self-efficacy in addition to facilitating a deep approach to learning. In general, self-assessment practices seem to improve particularly female students’ self-efficacy, although the effect size depends on the implementation (Panadero, Jonsson, and Botella 2017).

In peer and self-assessment, self-efficacy can be impacted through experiences of successful performance or through seeing others perform (Bandura 1997; Van Dinther, Dochy, and Segers 2011). Indeed, some students report that peer assessment improves learning because it lets them see alternative solutions, while others think seeing these solutions helps them better estimate their own level of understanding (Partanen 2020). Based on the results of Study I, students seem to possess a sense of peer and self-assessment efficacy, as exemplified by their perceptions of increased understanding of the assessment criteria and fears over their ability to evaluate others. This narrow sense of peer and self-assessment efficacy is likely correlated with the more general sense of self-efficacy.

As perceived workload is intimately connected to feelings of exhaustion, students’ perceptions of peer and self-assessment as time-consuming (Hanrahan and Isaacs 2001; Moneypenny, Evans, and Kraha 2018) likely contribute to study-related burnout. Inexperience, fear of hurting others, and the fear of being hurt can also make peer and self-assessment more stressful than traditional assessment, further contributing to feelings of burnout, especially for female students (Pope 2005). Study-related burnout is defined as a syndrome with three key components: emotional exhaustion that is the outcome of high perceived study demands, the development of a cynical and detached attitude towards one’s studies, and feelings of inadequacy (Salmela-Aro et al. 2009). These components bear a striking similarity the negative subcategories of Exhaustion, Cynicism, and Self-criticism identified from student perceptions in Study I. Thus, while peer and self-assessment can lead to improved performance, for students with negative perceptions it might also contribute to study-related burnout, especially for students applying an unreflective approach (Asikainen et al. 2020).

Based on the above literature review, we hypothesise first that negative experiences of peer and self-assessment are linked to the unreflective approach to learning and all three components of study-related burnout. Second, positive experiences should be connected to the deep approach to learning and high self-efficacy. We use these hypotheses to provide convergent validity for our questionnaire. Therefore, our central research questions in Study II are: ‘Do students’ peer and self-assessment perceptions form theoretically relevant factors?’ and ‘Do these factors relate to existing measures of learning approaches, self-efficacy, and study-related burnout in a reasonable way?’

Methods

Data collection

An outline of Study II can be found in Figure 1. The data was collected from a bachelor’s level Chemical thermodynamics course at Aalto University in Finland in 2020. The data collection practices and the peer and self-assessment tasks were similar to the ones in described in Study I. One notable difference between the years was that in 2020, each problem set was subdivided into smaller assessment tasks so that students didn’t have to assesses problems they had not attempted.

The students responded to a pre-survey at the beginning of the course, and a post-survey at the end of the course. The pre-survey contained a 5-item measure for general self-efficacy utilising a five-point Likert scale, which originated from the Motivated Strategies for Learning Questionnaire (MSLQ; Pintrich 1991). The post-survey included a 12-item measure for learning approaches with a five-point Likert scale (Parpala and Lindblom-Ylänne 2012), and a 9-item study-related burnout measure with a six-point scale (Salmela-Aro et al. 2009; Salmela-Aro and Read 2017). It also included the second version of the Perceptions of Peer and Self-Assessment -questionnaire (PPSA2). The PPSA2 contained 46 items with a five-point Likert scale and was based on the six open response subcategories of Study I. It is shown in Appendix 1. In all measures, higher values on the Likert scale indicate more agreement. The post-survey also included the open questions from Study I, but the questions were asked separately for peer and self-assessment.

Of the 155 enrolled students, 145 (94%) provided research permission. Meanwhile, 124 (83%) of 149 students who did not drop out during the course responded to both surveys. Of these, 68 and 48 provided open responses on their peer and self-assessment perceptions, respectively. Most (79%) of the participants were in their second year, and half (49%) were females. The mean age was 21.9, with a range or 19–43 years.

