Identification of pupils’ preferences of patterners and dramatists in secondary school computer science education

With the increasing integration of computer science into school curricula, a growing number of pupils are coming into contact with this subject. To get as many pupils as possible interested in computer science, the teaching strategies and methods must meet the heterogeneous prerequisites of the pupils. The purpose of this paper is to find out how to identify two different student behavioral preferences mentioned in the computer science education literature. Considering these preferences, the focus was set on the preferences for socially enriched learning on the one hand and task-oriented learning on the other hand. Even though related research has been found on the preferences of preschoolers and adults, research is missing in the field of computer science education at the secondary school level. Thus, this study focuses on the learning preferences of pupils in secondary school education regarding the two behavioral preferences. As a result, an instrument was developed and piloted that, validated by qualitative methods, measures these different preferences. Preliminary exploratory results from analyses with school classes show that preference clusters can be found in secondary education. Further research aims to use these findings to improve computer science education.


Introduction
There is no doubt that the need for comprehensive secondary school education in computer science is essential to society [1]. In our digital-driven society, computer science makes an essential contribution to general education. Due to digitization, it is also necessary for the discipline to establish itself in the education system to prepare future generations for studies and career prospects in this field [2]. Computer science provides the fundament for digitalization. As digitization also has an important part to play in pupils' lives, the aim is to enable them to use media in their private lives as responsible citizens, and also to qualify them for the job market. Thus, computer literacy is important in both private and professional life. Lack of computer literacy can become a significant barrier to students and society in general. In the long-term, a lack of computer skills could have a negative impact on people's life and could lead to a shortage of skilled workers. Hence, teachers in computer science face the challenge of adapting teaching methods to different groups of pupils to motivate as many pupils as possible. There are many reasons why students are more or less interested in computer science. Among the various sub-aspects of why some pupils feel more or less involved, are their previous experiences with computers [3] or the different upbringing of girls and boys [4]. Some pupils fail especially in science classes, e.g. physics, although they show overall age-appropriate learning behavior [5]. It is not yet clear why these bright students are failing. With significantly fewer women than men studying computer science, special research is being conducted on how to improve girls' engagement in computer science, e.g. in [6,7]. These studies highlight that girls tend to prefer a social, contextual approach to learning. It is also described that girls are more likely to choose narrative topics for programming projects [8]. Additionally, in the context of career choice, the interests of people were surveyed, and it was found that women are more interested in people and tend to prefer social professions [9]. Thus, interest in socialization and contextualization could influence girls' approaches to learning topics and career choices. Furthermore, independent of gender, the influence of social enrichment on learning is described in e.g. [10]. For example, computer science learning material could be socially enriched by adding background stories and figures to the topic, even if no social context is required for this subject. In this case, additional texts and illustrations are added to the task. However, this enhancement is discussed critically, as it could be a decisive factor for the motivation of some students, while others may feel hindered by added information [10][11][12]. With regard to the different preferences for tasks with more or less socialization, two groups of interest crystallize out: Learners with high interest in socialization and learners with high interest in objects, laid out in detail as follows [13,14]. These preferences should not be seen as conflicting, but as an amplitude that could each be equally well or little developed [14,15]. Students with high socialization interest may be motivated to STEAM topics through the use of socially enriched materials, while students with high interest in objects may be distracted by it [11]. Thus, it is suggested that social enrichment should be chosen carefully with respect to the target group because only learners who prefer social enrichment in learning benefit, while others perform better without it [10]. Hence, this study focuses on the identification of preferences in K-12 education and researching whether socially enriched informatics material improves motivation and learning for pupils with preferences for socialized tasks and if pupils with preferences for taskoriented material feel constrained by socialized tasks. To examine the effects on motivation and learning outcomes, instructional methods will be adapted to match the students' dispositions. The mentioned preferences can be found in various research areas using different terms and descriptions, as listed in Table 1 and described in more detail as follows.
The main part of this study focuses on the terms patterner (preference for spatial dimensions and object properties) and dramatist (preference for social aspects) [11,13] for mainly two reasons: they are well defined in the literature, and in comparison, to other studies focus on children and not on adults. Thus, the definition of these terms will be laid out in more detail as follows and serves as a basis of the research method.

