Need supportive collaborative learning: are teachers necessary or do students support each other’s basic psychological needs?

ABSTRACT

Some groups work together more successfully than other groups. These differences could depend on the extent to which students’ basic psychological needs for autonomy, competence, and relatedness are supported during collaborative learning. The aim of the present study was to examine associations between need satisfaction and performance, and to examine whether students experience more need satisfaction and performed better if their teacher was present to offer support during collaborative learning. Students (N = 466) were randomly assigned to groups that were led by a teacher (N = 20 groups) or worked independently without teacher support (N = 129 groups). Results revealed that satisfaction of the needs for autonomy and competence was positively associated with performance, whereas relatedness satisfaction was negatively related to performance. Teacher-led groups and groups without a teacher present did not differ in need satisfaction or performance, indicating students are also able to offer each other support.

KEYWORDS:

• Collaborative learning
• self-determination theory
• need supportive teaching

Many studies on collaborative learning (CL) have confirmed the positive impact of student collaboration in the classroom in terms of greater academic achievement (e.g., Kyndt et al. 2014). However, some groups of students work together more successfully and obtain better achievement outcomes than other groups. These differences could depend on the extent to which students’ basic psychological needs are supported during CL. That is, research from self-determination theory has shown that satisfaction of students’ basic psychological needs for autonomy, competence, and relatedness, promotes student motivation and engagement (Deci and Ryan 2000; Vansteenkiste, Ryan, and Soenens 2020). Given the evidence on the benefits of need support in whole classroom settings (e.g., Stroet, Opdenakker, and Minnaert 2013), it can be expected that need support will also be highly beneficial for student outcomes in the context of CL. The first aim of this study is therefore to examine whether need satisfaction during CL is related to student performance.

As studies on need support during CL are scarce, it is unclear however, whether need support needs to be offered by the teacher or whether students can also support each other’s basic psychological needs. Previous research shows the important role teachers play in promoting students’ satisfaction of basic psychological needs (for a review, see Stroet, Opdenakker, and Minnaert 2013). On the other hand, group members also function as an important social context during CL and it could be that through their interactions, students’ needs can also be supported by their group members (Hänze and Berger 2007). The second aim of this study is therefore to examine whether the presence of a teacher during CL enhances students’ need satisfaction and achievement or whether students experience similar levels of need satisfaction without a teacher present.

Need satisfaction during collaborative settings

The self-determination theory (Deci and Ryan 1985, 2000, 2004) is a theoretical framework on social development that posits that all people are inclined towards psychological growth. The social context in which people function can promote or hinder this growth, depending on whether it supports individual’s basic psychological needs. SDT states that each individual has three core basic psychological needs: the need for autonomy, competence, and relatedness. The need for autonomy refers to people’s innate desire to experience volition and psychological freedom when engaging in activities (Deci and Ryan 2004). The need for competence refers to the need to feel capable and in control and being able to develop one’s competences further (Deci and Ryan 2004). The need for relatedness refers to the desire to belong, and to develop and maintain meaningful and supportive relations with others (Deci and Ryan 2004). These needs apply to all areas of human functioning, including students’ functioning in a classroom context. Indeed, abundant research has indicated that need satisfaction is positively related to student engagement (for reviews, see Stroet, Opdenakker, and Minnaert 2013; Vansteenkiste, Ryan, and Soenens 2020).

Given the importance of psychological need satisfaction in all areas of human functioning, including education and collaboration at work (Deci and Ryan 2004; Vansteenkiste, Ryan, and Soenens 2020), our hypothesis is that this importance also applies to the context of collaborative learning (CL). In this paper, we regard CL as the activity in which students jointly work on a task, where students have the opportunity of mutual influence and equal participation (O’Donnell and Hmelo-Silver 2013). Meta-analyses on CL show the positive impact of student collaboration in the classroom in terms of greater academic achievement (Kyndt et al. 2014; Rohrbeck et al. 2003; Roseth, Johnson, and Johnson 2008). Research points out particular types of interaction between students during CL that are beneficial for the acquisition of knowledge and skills, namely explaining, elaborating, reflecting on and asking questions about the domain content (Asterhan and Schwarz 2016; Weinberger and Fischer 2006). For these beneficial collaborative processes to occur, students within a small group need to get to know each other and feel a sense of community and trust (Van den Bossche et al. 2006). Without it, there is a risk of students feeling anxious and defensive and unwilling to engage in the learning processes described above (Kreijns, Kirschner, and Jochems 2003). Also, each collaborating group needs to coordinate or regulate their activities (Järvelä et al. 2016). That is, the group needs to make sure their strategies for solving the task and for collaborating are adequate.