Analysis

The open response data from 2020 was abductively content analysed to verify the categorisation that had emerged in Study I. After this, the broader structures emerging from the PPSA2 were investigated with explorative factor analyses (EFAs). The data included the 124 respondents from the 2020 post-survey. There were no missing values. Factor stability was investigated using Bartlett’s test for sphericity (Bartlett 1954) and the Kaiser-Mayer-Olkin (KMO) measure for sampling adequacy (Kaiser 1974). We used both parallel analysis and the Velice’s minimum average partials (VMAP) -criterion to deduce the optimal number of factors to extract, as recommended by recent reviews in EFA (Gaskin and Happell 2014; Howard 2016; Watkins 2018). The final analysis utilised the principal axis method with direct oblimin (quartimin) rotation using polychoric correlations (Howard 2016; Watkins 2018).

After identifying the optimal number of factors, the number of scale items was reduced using weak and multiple factor loadings. The resulting scales were further refined based on the Cronbach’s α (Cronbach 1951), and McDonald’s ω_t (McDonald 1999) values of the individual factors and the test as a whole. McDonald’s ω_t indices were employed due to the multiple issues associated with the use of Cronbach’s α measure for reliability (Dunn, Baguley, and Brunsden 2014). A minimum of 3 items were required per factor, as recommended by Watkins (2018). Sum scales were formed based on the factor solutions and Pearson correlations were used to explore the relationship between these, learning approaches, self-efficacy, and study-related burnout.

Results

PPSA2 factor analysis

The statistical indicators for the EFAs are summarised in Table 2. Prior to the EFA, the PPSA2 included 24 items for peer assessment and 22 for self-assessment. The statistically significant values (i.e. below p = 0.05) on Bartlett’s test, combined with KMO values higher than 0.7 in Table 2 demonstrate that the sample was suitable for factor analysis (Watkins 2018). Meanwhile, the parallel analysis and VMAP values show that optimal number of factors was three. After the number of scale items was reduced, the final PPSA3 contained 15 items for peer and self-assessment each. The Cronbach’s $\alpha$ values were 0.88 for peer assessment and 0.90 for self-assessment, while the McDonald’s ω_t indices were 0.92 and 0.94, respectively.

Table 2. The relevant statistical EFA indicators for the PPSA2 and the PPSA3.

	No. items	Barlett	KMO	Parallel	VMAP	α	ωt
Peer assessment PPSA2	22	<0.001	0.90	3	3	0.92	0.94
Self-assessment PPSA2	24	<0.001	0.87	3	3	0.93	0.95
Peer assessment PPSA3	15	<0.001	0.88	3	3	0.88	0.92
Self-assessment PPSA3	15	<0.001	0.85	3	3	0.9	0.94

The final three-factor solutions were similar for peer and self-assessment experiences: Of the original six PPSA2 subcategories, the positive experience subcategories of Learning and reflection, and Motivation and interest collapsed into a single factor in PPSA3, as did all three negative experience subcategories. The factors for peer assessment were named Peer assessment as supporting learning (PF1), Self-efficacy in peer assessment (PF2), and Negative experiences of peer assessment (PF3). Analogous names and labels were used for self-assessment. The factor loadings, factor means, standard deviations, and McDonald’s ω_t reliabilities are reported in Table 3 for peer assessment and in Table 4 for self-assessment. For both peer and self-assessment perceptions, the emerging factors possessed ${\omega }_t$ values above 0.82.

Table 3. The factors structure of the peer assessment items in PPSA3, together with the means, standard deviations, and McDonald’s omega values of the factors.

	PF1	PF2	PF3
McDonald’s ωt	0.90	0.80	0.82
Mean (sd)	3.07 (0.89)	3.80 (0.70)	2.35 (0.72)
Peer assessment supported my learning in the course	0,795
Peer assessment motivated me to engage more deeply with course tasks.	0,860
Peer assessment made me participate more in course activities.	0,703
Peer assessment helped me recognise my errors.	0,735
Peer assessment increased my motivation to study in the course.	0,859
Peer assessment made me review the course contents more.	0,766
I was able to assess the performance of my peers with the provided assessment rubrics and model solutions		0,862
I was able to apply the assessment rubric to mark my peers’ solutions.		0,866
I understood how the tasks were supposed to be solved when I compared them to the assessment rubric and model solutions.	0,260	0,499
Peer assessment is unfair because the effort students put into it varies a lot			0,734
The criteria for peer assessment felt unjust when compared to the tasks.		−0,280	0,610
I felt that I could not assess others’ work reliably.			0,537
I feel that many students were not able to assess their peers’ work.			0,736
Peer assessment increased the workload too much	−0,243		0,612
In peer assessment, the students are forced to do the teacher’s work as well.			0,690

PF1 = Peer assessment as supporting learning –factor.