Research question
The overall goal of this work is to identify the extent to which pattern and dramatist preferences can be taken into account to improve motivation and learning outcomes in computer science. Thus, this overall research question drives this study: Can the preferences of patterners and dramatists or a mix of preferences be identified in the context of computer science education?
The remainder of this paper is structured as follows. First, we identify closely related work followed by the theoretical background. Then, we describe the methods applied, followed by the results. We conclude the paper with a discussion and outlook on further follow-up perspectives. Terms considering "preference for social aspects" Termsconsidering "Preference for spatial dimensions and object properties" [11,12,16] Dramatist Patterner [10,14] Person Oriented Thinking Spatially Oriented Thinking [17,18] Task and Relation Orientated Strictly Task Oriented Communication Style Communication Style [19] Interest in People Interest in Things 1 3

Work identification of preferences
Preferences for social or spatial aspects are expressed in different research projects and areas, in which, for example, adults are surveyed, or pre-school children are observed (e.g. [9,11,14,16]). A focus was laid on these preferences, because of their possible pedagogical impact as discussed in [10][11][12]. It is proposed that considering these preferences might be a crucial factor in engaging pupils, e.g. [11]. As research on preferences can be found in various areas, terms and description differ depending on domain and focus (exemplary listed in Table 1). Related research projects differ mainly in the research area (e.g. university, profession, or pre-school) and therefore also in the age of the subjects. For example, in [10,14], the expression of the subjects were determined e.g. by means of questionnaires and categorized in terms of spatially oriented thinking or person-oriented thinking. This means subjects with spatially oriented thinking show preferences for spatial representations and subjects with person-oriented thinking for interpersonal and interpersonal aspects. A link to task preferences is drawn in [18], where subjects are divided according to strictly task-orientated communication styles and task and relation-orientated communication styles. Learners with a strictly task-orientated communication style, rather prefer tasks and representations in objective ways, while learners with task and relation-orientated communication styles are more subjective and interested in personalizing tasks. In [19] interests for e.g. people or things are questioned in connections with professions. Inclinations for social or spatial dimensions were mainly researched in adults [14,19], but comparable preferences or so-called play styles were also observed in children at Harvard´s Project Zero [16]. In this project preschool children were observed and categorized according to their play behavior in patterners and dramatists: patterners prefer building and creating patterns, like puzzles or play with building bricks, while dramatists shine in social interaction, storytelling, and pretended play. This leads to the definition of Patterner: spatial oriented, interest in properties and functionality of objects, and Dramatist: personoriented, preference for social dimensions in experiences. It is stressed that both, thinking styles and play styles, should not be seen as conflicting, but as an amplitude that could each be equally well or little developed [14,15]. As we focus on pupils (and not adults) and as the terms patterners and dramatist are well defined, those will be used as a basis and described in more detail as follows.
In summary, preferences were mainly identified in adults [9,14] while dramatists and patterners were identified by observing preschool children [16] and interviewing educators [20]. Thus, those preferences or play styles were determined in pre-school children and in adults, but no validated method to identify preferences of pupils in K-12 education was found.