From the description of the interactions and conditions that make CL effective, the relation to psychological need satisfaction may be drawn. The sense of community and trust could enhance satisfaction of the need relatedness between students. Similarly, exchange of strategies for solving the task could foster feelings of competence, and regulation of the collaboration process could ensure that all group members experience autonomy. Thus, and aligning with prior research in a variety of settings, including work and education showing positive relations between need satisfaction and performance (Vansteenkiste, Ryan, and Soenens 2020), we hypothesise that increased need satisfaction also increases performance in the context of CL. Minnaert et al. (2011) indeed found that vocational students who experienced satisfaction of all three needs during a collaborative project experienced higher levels of situational interest. Likewise, a study by Liu et al. (2009) among 8th-grade students on collaborative project work indicated that high need satisfaction during group work was associated with higher levels of enjoyment, value, and perceived learning skills. Likewise, although these studies offer preliminary support for the idea that need satisfaction during CL is associated with higher performance, these studies included measures of interest and self-reported learning skills, rather than actual performance measures as outcomes. Moreover, these studies did not differentiate between group performance and individual learning gains as outcomes.

The role of the teacher and peers for students’ experienced need satisfaction

A subsequent question is how need satisfaction can be realised during CL. During CL, a large part of the interactions take place between students. Thus, in terms of psychological need satisfaction during CL, the social context of the small group a student is part of, can be considered an important source of need satisfaction. When the teacher is present during CL, he or she can also be an important source of need satisfaction for students. Many studies have examined how teachers, as important facilitators of learning in the context of whole class instruction, can support students’ basic needs (e.g. Assor and Kaplan 2001; Belmont et al. 1992; Reeve 2009; Jang, Reeve, and Deci 2010; Stroet, Opdenakker, and Minnaert 2013). Teachers can support students’ need for autonomy by aligning with students’ goals and interests, offering meaningful choices (Assor and Kaplan 2001; Reeve et al. 2004), fostering relevance (Assor and Kaplan 2001; Belmont et al. 1992), and being respectful towards students’ wishes and concerns (Assor and Kaplan 2001). Teachers can support students’ need for competence by providing students with structure (Jang, Reeve, and Deci 2010; Skinner and Belmont 1993). This includes providing clarity, guidance, encouragement, and informational feedback (Stroet, Opdenakker, and Minnaert 2013). Finally, students’ need for relatedness can be fostered when teachers show involvement by expressing affection, being attuned to the student, and by being dependable (Stroet, Opdenakker, and Minnaert 2013).

In the context of CL, the teacher’s choice of support must be specified to the group, and what the teacher says and does influences the interaction processes within the small group (Van Leeuwen and Janssen 2019). When offering support to collaborating students, the teacher acts as a model for how students can interact with each other within their group (e.g., Webb, Nemer, & Ing, 2006). Thereby, teachers can model how to put forward arguments or how to constructively respond to group members (e.g., Gillies 2011). Also, teachers do well by supporting students how to select appropriate strategies for solving the task or for collaborating (e.g., Oortwijn et al. 2008) without offering the answers or solutions to the task (e.g., Almasi, O’Flahavan, and Arya 2001). Compared to whole-class teaching, effective teacher behaviours during CL appear to be focused mostly on offering autonomy so that students are able to manage collaborative processes on their own (Van Leeuwen and Janssen 2019; Webb 2009), and on stimulating relatedness by acting as a model to show students how to be respectful and open to others to stimulate each group member’s feeling of autonomy (Gillies 2011). Enhancing the feeling of autonomy may thus be a prominent role for the teacher during CL. On the other hand, during collaboration students depend on each other for social support and for a feeling of trust and psychological safety (Jones and Issroff 2005; Van den Bossche et al. 2006). This means that compared to whole-class teaching, the teacher’s role in satisfying the need for relatedness may be smaller and to a large extent be taken up by the interaction between group members.