PF2 = Self-efficacy in peer assessment –factor.

PF3 = Negative experiences of peer assessment –factor.

Table 4. The factors structure of the self-assessment items in PPSA3, together with the means, standard deviations, and McDonald’s omega values of the factors.

	SF1	SF2	SF3
McDonald’s ωt	0.88	0.82	0.89
Mean (sd)	3.19 (0.81)	3.73 (0.71)	2.17 (0.82)
Self-assessment supported my learning in the course	0,902
Self-assessment made me review the course contents more.	0,826		0,217
Self-assessment made me participate more in course activities.	0,743
Self-assessment helped me to recognise my errors.	0,727		−0,230
Self-assessment motivated me to engage more deeply with course tasks.	0,693
Self-assessment increased my motivation to study in the course.	0,575
I was able to apply the assessment rubric to mark my solutions.		0,876
I was able to to assess my performance with the provided assessment rubrics and model solutions.		0,851
I understood how the tasks were supposed to be solved when I compared them to the assessment rubric and model solutions.	0,309	0,588
Self-assessment increased the workload too much			0,894
Self-assessment was unpleasant and stressful			0,823
Performing self-assessment felt frustrating.			0,880
It was difficult to assess my performance if I had not mastered the tasks myself.			0,772
I felt that I could not assess my work reliably.	0,252	−0,258	0,694
There is little benefit to self-assessment compared to the workload.			0,556

SF1 = Self-assessment as supporting learning –factor.

SF2 = Self-efficacy in self-assessment –factor.

SF3 = Negative experiences of self-assessment –factor.

Correlations between PPSA3 factors and other pedagogical measures

Correlations between the PPSA3 factors and learning approaches, self-efficacy, and the three components of study-related burnout are reported in Table 5. As hypothesised, negative perceptions of both peer and self-assessment were positively associated with the unreflective learning approach. We also observe statistically significant connections between negative perceptions of both peer and self-assessment and the different components of study-related burnout, except for PF3 and exhaustion.

Table 5. Correlations between the peer and self-assessment factors in PPSA3 and the measures of learning approaches, study-related burnout, and general self-efficacy.

	Mean	SD	PF1	PF2	PF3	SF1	SF2	SF3
Self-efficacy	3.94	0.96		0.18*			0.23*
Unreflective	3.31	0.78			0.23*		−0.21*	0.27**
Deep	3.24	0.69		0.26**			0.22*
Exhaustion	3.04	0.98	0.18*					0.29**
Cynicism	2.23	1.09			0.23*		−0.20*	0.25**
Inadequacy	3.20	1.10	0.18*		0.22*			0.31**

Note: * p < 0.05; ** p < 0.01.

PF1 = Peer assessment as supporting learning –factor.

PF2 = Self-efficacy in peer assessment –factor.

PF3 = Negative experiences of peer assessment –factor.

SF1 = Self-assessment as supporting learning –factor.

SF2 = Self-efficacy in self-assessment –factor.

SF3 = Negative experiences of self-assessment –factor.

Unreflective = Unreflective approach to learning.

Deep = Deep approach to learning.

On the positive side, self-efficacies in both peer and self-assessment were linked to general self-efficacy, as expected. In addition, self-efficacy in self-assessment was positively correlated with the deep approach and negatively correlated with both the unreflective approach and the cynicism component of study-related burnout. Meanwhile, self-efficacy in peer assessment correlated positively with the deep approach but did not correlate with the unreflective approach. Almost no correlations were observed for the supporting learning factors: Only PF1 showed positive correlations with exhaustion and inadequacy. While statistically significant, all observed correlations were small.