Educational relevance
The impact of preferences for social or spatial dimensions on educational processes is discussed in e.g. [10][11][12]. For example, [10] first identified thinking styles in students and afterward presented adapted learning material in a course of information technology. The learning material was presented twice to students, once socially enriched by added figures and personal details of the figures, like age and hobbies, and a second time in a neutral way with a focus on buildings and their location. Learning effects in terms of motivation and memory retention were measured. Results show that only some students benefit from social enrichment and that there is evidence that those results correlate with thinking styles. In [18] students were subdivided in terms of either strictly task-oriented communicative style and task and relation orientation communicative style and subjects had the choice between characters in two distinct styles: one rather realistic, the other iconic. Results show that subjects with a strictly task-oriented communicative style rather prefer photo-realistic characters, while subjects with a task and relation orientation communicative style rather chose an iconic character. The study stresses the variability in users' preferences and suggests consideration of communication styles. In addition, [17] discusses the social context in learning materials and highlights several studies suggesting that the use of virtual characters in learning materials increases motivation and feelings of ease or simulates essential learning behaviors, but these benefits are not well researched empirically. It is stated that there is variability in how learners perceive virtual characters and that this variability should be respected in the design of learning material.
From experiences of robotics workshops [11,12] and a preliminary study in prior work from this paper's author [21] it became visible that the motivation of the pupils with dramatists' preferences, named dramatists as follows, improved, after teachers adapted the assignments on pupils' preferences, for example, storytelling or social interaction. To engage pupils, robotics can be used in various learning domains as discussed in [22]. Nevertheless, some pupils do not benefit from using robots as an object but prefer the contextualization of robotics workshops as laid out in [12]. The authors point out that challenges like the very successful LEGO Mindstorm League are attractive for some people, but others prefer working with framework stories, such as creating objects for a model of an amusement park, and enjoy not the competition but the exhibition of their art. At MIT Media Lab, computer science is embedded and contextualized in stories so that students have the opportunity to learn programming by designing interactive objects for a given story. With the title A Day in the Park [12], learners designed interactive model objects that symbolized a park, for example, carousels, insects, flowers, and hanging lanterns, by using programmable blocks enriched with craft materials. It is stressed that for some learners' social enrichment was a main motivational factor to perform programming while others preferred neutral material. From this, it was concluded that socialization of the learning material is crucial for some pupils, as they might never become engaged in robotics otherwise. Analogous experiences are described in [11,23] where it is indicated that dramatists need socialized learning paths. [11] describes in "Lifelong Kindergarten" a group of elementary school children, who were supposed to build an object with a LEGO Robotic Kit and should gain first programming experience. The group with patterner preferences, immediately built an impressive object, the group of dramatists started a project but interrupted the work to first develop a framework story. Only when the framing story was satisfactory, they created an object that was as complicated as the one of the patterners, but they needed much more time. Moreover, [11] stresses that the dramatists would not have achieved a result if the workshop had ended earlier. He also points out that not taking play styles into account can lead to a negative attitude towards science for dramatists, when classes are designed for patterners only and that those play styles can be found among learners of all ages. In a pilot study by this paper's author, pupils with dramatist preferences were asked to navigate a robot through a maze [21]. The pupils were not motivated to perform the task at first and enjoyed it much more after we added social interaction and focused not on working with the robot, but on using the robot as a toy to play against another pupil.

Theoretical framework
In [13,16,20,24] preschool children were observed and categorized into two groups according to their play behavior. One group of pre-school children were mainly interested in objects and were named patterners, while the other group showed more interest in socio-dramatic play and was called dramatists. In [13] the types of patterner and dramatist are described as follows: (a) The patterner who exhibits a sensitivity to the objective dimensions of experience like quantity, shape, size and color.
As a result, a patterner excells in visual-spatial tasks and in the creation of designs and systems. Individuals of this bent frequently display a complementary reticence, if not a disability, in domains which demand the construction and exploration of more social and subjective forms of experience like discussion, story-telling, and pure fantasy; (b) The dramatist who is alive to many of the personal and social dimensions of experience, like roles, attitudes and strategies. Consequently, dramatists shine in games, sociodramatic play, arguments and negotiations. However, these individuals typically grow bored and frustrated, performing poorly when asked to pursue highly organized tasks which demand close attention to the properties of objects (e.g., puzzles, counting, classification, etc.
Furthermore, in [13], object-dependency is discussed where patterners who are highly interested in the properties of objects prefer to play object dependent and dramatists with their stronger interest in their environment and storytelling prefer an object-independent play. Based on this, [15] suggested that patterner and dramatist should not be seen as a label for types of children, but as a dimension, where a child could rate high, low, or balanced on each dimension. For example, a child could rate high in the dimension dramatist and low in the dimension patterner but could also rate high in both dimensions. Related observations were found for example in [25] where high and low fantasy predispositions of children are researched by also observing and questioning their play behavior. [26] describes pupils who are more likely to have preferences for either linguistic or logical topics and suggests that these preferences for teaching should be used and the teaching methods should be adapted to meet the preferences of the pupils. Besides the mentioned research on children, further indicators of evidence can be found according to the preferences of adults. In [10,14] thinking styles of adults were determined by means of a questionnaire and interviews, and the perception of socially enriched material of each group was researched. Also, in 1 3 [19], the different interests of women and men in things are discussed. The authors conducted a meta-study including 47 interest inventories and 503,188 respondents, showing that men prefer working with things and women with people.