Research concerning the relation between teacher support and students’ need satisfaction during collaboration is scarce. A notable exception is a quasi-experimental study by Hänze and Berger (2007), in which students either received direct instruction or worked collaboratively on an physics assignment. Students in the CL condition experienced more need satisfaction, yet their performance outcomes were lower. However, in the cooperative condition, no teachers were present. It, therefore, remains unclear whether the differences in need satisfaction between both conditions were actually due to the (collaborative) nature of the task, or due to the presence versus absence of the teacher. As such, it is unclear whether the positive effects of CL on need satisfaction may have been even more beneficial and whether performance outcomes may have been better if the cooperative groups would have also received teacher guidance. Thus far, to our knowledge, no studies have directly compared similar classroom situations with and without the teacher present and compared whether the presence of the teacher during CL affects need satisfaction and performance outcomes. The question, therefore, remains what the influence of teacher support during CL is on students’ experienced need satisfaction and subsequent performance, and whether students are able to support each other sufficiently without a teacher present.

The present study

To summarise, research from the tradition of SDT has mostly focused on the role of need satisfaction and need supportive teaching in whole classroom settings, whereas only few studies focused on the potential benefits of need satisfaction on performance during collaboration. Moreover, there is a lack of research concerning the role of the teacher and peers concerning students’ experienced need satisfaction during collaboration. Therefore, this study will focus on the role of experienced need satisfaction for the outcomes of CL, as well as on the effect of teacher presence on experienced need satisfaction and student group and individual performance. As such, this study addresses the following questions:

1. Is students’ experienced need satisfaction during CL, in terms of autonomy, competence and relatedness, associated with students’ performance?

2. Is the presence of a teacher during CL associated with students’ experienced need satisfaction and their performance?

Based on studies about the role of SDT and need supportive teaching in whole classroom settings, our hypotheses for these two questions are as follows:

H1. In the context of CL, students’ experienced need satisfaction is associated with their performance.

H2. Students in groups guided by a teacher will experience a higher degree of need satisfaction and will show higher performance than students in groups without teacher guidance.

A distinction will be made between students’ performance as a group and their individual performance, the latter referring to their individual learning gains.

Method

Design

The study had an experimental design, in which students collaborated in small groups on a biology assignment either with or without a teacher present. Students filled out several questionnaires concerning background variables and their experiences during collaboration, and made an individual performance test afterwards.

Participants

A total of 466 students (52.45% girls) in year 2 of secondary school (cf. grade 8 in USA) participated. Their mean age was 13.66 years (SD = 0.56). They were from 20 classes from 12 different secondary schools. From grade 7 onwards, students in the Netherlands are tracked into three different levels of secondary education. The distribution of students in our sample was as follows: 20.8% students were in vocational track (“vmbo”), 72.5% were in the general secondary education track (“havo”), and 6.7% were in the pre-university track (“vwo”). One teacher per class participated (N = 20), eight teachers were male. The students were taught by these teachers for 2–4 hours per week for at least 8 months at the time of the study. The teachers’ mean age was 41.6 (SD = 11.0). Their mean number of years of teaching experience was 11.6 (SD = 7.4).

Experimental condition and procedure

Students were randomly divided into small groups of 3–4 students (N = 149 groups). In each class, the teacher provided guidance to one randomly selected group while they were collaborating. Teachers were not informed about the purpose of the study. They received no instructions on how to provide guidance. They were only told that they could join this group and could not attend to the other groups. The other groups worked on the assignment independently.

The data for this study were collected by university students in the context of a course on study designs and data analyses. They were very carefully instructed about the procedure and the materials. The procedure of the study during data collection was as follows.