Discussion

In Study II, a new version of the PPSA (PPSA2) was developed based on the results from Study I. Exploratory factor analysis was then used to reduce this 46-item PPSA2 into the final 30-item PPSA3. Three similar factors emerged from the peer and self-assessment perception data: The first factor had to do with peer or self-assessment as supporting learning. It combined items in the open response categories of Learning and reflection and Motivation and participation from Study I. The second factor was analogous to the self-efficacy open response category for both peer and self-assessment. The third factor subsumed all items in the three negative experience open response subcategories. For both peer and self-assessment perceptions, the emerging factors showed ${\omega }_t$ values greater than 0.82 and contained substantial loadings only from PPSA3 items that were a part of that factor, underscoring the reliability of the instrument. At the same time, the ${\omega }_t$ values were not extremely close to one, which is good as such values could indicate excessive redundancies between factor items. As expected from the open responses of Study I, a two-factor solution further collapsed the supporting learning and self-efficacy factors into a single positive experience factor, recovering the positive versus negative perception dichotomy that formed the basis of the open response categorisation in Study I.

To investigate the convergent validity of the PPSA3, we studied correlations between the identified factors and learning approaches, study-related burnout, and self-efficacy. As expected, based on the finding that self-assessment can promote self-efficacy (Nieminen, Asikainen, and Rämö 2021; Panadero, Jonsson, and Botella 2017), the self-efficacy factors in both peer and self-assessment were correlated with global self-efficacy. Our results also aligned Parpala et al. (2010), in that students with the deep approach tended to perceive the teaching and learning environment more positively as evidenced by the positive correlation between the deep approach and the self-efficacy factors for peer and self-assessment. At the same time, the unreflective approach was positively correlated with negative views of peer and self-assessment and negatively correlated with perceptions of self-efficacy in self-assessment.

Surprisingly, the self-assessment as supporting learning factor showed no correlations at all, while the corresponding peer assessment factor showed weak positive correlations with the measures of exhaustion and inadequacy. These correlations are small, so they might just be a statistical artefact caused by the relatively small sample size. However, many students praised peer and self-assessment as supporting learning while simultaneously berating them for the extra workload in their open responses in Study I. Perhaps this phenomenon is visible here, as it is the more work-intensive peer assessment that is connected to the study-related burnout variables.

Besides their connections to the unreflective approach, negative experiences of both peer and self-assessment were associated with the exhaustion, cynicism, and inadequacy facets of study-related burnout. These connections are in line with the hypotheses drawn from the literature, as students applying the unreflective approach tend to experience more study-related burnout symptoms than those applying the deep approach (Asikainen et al. 2020). The strongest correlations were observed between students’ negative experiences of self-assessment and the measures of inadequacy, underlining how negative perceptions of one’s ability to succeed in one’s studies are tied to negative perceptions of self-assessment. This stands to reason, given that the PPSA3 included items from the Self-criticism category of the PPSA2, and self-criticism in general is closely related to both depression and burnout (Hashem and Zeinoun 2020; López, Sanderman, and Schroevers 2018).

Limitations

There were several limitations in our study: First, the numbers of open responses of peer and self-assessment perceptions were relatively small. A second limitation is that we only used self-reports in this study, although this seems appropriate given our aim of exploring students’ perceptions. Third, our data consisted solely of students assessing course problems in one course at one university. Consequently, the PPSA needs to be further tested with other types tasks, in other contexts, and with larger populations.

Conclusions and practical implications

This study explored Finnish chemical engineering students’ perceptions of peer and self-assessment. In the students’ view, peer and self-assessment possess both shared and unique advantages, with different students preferring and benefitting more from one or the other activity. Negative perceptions of peer and self-assessment seem to be generally linked to students’ doubts in their abilities, unreflective studying, and lack of well-being. Meanwhile, positive experiences of assessing own or others’ work may enhance self-efficacy and well-being and facilitate the adoption of a deep approach to learning. Teachers should thus work towards creating positive assessment experiences for their students. Future studies should also carefully consider whether treating peer and self-assessment as two different but complementary assessment practices rather than a single entity makes sense in their learning context. At present, peer assessment appears somewhat more prevalent in engineering education than self-assessment based on our literature review. Therefore, teachers should consider implementing self-assessment activities together with peer assessment to facilitate learning.