Patterners preferences
• Display considerable skill and interest in making patterns, structures, and spatial arrangements with objects and materials.
• Show considerable skill and interest in making patterns, structures, and spatial arrangements with objects and materials.
• Interest in an object's mechanical and design possibilities, rather than in communication or interpersonal events.

Dramatists preferences
• Exhibit a strong interest in human surroundings: what others do, feel, and how they can be known.
• Prefer and do well in games, socio-dramatic play, and storytelling.
Based on these preferences, the research approach and instrument were developed. Thereby, it is investigated if pupils show patterner, dramatists, or a mix of propensity. As explained in detail as follows, this is done using a mixed method approach triangulating surveys and interviews.

Methods
This article addresses the question of how to engage as many learners as possible in computer science and leave no one behind. To this end, we develop an instrument based on [10,11,16] to identify the different preferences of pupils according to the definitions of patterners and dramatists. The study follows a sequential explorative mixed-methods approach aiming at investigating the research question and generating hypotheses [27]. For that purpose, first, the quantitative data is collected, and the qualitative data is used to explain the quantitative data and to give additional information about the preferences (compare Fig. 1). The methods are supplementary, whereby quantitative surveys are conducted and afterward triangulated with qualitative interviews.

Participants
The intervention was conducted in a private school with 34 pupils (15 females and 19 males, age range = 10-14 years) in grades 5-8. In this school, pupils learn together in grades 5 and 6 and in 7 and 8. The youngest student in grade 5 was 10 years old and the oldest in grade 8 was 14. According to the teacher, there is a high achievement gap at the school, where high ability pupils but also pupils with learning and concentration difficulties learn together. The study was conducted during a Computer Science course offered by our department.

Research design
In the quantitative part of the study, a survey is conducted, and in the qualitative part a guided interview. All parts are triangulated to answer the research question. In Fig. 2 the research instruments are depicted. All components are laid out as follows in detail.

Procedures
Considering previous approaches used in the literature to classify preschool children or adults, e.g. in [13,14], it turned out that those approaches cannot be transferred one-to-one to interviewing teenagers. They are therefore

Research
Discover Education (2022) 1:11 | https://doi.org/10.1007/s44217-022-00011-0 1 3 adjusted to match the age of pupils in K12 education and school context. For example, in [13], pre-school children were observed in their play behavior, with a focus on their preferences considering inventing stories using dolls or showing interest in objects with building bricks. Naturally, the inclinations of the pupils' change, as they grow up. In order to research the preferences based on the original definitions, the approaches were adapted. In the following, two preferences that are stressed, e.g. in [13], are listed and their assessment for pupils in K12 education is discussed.
• "Patterner preference: display considerable skill and interest in making patterns, structures" As in literature, e.g. in [13], a focus is set on the preference for making patterns and structures, this preference seems to be a main patterner preference. But no validated tool was found to measure this inclination reliably in K12 education. Former methods e.g. observing play behavior were not practical as explained above. However, it was noticeable that in preliminary studies, patterners showed a tendency to be interested in mathematical tasks. This could be due to the possibility that working with abstract numbers might attract pupils showing a propensity for patterns and structures. Considering the common background of the pupils in mathematics and the existence of evaluated research methods for mathematics preferences, the amplitude of mathematics preferences seems to be a decisive variable for our study. In order to investigate this, mathematics preferences are evaluated in the survey and verified in the interview. • "Dramatistic preference: prefer and do well in games, socio-dramatic play, and story telling" Furthermore, in literature, e.g. in [13], the storytelling preferences of dramatists are stressed. Precisely like the patterner preference above, it was considered as not practical to question former preferences and no validated tool was found to assess this preference. Nevertheless, preliminary studies revealed that dramatists in K12 education show an inclination toward writing essays. In contrast to the mathematics preferences, no validated instrument was found to assess a propensity for story writing. Therefore, the questionnaire for mathematics was adapted for story writing, as listed in Table 1. To verify the questionnaire and gain deeper insights, we ask about storytelling experiences also in the interview.