Table

	1	2	3	4
Students	Read information about heredity individually (12 min)	Collaborate on group assignment (12 min)	Questionnaire on experienced SDT and background characteristics (5 min)	Individual post-test (12 min)
Teacher	Questionnaire on teachers’ background characteristics	Provides guidance to one of the groups in the classroom	-	-

Task materials and performance

The study was conducted in the context of a biology assignment concerning the topic of heredity, and all materials (the individual information, the group assignment, and the individual post-test) were derived from an earlier study (Kirschner, Paas, and Kirschner 2009). In the group assignment, a number of statements about a family of animals were given, and the groups were asked to determine the possible genotypes and phenotypes of a particular family member. The individual post-test was similar to the group assignment (determining possible genotypes and phenotypes of a particular family member), and was used to determine whether individual learning had occurred.

Group performance was measured by the score on the group assignment, with a possible range of 0–6 points. Individual performance was measured by the score on the individual post-test, with a possible range of 0–20 points.

Experienced need satisfaction

Immediately after the collaborative assignment, students filled out a questionnaire (Broeck et al. 2010) on experienced need satisfaction during group work. The original scale is aimed at collaboration in work settings. Minor adaptations were made so that the scale also applied to CL in school (e.g. “I feel free to express my ideas and opinions in this job” was adapted to “I felt free to express my ideas and opinions when working on this task”). A principal components analysis confirmed the three-factor structure of the questionnaire. Each subscale consisted of six items. An example item of autonomy satisfaction was “I felt I had to follow other group members’ commands.” (reverse coded). The reliability of this subscale was α = 0.82. An example item of competence satisfaction was “I felt competent while working on this task.” The reliability of this subscale was α = 0.84. An example item of relatedness satisfaction was “I felt part of the group while working on this task.” The reliability of the subscale competence was α = 0.75.

Data analyses

The data had a nested structure: students nested in groups and groups nested in classes. To take the hierarchical structure of the data into account, we performed multilevel regression analyses in MPlus (Version 7; Muthén and Muthén 2017) to examine association between need satisfaction and performance, as well as the effect of teacher presence on need satisfaction and performance. Prior to these analyses, the distribution of variance across the individual, group, and class level was examined. For the analyses with group performance as the dependent variable, we analysed the data at the group level, and controlled for the classroom level (type = complex). For the analyses with need support and individual performance as the dependent variables, we estimated a two-level model (group level and individual level) and controlled for the classroom level (type = complex). Because students’ performance and their need satisfaction could be affected by students’ gender, educational track, their familiarity with the topic, the degree to which they had experience with CL, their interest in biology, and their average grade in biology, we included these variables as covariates in the analyses.

Results

Table 1 reports the descriptive statistics of the variables in this study. The means of the need satisfaction dimensions are all above the theoretical mean of the scale (3.00), suggesting that students experienced a relatively high level of need satisfaction. Furthermore, the distributions of the variance components of the variables were examined prior to further analyses. Table 1 reports the intraclass correlations as an indication of the proportion of variance situated at each level. For the group assignment, the variance could only be distributed across two levels: the group and class level. Almost half of the variance of the group assignment was situated at the class level (ICC_class =.44). For autonomy, competence, relatedness, and individual performance, the variance could be distributed at three levels: the class, group, and individual level. The intraclass correlations (ICCs) at the class and group level are reported. The remaining variance is situated at the individual level. Table 1 shows that for these variables, a substantial proportion of variance is situated at the group level (ICC_group = .14-.37). For autonomy and relatedness, the class level variance is low (ICC_class = .03 and .01), while it is more substantial for competence and performance (ICC_class = .20 and .26). In all, for each of these variables, these results indicate substantial variation at the group and individual level, allowing for the subsequent analyses focusing on these two levels, while correcting for the class level.

Table 1. Descriptive statistics and intraclass correlations.