We developed the PPSA questionnaire based on previous literature and our qualitative investigation of students’ perceptions of peer and self-assessment in Study I. In Study II, we showed that the PPSA measures student peer and self-assessment perceptions in higher education chemical engineering courses both reliably and validly. Next, the PPSA should be tested with larger populations and in different learning contexts. However, due to the relatively large number of items within the PPSA, we recommend that teachers add a time estimate at the beginning of the questionnaire.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Notes on contributors

Lauri J. Partanen

Dr. Lauri J. Partanen is a university lecturer in physical chemistry at Aalto University in the School of Chemical Engineering. His research interests span computational and theoretical chemistry, particularly in the fields of quantum mechanics and thermodynamics but the principal focus is on chemistry education, learning, and assessment.

Viivi Virtanen

Dr. Viivi Virtanen is a principal research scientist in the EDU-research unit at the Häme University of Applied Sciences in Finland. She also has a docentship/adjunct professorship at the University of Helsinki. Her research interest include broadly teaching and learning in higher education with specific interests in well-being, and researcher education and careers especially in STEM fields.

Henna Asikainen

Dr. Henna Asikainen is a senior university lecturer in university pedagogy at the University of Helsinki in the Centre for University Teaching and Learning and a research director in the WELLS-project. She has the title of docent in university pedagogy. Her research interests comprise student and teacher wellbeing and its relation to learning and assessment in learning.

Liisa Myyry

Dr. Liisa Myyry is a senior university lecturer in university pedagogy at the University of Helsinki in the Centre for University Teaching and Learning. She has the title of docent in social psychology. Her research interests are moral and adult development, personal values, digitalization of teaching and learning and assessment practices in higher education.

Appendix 1.

The PPSA2 translated from Finnish into English.

Table

1A. Learning and reflection Peer assessment supported my learning in the course. Peer assessment made me review the course contents more. Peer assessment helped me understand what I know and do not know. Peer assessment helped me recognise my errors. Peer assessment opened new perspectives to the course tasks and their solutions. I feel like I improved in peer assessment during the course.	1B Learning and reflection Self-assessment supported my learning in the course. Self-assessment made me review the course contents more. Self-assessment helped me understand what I know and do not know. Self-assessment helped me to recognise my errors. I feel like I improved in self-assessment during the course.
2A Motivation and participation It was good that the students participated in assessment through peer assessment. Peer assessment motivated me to engage more deeply with course tasks. Peer assessment increased my motivation to study in the course. Peer assessment made me participate more in course activities. Peer assessment brought nice variation to how courses are usually assessed	2B Motivation and participation It was good that the students participated in assessment through self-assessment. Self-assessment motivated me to engage more deeply with course tasks. Self-assessment increased my motivation to study in the course. Self-assessment made me participate more in course activities. Self-assessment brought nice variation to how courses are usually assessed.
3A Self-efficacy in peer assessment I was able to assess the performance of my peers with the provided assessment rubrics and model solutions. I understood how the tasks were supposed to be solved when I compared them to the assessment rubric and model solutions. I was able to apply the assessment rubric to mark my peers‘ solutions.	3B Self-efficacy in self-assessment I was able to assess my performance with the provided assessment rubrics and model solutions. I understood how the tasks were supposed to be solved when I compared them to the assessment rubric and model solutions. I was able to apply the assessment rubric to mark my solutions.
4A Exhaustion and workload There is little benefit to peer assessment compared to the workload. In peer assessment, the students are forced to do the teacher’s work as well. Peer assessment increased the workload too much.	4B. Exhaustion and workload There is little benefit to self-assessment compared to the workload. In self-assessment, students are forced to do the teacher’s work as well. Self-assessment increased the workload too much.
5A Cynisism I feel that many students were not able to assess their peers’ work. Peer assessment is unfair because the effort students put into it varies a lot. The criteria for peer assessment felt unjust when compared to the tasks. Peer assessment was a waste of time and felt useless.	5B. Cynisism­­­ Performing self-assessment felt frustrating. The criteria for self-assessment felt unjust when compared to the tasks. Self-assessment was a waste of time and felt useless.
6A Self-critisism I was ashamed to show my solutions to others. I felt that I could not assess others’ work reliably. It was difficult to give feedback if I had not mastered the tasks myself.	6B. Self-criticism Self-assessment was unpleasant and stressful. I felt that I could not assess my work reliably. It was difficult to assess my performance if I had not mastered the tasks myself.