Data collection
To assess patterner and dramatists' preference by means of a quantitative method, a survey was conducted, which bases on the considerations explained above in the caption Procedures. Hence, the preference for story writing and mathematics are measured. First, mathematics preferences are collected, based on TIMMS [28], afterwards data about preferences for story writing are determined, as depicted in Table 2. Each category includes nine questions that pupils can strongly agree, agree, disagree, and strongly disagree with on a 5-Point Likert scale.

Qualitative data collection
The interview consisted of two parts: first the pupils were asked to organize figures and second, they were asked questions along a guided interview, both explained as follows. The pupils are individually interviewed by one interviewer, the interview took about 15 minutes and was voice recorded. According to [29] pupils can be categorized by observing how they arrange figures on sheets of paper. This means, that the pupils face twenty figures and four sheets, and they are asked to arrange the figures on the sheets. Using this setting, patterners tend to organize characters by categories (compare Fig. 3). On the other hand, dramatists tend to invent stories (compare Fig. 4). To validate this approach, the pupils were asked to distribute twenty Playmobil figures on four pieces of paper. The Playmobil figures were five play figures, five bicycles, five media devices and five animals.

Guided interview
Afterwards, a guided interview was conducted, the items of the interview are listed as follows. The structure of the interview is following: First, the results of the quantitative survey are questioned to gain deeper insights into the motivations of the responses; second, based on the literature, preferences for dramatist and patterner are queried; and third, also based on the literature, fantasy disposition and play behavior are queried. All three parts are explained in detail as follows. In the first part of the interview, the results of the survey are questioned in order to triangulate survey and interviews. For example, the interviewer asked if this opinion has changed recently, whether the student really dislikes the subject or just certain parts of it. After the verification of the quantitative survey, interest towards patterner and dramatists' preferences are questioned, based on literature. For example, as described in e.g. [11,14], one indicator could be social preferences in learning and preferences for contextual tasks. Thus, the pupils are asked how much they like learning with other pupils and if they rather prefer contextualized tasks or factual tasks. In literature, the fantasy predisposition and play behavior are also discussed in the context of patterner and dramatist preferences e.g. [16], thus questions about play behavior and preferences for fantasy topics are added. All questions were open-ended, and students were motivated to speak freely about their preferences. The interviews were voice recorded and transcribed. The answers to the questioned preferences were rated high, medium, low, and undecided by the interviewer. For example, if a pupil stated he likes mathematics very much, the interviewer rated the mathematics preference as"high". The guided interview has the following structure:

Analysis
First, the quantitatively collected data are analyzed and presented using descriptive methods. The qualitative interviews were analyzed in a deductive qualitative content analysis [30] as follows: the voice recorded interviews were transcribed and the answers were ordered according to the interview guideline listed above. Responses to the following questions were categorized as high, medium, or low. Thus, pupils could show high, medium, or low preference for each of the categories. As laid out in the section Procedures, mathematics preferences seem to be a decisive variable. Therefore, pupils were categorized in "Low Math Preferences" "and" Medium/High Math Preferences". After this presorting, the remaining results were summarized along the guided interview. First, the quantified data was listed in an overview table. Afterward, to gain deeper insights into the responses, the results were described textually along the guided interview. From both approaches, the quantitative and the qualitative, we draw conclusions in the analysis by triangulating how the differences are situated with non-parametric statistical tests in the differences identified from questionnaires and from interviews.

Findings
This section presents the main findings of this empirical study, the goal of which was to determine whether patterner and dramatist preferences, or a mix thereof, can be established in the context of computer science education. To this end, the results of the questionnaire survey are reported first, followed by the results of the interviews. In a third step, the differences between these two results are statistically tested.

Assessment of mathematics and story telling preferences
The results of the questionnaire are illustrated in Table 3 and show the preferences in the areas of mathematics and storytelling. It shows that the mathematics preferences in almost all items are rated higher with a similar standard deviation than in the stories telling. The exception is item MP4/SP4, where the stories telling preferences show that interesting things were more likely to be noticed than in the mathematics preferences.

Reliability of the instrument
In order to test whether the items worked, we examined the reliability of the items and the scales. Analysis of the internal consistency of the assessment instruments are presented in Table 4. The scales Preference for Mathematics, Performance in Mathematics, Preference in Story Writing and Performance in Story Writing worked well.

Interviews
In the following, first the results of the interviews are summarized and depicted in tables, afterwards, the guided interview is described in detail.