	M	SD	Min	Max	ICCclass	ICCgroup
Autonomy	4.15	0.69	1.67	5.00	.03	.37
Competence	3.31	0.79	1.33	5.00	.19	.14
Relatedness	4.10	0.66	1.50	5.00	.01	.34
Individual performance	9.66	5.55	0.00	20.00	.25	.16
Group performance	3.41	2.04	0.00	6.00	.44	.56

Research question 1: need satisfaction and individual and group performance

Table 2 shows the results of the analyses on the associations between need satisfaction and group performance. These analyses were conducted at the group level while controlling for the classroom level. Model 1 only includes the control variables. These are students’ gender, educational track, their familiarity with the topic, experience with CL, their interest in biology, and their average grade in biology. In model 2, the three aspects of need satisfaction were added. The results of model 2 indicate that competence satisfaction is a significant positive predictor of group performance (b = 0.67, p < 0.001). That is, if group members felt more competent, they performed better on the group assignment. The associations between autonomy and relatedness satisfaction with group performance both just failed to reach significance (b = 0.34, p = 0.056; b = −0.29, p = 0.093, respectively). Need satisfaction explained 6% of the variance in group performance in addition to the control variables.

Table 2. Regression analyses on experienced need satisfaction as a predictor of group performance.

	Model 1			Model 2
	B		SE B	B		SE B
Control variables
General track (vs vocational track)	−2.19	***	0.30	−1.83	***	0.31
Pre-university track (vs vocational track)	−0.92		0.90	−0.65		0.77
Familiarity with genetics	1.11	***	0.30	0.86	**	0.27
Experience with collaboration	0.41		0.25	0.51	*	0.23
Interest in biology	−0.10		0.09	−0.18	*	0.09
Biology grade	0.34	**	0.11	0.18		0.11
Gender	0.00		0.15	0.07		0.15
Need satisfaction
Autonomy				0.34	^Ϯ	0.18
Competence				0.67	***	0.15
Relatedness				−0.29	^Ϯ	0.17
Variance
R^2	0.37			0.43

Unstandardised regression coefficients are reported.

^Ϯp <.10 * p <.05 ** p <. 01 *** p <.001

The results regarding the relation between need satisfaction and individual performance are reported in Table 3. These analyses included two levels, the group level and individual level, while controlling for the classroom level. The variables at the group level were aggregated per group. Control variables were students’ gender, educational track, their familiarity with the topic, the degree to which they had experience with CL, their interest in biology, and their average grade in biology. After adding the three dimensions of need satisfaction, it was found that at the group (between) level, all three dimensions of need satisfaction were associated with individual performance. That is, students performed better on the individual tasks if they had collaborated in groups which on average experienced more autonomy (b = 2.10, p = 0.008) and competence (b = 2.60, p < 0.001), and less relatedness (b = −1.96, p = 0.010). Furthermore, results at the within level indicated that satisfaction of the needs for autonomy and relatedness did not predict differences in individual performance within the group (b = 0.87, p = 0.151; b = −1.23, p = 0.059, respectively). The association between competence satisfaction and individual performance was significant at the within level (b = 0.94, p = 0.049). This indicates that students who felt more competent, performed better on the individual assignment than their group members.

Table 3. Multilevel regression analyses on experienced need satisfaction as a predictor of individual performance.

	Model 1			Model 2
	B		SE B	B		SE B
Within groups level
Control variables
General track (vs vocational track)	−4.41	***	0.73	−2.70	***	0.76
Pre-university track (vs vocational track)	−2.14	*	0.92	−0.85		1.07
Familiarity with genetics	2.76	***	0.64	1.78	**	0.65
Experience with collaboration	−2.27	***	0.65	−2.02	**	0.60
Interest in biology	0.46	^Ϯ	0.25	0.32		0.23
Biology grade	0.64	*	0.28	0.34		0.29
Gender	0.18		0.51	0.32		0.51
Need satisfaction
Autonomy				0.87		0.61
Competence				0.94	*	0.47
Relatedness				−1.13	^Ϯ	0.60
Between groups level
Autonomy				2.10	**	0.79
Competence				2.60	***	0.66
Relatedness				−1.96	*	0.76
Variance components
R^2 students	0.29			0.18
R^2 class	-			0.49

Unstandardised regression coefficients are reported.

