The association between student–teacher relationship quality and school liking: A small-scale 1-year longitudinal study

Abstract This 1-year longitudinal study examined the association between student—teacher relationship quality and school liking in a sample of 234 students from two public schools in Sweden, who completed an online questionnaire on two separate occasions. The age range was 9–15 years in Time 1 and 10–16 years in Time 2. A path analysis showed that students who were younger, liked their school more, and had more positive, warm, and supportive relationships with their teachers were more inclined to score high in school liking one year later. In addition, younger students and students who liked their school and had better relationships with their teachers at Time 1 were inclined to have better relationships with their teachers one year later.


Introduction
Considering that students spend a notable amount of their time in schools that affects their learning and development in many ways, the degree to which they thrive in and like their school should not be underestimated. School liking is an important aspect of students' school experiences that has been associated with a range of positive student outcomes, and learning more about how teachers can affect students' school liking is therefore desirable. In this study, we examined whether good student-teacher relationships and school liking are longitudinally interconnected among students over the course of an academic year.

Student-teacher relationship quality
How teachers treat their students and establish relationships with them in school is crucial in many ways. Self-determination theory (Ryan & Deci, 2017) assumes that relatedness is one of the three human basic needs (the other two are autonomy and competence) that must be satisfied for motivation, development and well-being to be sustained. In a school context, relatedness refers to when students feel a sense of belonging and are involved in relationships where they experience themselves and others as mutually caring, responsive, and sensitive. When teachers manage to establish such relationships with their students, they contribute to satisfying students' basic needs, which in turn will increase student motivation, engagement, achievement, and well-being.
According to the extended attachment theory (Pianta, 1999;Sabol & Pianta, 2012), close, warm, and supportive student-teacher relationships provide students with emotional security and trust where teachers function as a "secure base" that allows students to explore and learn in school in a more effective way, and to develop confidence, competence, and well-being. The theory utilizes three dimensions of student-teacher relationship quality: closeness (the degree of warmth, support, trust, and positive affect), conflict (the degree of conflicts, poor and harsh relationships, and unpleasant and negative affect), and dependency (the degree of clinginess, overreliance, and possessiveness). Close or positive student-teacher relationships and conflictual or negative student-teacher relationships are much more studied in the literature than dependent studentteacher relationships, which are largely unexplored (Roorda et al., 2021), partly because the validity of measures of dependency has been questioned (Lei et al., 2016).
While it is common to make a distinction between positive (close) student-teacher relationship quality and negative (conflictual) student-teacher relationship quality in the literature (e.g., Roorda et al., 2017), student-teacher relationship quality can also be operationalized as a single dimension with closeness (positive relationship) and conflict (negative relationship) as two oppositive ends of a continuum (e.g., Goetz et al., 2021). In the current study, we operationalized student-teacher relationship quality in line with the latter. Thus, higher quality means that the student-teacher relationship is characterized by being more positive, warm, caring, supportive and close, and less negative, conflicting, harsh, and poor.
Previous research supports self-determination theory and extended attachment theory in terms of student-teacher relationship quality and student outcomes. Students who experience warm, caring, supportive and respectful student-teacher relationships are more likely to feel higher levels of well-being (García-Moya, 2020; Markkanen et al., 2019;Sarkova et al., 2014) and more positive emotions (enjoyment and pride) in the classroom (Goetz et al., 2021), and to be less engaged in disruptive behaviors (Quin, 2017) and other forms of behavioral problems (Lei et al., 2016). They tend to feel fewer negative emotions (anxiety, anger, boredom, and shame) in the classroom (Goetz et al., 2021), and be more engaged in their schoolwork and learning and more successful regarding academic achievement (Quin, 2017;Roorda et al., 2011Roorda et al., , 2017)-an effect which is even stronger for students from lower socioeconomic backgrounds (Roorda et al., 2011(Roorda et al., , 2017. In addition, studies have shown that positive, warm, and supportive relationships decrease the risk of school bullying and peer victimization, while negative, poor, and harsher relationships increase this risk (Banzon-Librojo et al., 2017;Chen et al., 2021;Dietrich & Cohen, 2021;Espelage et al., 2015;Longobardi et al., 2022;Thornberg et al., 2018Thornberg et al., , 2022. Furthermore, qualitative and mixed-methods studies on students' perspectives on their school experiences have shown that students prefer teachers who are kind, friendly, fair, caring, and trustworthy; treat students as individuals; show them respect; and build positive relationships with them (Egeberg & McConney, 2017;Krane et al., 2017;Power et al., 2018;Raufelder et al., 2016;Strikwerda-Brown et al., 2008;Thornberg et al., 2020). Altogether, our review of the empirical research suggests that both school liking and student-teacher relationship quality are important for students in their schooling. In the current study, we wanted to investigate whether school liking and student-teacher relationship quality were interrelated over time.