Summary of the interview
Considering the overall results in the interview, pupils showed inclinations for patterner preferences, for dramatists' preferences and for a mix of preferences, the propensities varied in their amplitude. Also, some pupils could not be categorized. The clustering is listed in Table 5.
The results of this overview are depicted in Tables 6 and 7. Pupils with low preferences for mathematics showed a tendency for preferring writing stories, a higher preference for imagination, preferred collaborative learning, and rather preferred framework stories in assignments, compare Table 6. As depicted in Table 7 pupils with a high preference for mathematics were more likely to indicate that they dislike writing stories, prefer to learn on their own, have a rather low preference for imagination, and prefer more factual tasks. Additionally, in the guided interview, the pupils were asked to arrange figures on paper, as described above. Five lower schoolboys (5th and 6th grade) with low math preference created a story with characters. All other pupils arranged the characters by category. Some pupils indicated medium preferences for learning, showed learning problems, or were not motivated to participate in the interview. The responses of these pupils are summarized at the end of this chapter.

Detailed description of the interviews: pupils with low preference for mathematics
Seventeen pupils (nine girls and eight boys) reported low or no preferences for mathematics in the interview. Most of them indicated moderate or high motivation in story writing. Most of these pupils indicated that they prefer to learn   with others. A detailed review of the responses is listed below, following the setup of the guided interview as described above in the section qualitative assessment.
1. Mathematics preferences: All seventeen pupils indicated that they dislike mathematics. Twelve pupils showed strong dislikes and used strong words such as"I hate math, " seven pupils said they wished they didn't have to learn math, five pupils mentioned they did well in math but still found it boring, superficial, or they just didn't like working with numbers. 2. Story writing preferences: Sixteen pupils indicated that they liked stories. Twelve pupils said that they sometimes write stories at home. Two girls talked about how they write stories at home and then add pictures. One girl mentioned:" I have a storybook, on the left the lines and on the right, I draw something to it". Two pupils reported that they like to write stories as Christmas gifts. Two pupils said they started writing books or scripts just for fun. Two pupils mentioned they often go around making up stories. One boy said he used to like making up stories, but he doesn't want to do it anymore because it feels weird, he also said:"I used to like writing stories, but I do not write anymore even though I still like it. I want to concentrate more on school because I think that I don't learn enough when I just keep writing. " Two pupils indicated that they like to make up stories but don't like to write them down. One boy said that he did not like writing or making up stories. 3. Joint learning: All seventeen pupils said they rather prefer to learn with others or at least with others in their room.
Eight pupils indicated that they strongly prefer joint learning, they mentioned, that if alone: they feel uncomfortable, bored, do not have fun doing tasks, do not like to sit in a room by themselves, and cannot concentrate or motivate themselves. One boy said,"I don't like learning alone, I feel uncomfortable", and one boy said,"I don't like learning alone. That's why I always want to have a partner to do it with me. Then you can help each other, secondly, you get to know each other even better and build an even better friendship, and it's just a lot more fun to learn when you learn together. " Six pupils stated that they like joint learning but did not give closer explanations. Three pupils said, they usually like it, but for them, it depends strongly on the subject and the team. 4. Play behavior: Fifteen pupils remembered doing socio-dramatic play using plastic figures, dolls, or stuffed animals.
Five of them said, they also liked to build objects, but mostly for their figures to play with them. One boy said, "When I was younger, I only played with the figures, but that changed over time, later I cared only about building and not about playing anymore. " One girl talked about inventing games with her friend and one boy did not remember much of his play behavior. 5. Open questions towards fantasy preferences: Twelve pupils indicated that they have a lot of imagination and told stories in which they invent things or create artwork. Six of them also stated that they often daydream. Four pupils said they think they sometimes have a fondness for fantasy, but not always. One pupil said he didn't have much of a fantasy preference. 6. Task preferences: Pupils were asked if they preferred tasks that included a framework story or more fact-based tasks.
Eight pupils indicated that they strongly preferred framework stories in tasks. One boy was quite excited about the question and immediately started making up a story for a task. One pupil said, "I prefer assignments with stories because I also like to read and I am happy when there are assignments with stories, I think it is really cool." Seven pupils preferred a mix of tasks and stated that framework stories are usually more fun, but sometimes the stories are boring and don't make sense. Also, sometimes it takes more time to understand the task. One pupil said, "It depends, I actually like both tasks, sometimes I think it's good, when there's a story and you know what you can use it for (…) but, when there's a story you have to look at it and read through it first and then you need more time. " One girl said, that in a group she prefers structured tasks because the work is done faster, but on her own she prefers tasks with stories because it motivates her. Three pupils showed no preferences. 7. Additional Task: Figure Arrangement: Five boys (all in class level 5/6) created a story using the figures. For example, a boy commented on the process like this: "There is something like a zoo, the media things, I put here because it should be like filming the animals and the children, in means of a documentary. The girl has a bike and her boyfriend, and she was looking at the animals. Now I don't really know, there is a bicycle lying around and the cat looks at it. Now the old man, maybe he has some kind of work with the telephone (…) and the two of them here arranged to go for a bike ride. " All girls in this group and three boys ´ (class level 7/8) chose an object-oriented approach by organizing the figures according to categories, for example, vehicles, and devices. One girl from class 7/8 was asked if she could imagine inventing a story with the figures. She said, "I could have done that, but if I have to do it quickly, I prefer to do it that way. I could have thought up a huge story now, but that would have taken longer, I thought I should do it quickly".