^Ϯp <.10 * p <.05 ** p <. 01 *** p <.001

Research question 2: effects of teacher presence on need satisfaction, individual and group performance

Table 4 presents descriptive statistics for experienced need satisfaction separately for students working in groups with the teacher present (left), and students working in groups without the teacher present (right). Table 5 presents the results for the effect of teacher presence on need satisfaction, taking into account the same covariates as used to answer research question 1. The results indicate that the presence of the teacher did not affect satisfaction of students’ need for autonomy (b = −0.04, p = 0.707), competence (b = 0.20, p = 0.171), nor relatedness (b = 0.01, p = 0.950). Moreover, the results in Table 6 show that presence of the teacher also did not affect group performance (b = 0.02, p = 0.955), nor individual performance (b = 0.24, p = 0.819). Hence, the presence of a teacher did not affect need satisfaction or performance.

Table 4. Descriptive statistics for students working in groups with the teacher present and students working in groups without the teacher present.

	Teacher present (N = 52)				Teacher not present (N = 379)
	M	SD	Min	Max	M	SD	Min	Max
Autonomy	4.05	0.53	2.83	5.00	4.16	0.71	1.67	5.00
Competence	3.40	0.73	2.00	4.83	3.30	0.79	1.33	5.00
Relatedness	4.05	0.52	3.17	5.00	4.11	0.68	1.50	5.00
Individual performance	10.14	5.07	0.00	20.00	9.68	5.61	0.00	30.00
Group performance	3.57	1.75	0.00	6.00	3.39	2.08	0.00	6.00

Table 5. Multilevel regression analyses on teacher presence as a predictor of students’ experienced need satisfaction.

	Autonomy			Competence			Relatedness
	B		SE B	B		SE B	B		SE B
Within groups level
Control variables
General track (vs vocational track)	0.00		0.01	0.06	**	0.02	0.01		0.01
Pre-university track (vs vocational track)	−0.01		0.01	0.03		0.03	0.01		0.01
Familiarity with genetics	0.00		0.01	−0.00		0.01	0.00		0.01
Experience with collaboration	0.01		0.01	−0.03	*	0.02	−0.00		0.01
Interest in biology	0.01		0.02	0.06	*	0.03	0.00		0.02
Biology grade	−0.02		0.03	0.10	**	0.04	0.01		0.03
Gender	−0.03		0.05	−0.08		0.06	0.02		0.05
Between groups level
Teacher present	−0.04		0.11	0.20		0.14	0.01		0.11
Variance
R^2 within	0.00			0.06			0.00
R^2 between	0.00			0.01			0.00

Unstandardised regression coefficients are reported.

^Ϯp <.10 * p <.05 ** p <. 01 *** p <.001

Table 6. Multilevel regression analyses on teacher presence as a predictor of group and individual performance.

	Group performance			Individual performance
	B		SE	B		SE
Covariates
General track (vs vocational)	−2.19	***	0.30	−4.26	***	0.73
Pre-university track (vs vocational track)	−0.91		0.90	−2.37	**	0.79
Familiarity with genetics	1.11	***	0.30	2.81	***	0.63
Experience with collaboration	0.42		0.26	−2.43	***	0.63
Interest in biology	−0.11		0.09	0.43	^Ϯ	0.25
Biology grade	0.34	**	0.11	0.88	**	0.29
Gender	0.00		0.15	0.21		0.54
Experimental condition
Teacher present	0.02		0.40	0.24		1.05
Variance
R^2	0.37			0.24

Unstandardised regression coefficients are reported.

^Ϯp <.10 * p <.05 ** p <. 01 *** p <.001

Discussion

The present study had two goals. First, the relation between students’ need satisfaction and student achievement was investigated in the context of student collaboration. Similar to findings from research on need supportive teaching in whole classroom settings (e.g., Stroet, Opdenakker, and Minnaert 2013), satisfaction of students’ basic needs for autonomy and competence during CL was found to be a positive predictor of performance. Satisfaction of the need for relatedness, however, was negatively related to individual performance. Second, the role of teacher support in achieving satisfaction of students’ basic psychological needs and student achievement was examined. Students who were randomly assigned to collaborative groups that were guided by a teacher experienced similar levels of need satisfaction and showed similar performance outcomes compared to groups without a teacher. This study thereby combines and extends the bodies of research concerning SDT and CL. Below, the results will be discussed in more detail.