The link between student-teacher relationship quality and school liking
Research has shown that students who experience better student-teacher relationships tend to have a stronger sense of belongingness to their school (Allen et al., 2018). By contrast, not feeling supported, such as feeling like teachers do not like them, has been reported by students as aspects in school that prevent them from feeling part of the school (Porter et al., 2021). In a qualitative interview study of boys who had a record of school problems and disruptive behavior, the most common reasons they gave for why they began to dislike school in their early years were difficulties with schoolwork and negative student-teacher relationships (Graham et al., 2016). Very few quantitative studies have investigated and analyzed the link statistically, but according to their findings, warm and supportive children-teacher relationships are associated with school liking (Bear et al., 2011;Graham et al., 2022;Hallinan, 2008). In other words, students who have good relationships with their teachers and experience them as kind, caring, warm, respectful, supportive, and fair, tend to like their school.
For example, Hallinan (2008) found that students who feel that their teachers respect, support, and care about them were more inclined to like school. Hallinan also showed that studentteacher relationship quality predicted school liking over time, but whether school liking predicted student-teacher relationship quality was not tested. In comparison with students who disliked school, Graham et al. (2022) found that students who liked school were more prone to perceive that their teachers really cared about them, told them when they did a good job, noticed when they were not present, listened to them when they had something to say, believed that they will be a success, and made time for them. Graham et al. (2022) argue that "students who dislike school may be more likely to experiences conflict and have poorer quality relationships with teachers, feel less connected to school, and be subject to higher rates of exclusionary discipline, which is precisely the opposite of what these students need to stay in and succeed at school" (p. 2). By contrast, students who feel that their teachers are fair, give praise, and care about them as students, tend to like their school; and students who like their school tend to feel safer, have better peer relationships, and be more engaged in schoolwork and academic learning.

Current study
With reference to the relevant literature, we wanted to further investigate the longitudinal association between student-teacher relationship quality and school liking among students at two public schools in Sweden. Quantitative studies testing this link are scarce, and only one study, as far as we know, has used a longitudinal design to investigate the association between the two variables over time, and this study only examined whether student-teacher relationship quality predicted school liking. It did not investigate whether school liking predicted student-teacher relationship quality. Therefore, the aim of the current study was to examine the longitudinal association between student-teacher relationship quality and school liking. Three research questions were raised: (1) Does student-teacher relationship quality predict school liking?
(2) Does school liking predict student-teacher relationship quality?
(3) Is the longitudinal link between student-teacher relationship quality and school liking unidirectional (i.e., only one of them predicts the other) or bidirectional (i.e., both predict the other)?
Following Hallinan's (2008) study, we expected that student-teacher relationship quality would predict school liking one year later, while the second and third research questions were examined in an exploratory manner to address the gap in the literature. Gender and age were included in the study as control variables. Previous studies report that girls tend to like school more than boys (Coelho et al., 2020;Ding & Hall, 2007;Erath et al., 2008;Riglin et al., 2013;Rönkä et al., 2017; for an exception that shows no gender difference, see Boulton et al., 2011). In addition, students tend to like school less as they get older (Arens & Niepel, 2019;Coelho et al., 2020;Ding & Hall, 2007).

Methods
We adopted a short-term longitudinal design with two measurement times (spring 2018 and spring 2019).

Participants and procedure
Participants consisted of 234 students (113 girls and 121 boys) from two public schools in southeast Sweden who completed an online questionnaire twice in their ordinary classroom or in a computer lab in their school. The first school was a K-9 school located in a small village where 61% of students had highly-educated parents, and 10%-12% had a foreign background (i.e., either they were born in another country, or both their parents were born in another country). The second school was a K-6 school located in a medium-sized city, where 86% of the students had highlyeducated parents, and only 5% had a foreign background. At the end of the spring semester in 2018, the students in the sample were in grades 3-8 (age range = 9-15, mean age = 11.3, SD = 1.7) when they filled in the questionnaire, and they completed the questionnaire again at the end of the spring semester in 2019, when they were in grades 4-9 (age range = 10-16, age mean = 12.3, SD = 1.7). The study received ethical approval from the Regional Ethical Review Board in Linköping, and we obtained parental consent and student assent for all students who participated in the study.