Need satisfaction and performance

Based on studies in work settings and whole classroom settings (e.g., Stroet, Opdenakker, and Minnaert 2013; Vansteenkiste, Ryan, and Soenens 2020), our hypothesis was that need satisfaction during CL would be positively associated with performance. Our hypothesis was partly supported by the results: satisfaction of feelings of competence was related to both individual and group performance, and satisfaction of the need for autonomy was related to individual performance. Satisfaction of those needs was related to more favourable performance even after taking into account prior grades, familiarity with the topic, interest, and other relevant background characteristics. For these two psychological needs, the positive relation with performance is thus similar to results in whole classroom settings (Stroet, Opdenakker, and Minnaert 2013) and similar to results of the few studies that examined need satisfaction during CL (Hänze and Berger 2007; Liu et al. 2009; Minnaert et al. 2011). Hence, in combination with earlier studies, there is strong indication that a feeling of competence plays an important role for the performance of collaborating groups as well as for the learning gains obtained by its individual members.

Concerning autonomy, it was found that experienced autonomy did not predict group performance, although there was indeed a trend towards this finding. We did find evidence that a high sense of autonomy experienced by all group members, which is indicative of equal involvement of all group members, appears to be important for obtaining successful individual performance outcomes for all group members. This finding is also in line with a body of research in the field of CL that stresses the importance of groups structuring or regulating their collaborative process (Järvelä et al. 2016). For educational purposes, it seems to be advisable to help students to develop strategies to effectively regulate the collaboration in order to feel competent and autonomous.

Interestingly, in contrast to findings in whole classroom settings (Stroet, Opdenakker, and Minnaert 2013) and the findings by Liu et al. (2009) and Minnaert et al. (2011), satisfaction of the psychological need for relatedness was found to be a negative predictor of group achievement. Based on research in the fields of SDT and CL, it was expected that a feeling of trust and community would form the basis to engage in constructive argumentation that can lead to learning (Asterhan and Schwarz 2016). One might argue that feelings of relatedness led students to be more inclined to engage in off-topic discussion, which means less time was spent on the group assignment. Another explanation could be that increased feelings of relatedness may have led to a version of social loafing (Simms and Nichols 2014), which means that as a result of increased trust in each other, students may have relied heavily on their peers to solve the task instead of putting in all their effort themselves. It would be good to examine the concept of relatedness in the context of CL in further research to distinguish between its various aspects (such as familiarity, trust, and group atmosphere) and to examine under which conditions relatedness contributes to or undermines performance.

Influence of teacher presence during CL

The second research question concerned the influence of the presence of a teacher on students’ experienced need satisfaction and students’ performance in the context of CL. We compared collaborating groups that were guided by a teacher to groups in which no teacher was present. Our hypothesis was that the presence of a teacher during CL would enhance students’ need satisfaction thereby stimulating student performance (Van Leeuwen and Janssen 2019). However, the results show that teacher presence was not associated with either the satisfaction of students’ needs, nor with performance. This finding is surprising, because we expected the presence of a knowledgeable teacher during the discussions to have a profound influence on the strategies students used for solving the task and for the level of understanding the groups would reach about the task. Two potential explanations can be offered.

First, one possible explanation for the finding that teacher presence had no influence on students’ experienced need satisfaction nor on performance may be that students in the groups without a teacher offered the same kind of support as the teacher did. As discussed, during CL, students within a small group form each other’s social context (Webb, Farivar, &Mastergeorge, 2002) and thus they can be considered a source for fulfiling each other’s psychological needs for autonomy, competence, and relatedness. It could be that in the absence of a teacher, one or more of the group members take on a teacher-like role. If the degree of support that was offered in both types of groups was roughly the same, it could explain why teacher presence did not affect students’ experienced need satisfaction. That is, groups without a teacher may have been able to manage their group processes equally effective as the groups with a teacher present.

Second, the specific nature of the task used in this study may also account for some of the findings. The genetics task that students collaborated on was relatively short and had clear-cut correct and incorrect solutions, whereas CL is also often employed for students to work on complex, so-called “wicked” tasks that take more time to solve (Munneke et al. 2007). In case of a more complex task, the support of the teacher may have contributed to need satisfaction or performance. More research is needed to examine whether the effect of the presence of a teacher depends on task-related factors such as task complexity.

Limitations and directions for future research

The results of this study must be regarded in light of some limitations. First, while the effects of teacher presence on need satisfaction and achievement were examined with an experimental design, the relations between need satisfaction and achievement were examined by means of a correlational approach. The results of the present study can therefore not demonstrate the causal direction of these relations.

Furthermore, the findings were obtained in the specific context of secondary school education, in the domain of biology, and for a specific task. One might argue that in domains other than natural sciences or in other types of tasks, the nature of collaboration and the factors that determine whether collaboration is successful, are different. Also, the present study used a short period of time (one experimental session) so that potential differences in need satisfaction might not have unfolded. Thus, it is recommended that future research investigates the relation between need satisfaction and teacher presence in the context of CL in other domains and with other types of tasks, including longer-lasting and more ill-structured tasks.

A contextual factor that may have affected the ecological validity of the present study concerns how teachers enacted their guidance. That is, for research purposes, each participating teacher was continuously present within one specific group. In a naturally occurring classroom situations, teachers typically walk around between groups and engage in a few episodes of discussion for each group (Greiffenhagen 2012). The setup of the current study was chosen as it maximised the difference between groups with and without teacher support, which means it is even more interesting that no effect of teacher presence was found. More research is needed concerning the role of the teacher in supporting students psychological needs during naturally occurring forms of CL.

In addition, no specific instruction was given to teachers with regard to the type of support they should offer their students. Hence, teachers may have varied in their strategies and some teachers may have used need-thwarting strategies. If they were instructed about how to offer support in need-supportive ways, their presence might have contributed to the experience of need satisfaction by students. Future research could aim to systematically vary the type of support students receive from the teacher. Some studies have investigated the effect of teacher guidance by purposefully instructing the teacher to focus on a specific aspect of collaboration, for example either the product versus the process of collaboration (e.g., Asterhan, Schwarz, and Gil 2012; Pijls, Dekker, and Hout-Wolters 2007). A similar methodological setup could be used to compare teacher guidance that focused specifically on one or multiple aspects of psychological need satisfaction. Furthermore, future research could zoom in on the exact behaviours teachers and students display during CL to explain which specific behaviours underly need support and experienced fulfilment of those needs.

Conclusion

In this study, we aimed to integrate research on CL with research from SDT. We extended previous findings of SDT by showing that its main findings regarding the importance of autonomy and competence satisfaction can be extended to the domain of CL, while also showing that relatedness can in case of CL be counterproductive. In addition, we extended research on CL by showing the importance of autonomy and competence satisfaction for successful collaboration and exploring the role of teachers with regard to need support during CL. We were able to show that experience of autonomy and competence during CL contributes to successful CL and that the presence of a teacher during CL does not automatically enhance need satisfaction or performance. In order to obtain successful CL in classrooms, it is worthwhile to examine how teachers can optimally foster need-supportive interactions between group members.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Notes on contributors

Anouschka van Leeuwen

Anouschka van Leeuwen is assistant professor at Utrecht University. Her research interests include collaborative learning, learning analytics, and blended learning.

Lisette Hornstra

Lisette Hornstra is associate professor at Utrecht University. Her research interests include students’ motivation for school, equal opportunities, social and ethnic diversity in schools, and teaching high-ability students.

Barbara Flunger

Barbara Flunger is assistant professor at Utrecht University. Her research focuses on the question how students’ motivation and learning behavior can be improved, considering the impact of students’ individual differences.