Student-teacher relationship quality
We used a student-teacher relationship quality scale (Sjögren et al., 2021) to assess how students perceive the quality of their individual relationships with their teachers, because the scale has been developed in Sweden and been found to be reliable and valid. Sjögren et al. (2021) demonstrated a high reliability (Cronbach's α = .92). In addition, Thornberg et al. (2020) found that the scale was concurrently correlated with student engagement at two time points (r = .53, p < .001; r = .51, p < .001) and predictive of student engagement one year later (r = .46, p < .001); even when controlling for gender, age, and student engagement at the first time point (β = 0.15, p < .01). This can be interpreted in terms of criterion validity, both as concurrent validity and predictive validity. Using this scale, we asked the students to estimate nine items (e.g., "My teachers care about me", "My teachers listen to me when I have something to say", and "I feel that I can trust my teachers"), from which we calculated a mean value. Some items were inverted (e.g., "My teachers yell at me" and "My teachers dislike me") before issuing the student-teacher relationship quality variable. Students rated each item on a five-point scale (4= always, 3= often, 2= sometimes, 1= seldom, 0= never). Thus, a higher score means that the students perceived their teachers as showing a higher degree of support, caring, responsiveness, and respect toward them, and a less conflict, dislike, hostility, and anger toward them. Cronbach's α for the scale was .88 at Time 1 and .92 at Time 2, which suggest that the reliability (internal consistency) of the scale was good to excellent.

School liking
We developed a short scale to measure school liking for this study. We asked the students to estimate three items ("I like my school", "I hate school" [reverse], "The school is important to me"). Students rated each item on a five-point scale (4= always, 3= often, 2= sometimes, 1= seldom, 0= never). While the widespread "School Liking and Avoidance Questionnaire" (SLAQ; Ladd & Price, 1987;Ladd et al., 1996) has been developed for younger students, and other measurements of school liking include only one item (e.g., "Do you like school?"; Graham et al., 2016) or are incorporated into a broader scale (e.g., Bear et al., 2011), the current scale was constructed to be short (but longer than 1 item to check for internal reliability) and appropriate for students at both elementary and secondary schools in Sweden. Cronbach's α for the scale was .77 at Time 1 and .80 at Time 2, which suggest that the reliability (internal consistency) of the scale was acceptable to good.

Descriptive statistics
Descriptive statistics for the study variables student-teacher relationship quality and school liking at Times 1 and 2, are shown in Table 1, and correlations are shown in Table 2. All study variables were positively associated with each other. Student-teacher relationship quality was positively correlated with school liking within each timepoint. Student-teacher relationship quality at Time 1 was also positively correlated with school liking at Time 2, and school liking at Time 1 was positively correlated with student-teacher relationship quality at Time 2. Furthermore, the correlations showed that those who liked their school at Time 1 tended to  like the school one year later (and vice versa), and those who reported positive, warm, and supportive student-teacher relationships at Time 1 tended to do the same one year later (and vice versa).

Path analysis
Because we wanted to examine the effects of student teacher relationship quality and school liking, at Time 1, on student teacher relationship quality and school liking, at Time 2, we estimated a path model using the ML estimator in the Lavaan package version 0.6-11 in R. Prior to estimation, we transformed the dependent variable student teacher relationship quality at Time 2 by reflecting it and then log transforming it, which resulted in a skewness of 0.290 and a kurtosis of −0.782. We transformed school liking at Time 2 by reflecting it and then taking the square root, resulting in a skewness of 0.236 and a kurtosis of −0.666. In order to ease interpretations of the path coefficients, we reflected the variables again, so that high values on the original variables would correspond to high values on the transformed variables. There was no need to transform the study variables at time 1, because they served as independent variables in the analysis. Figure 1 shows the standardized coefficients from the estimation of the path analysis. The model explains 45% of the variance in student teacher relationship quality at Time 2, and 34% of the variance in school liking at Time 2, and the residual correlation between these variables is only 0.02. As shown in the figure, gender did not predict school liking nor student teacher relationship quality. Younger students however liked their schools more than older students, and they reported higher student-teacher relationship quality scores than older students. Our results also revealed that student-teacher relationship quality at Time 1 predicted both and student-teacher relationship quality and school liking at Time 2, and school liking at Time 1 predicted student-teacher relationship quality and school liking at Time 2. In other words, students who had more positive, warm, and supportive relationships with their teachers were more inclined to score high in school liking and on student teacher relationship quality one year later. Similarly, students who liked their school more tended to score high in school liking and on student teacher relationship quality one year later.

Discussion
Although a few studies have analyzed the link between student-teacher relationship quality and school liking (Bear et al., 2011;Graham et al., 2022;Hallinan, 2008), the possible longitudinal association between these variables is understudied. Only one study, as far as we know, has adopted a longitudinal design and its findings showed that positive and caring student-teacher relationships predicted students' school liking two years later (Hallinan, 2008). However, it did not examine whether school liking, at the same time, predicted student-teacher relationship quality. To contribute to filling these gaps, in the present study student-teacher relationship quality and school liking were both considered to be potential longitudinal predictors of each other.
In this 1-year longitudinal study including students from two Swedish schools, we found that student-teacher relationship quality at Time 1 predicted school liking one year later (Time 2) when controlling for school liking at Time 1. At the same time, our results showed that school liking at Time 1 predicted student-teacher relationship quality one year later (Time 2) when controlling for student-teacher relationship quality at Time 1. Thus, both study variables were interrelated over time. After one year, they predicted not only themselves but also each other. In other words, students who experienced warm and supportive relationships with their teachers liked their school one year later, and students who liked their school experienced warm and supportive studentteacher relationships one year later. This bidirectional, longitudinal association between studentteacher relationship quality and school liking found in our study highlights the importance of teachers in building positive relationships with all students to increase the probability of having students who like being at school.
The longitudinal association between student-teacher relationship quality and school liking found in the present study supports extended attachment theory (Pianta, 1999;Sabol & Pianta, 2012), and self-determination theory, and its assumption of relatedness as a human basic need applied to the school context (Ryan & Deci, 2017). The findings further contribute to empirical literature on student-teacher relationship quality and student outcomes (e.g., Lei et al., 2016;Quin, 2017;Roorda et al., 2011Roorda et al., , 2017 by demonstrating how student-teacher relationship quality and school liking are predicting each other over time in a sample of Swedish students. Unlike several studies (e.g., Coelho et al., 2020;Lietaert et al., 2015) but in line with a few others (e.g., Sjögren et al., 2021), gender was not significantly associated with student-teacher relationship quality. Neither was gender significantly linked with school liking, which has been found to be the case in several studies (e.g., Coelho et al., 2020;Rönkä et al., 2017) but not in all (e.g., Boulton et al., 2011). Possible explanations for these inconsistent findings may involve cultural, sampling, or methodological differences across studies. In addition, the small and very local sample in our study might have affected the statistical analyses, rendering them less able to detect gender differences.
In accordance with previous research (e.g., Arens & Niepel, 2019;Coelho et al., 2020;Jerome et al., 2009), we found that both student-teacher relationship quality and school liking declined with age. One possible explanation might be the school system. It may be easier to establish and maintain positive, close, warm, and supportive student-teacher relationships in primary school where students have only a few, close teachers, as compared to secondary school where students have several specialized school subject teachers and classrooms. Better student-teacher relationships in primary school might then influence younger students to like school more, at least compared to older secondary school students.

Study limitations
Some of the limitations of the current study are worth mentioning. One limitation is the use of selfreported data, which are susceptible to careless marking, intentionally exaggerated responding, social desirability, perception and recall biases. Another limitation is that our longitudinal analysis only covered one year and two measurement time points. Future research should examine the longitudinal association between school liking and student-teacher relationship quality more in depth by using cross-lagged models over longer periods of time and with several time points. A further limitation in our study is that we did not investigate contextual factors-such as socioeconomic and ethnic background and neighborhood compositions-beyond the studentteacher relationship dyads.
Finally, a note of caution needs to be sounded regarding generalizability. Our small and nonprobabilistic sample of students from two public schools within a particular area in Sweden may or may not be like the population of students with whom readers primarily work or are interested in. Nevertheless, and in contrast to naïve realism but in line with post-positivism (Phillips & Burbules, 2000;Popper, 1959) and pragmatism (Biesta & Burbules, 2003;Dewey, 1929;Peirce, 1960), we do not claim that our variables and study findings are exact representations of reality but partial, fallible, and provisional estimations and approximations.

Practical implications
These limitations aside, the present results showed that students' relationships with their teachers play an important role in school liking. Thus, teachers' social-emotional competence must be considered as essential for promoting a positive social context for their students' learning and wellbeing in school (cf., Hen et al., 2016). Our study contributes to the literature by proposing that student-teacher relationship quality and school liking seem to mutually affect each other over time. Enriched school liking experiences, and positive, caring, warm and supportive studentteacher relationships can, therefore, be interpreted as parts of effective teaching. School policy and practice should include goals and strategies designed to increase student-teacher relationship quality and students' school liking.
Policymakers and practitioners need to understand that student-teacher relationship quality and students' school liking are intercorrelated. Our study contributes to this understanding by demonstrating that these two aspects of students' school experiences predict each other over time; at least in our small sample of Swedish students. Thus, how schools promote good studentteacher relationships and school liking-which are associated with each other and with students' well-being and educational achievement-must be considered and addressed in national and local efforts to realize effective teaching.