Detailed description of the Interviews: pupils with high or medium preference for mathematics
Seventeen pupils showed high or medium preferences for mathematics. The pupils split up in two groups, nine pupils showed a very high tendency towards patterner preferences and eight showed mixed or undefined preferences. Not all pupils responded to all questions in the interview, which is commented in the description of the responses. First, the answers of the pupils with high patterner preferences are listed and afterwards the pupils with mixed preferences. Pupils with patterner preferences (three girls, six boys) 1. Mathematics preferences: Seven pupils stated they have high preferences for mathematics and six of them said, it is their favorite subject. One boy said:" I don't know, but I've always enjoyed math. " Two pupils think they like math, but it's not necessarily their favorite subject. 2. Story writing preferences: Eight pupils said they did not like writing stories at all. Some of them said it is hard to motivate themselves to write a story, it is not fun, and sometimes they got stuck in the middle and don´t know how to continue writing. Two of them stated to have problems with spelling. One boy said he couldn't imagine anyone enjoying writing a story. One boy mentioned he kind of likes writing at school but would not do it at home. 3. Joint learning: All nine pupils said they can learn well on their own. Five of them stated they would rather learn on their own. One of them said"I prefer to learn alone, group work is rather difficult for me, I get distracted then and it is better if I have my peace and concentrate on one task. " Four pupils said, they do well on their own but do sometimes learn with others, depending on the pupils and the tasks. 4. Play behavior: Seven pupils in this group answered the question about play behavior. Six recalled that they liked to build objects, for example with Lego bricks. Two of them stated that they also played with objects. One pupil did not remember if she had played or built with toys. 5. Open questions towards fantasy preferences: Seven pupils said that they do not consider themselves a"fantasy-type".
Four of them told stories about building objects. Two pupils said that they sometimes think they are creative. 6. Task preferences: Eight pupils answered this question. Five pupils said they would prefer factual tasks without a story.
Three pupils were undecided and said they sometimes like a context in their tasks. 7. Additional Task: Figure set up: Eight of nine pupils organized the figures according to categories. One boy created a story.
Pupils with mixed/undefined preferences (three girls, five boys) Eight pupils had mixed preference for mathematics and gave undefined answers compared to the rest of the group. Six of them did not show a profile as they either liked all tasks or disliked any kind of task which were questioned. They said, for example "I just like all subjects". Four of them showed high motivation in story writing and mathematics, high fantasy preferences and were undecided if they prefer joint learning or factual tasks. Two of them stated that they do not like any tasks, showed low fantasy preference and were also undecided about joint learning and factual tasks. Additionally, two boys mentioned they rate learning low because they have learning problems.

Differences between survey and interviews
In order to further investigate the relationship between the quantitative and qualitative data and to compare the two groups of patterners and dramatists, we have added the categories identified in the interviews to the results of the questionnaire data. In a next step, we compare these two groups. Since the K-S test did not show a normal distribution 1 3 (p ¡ 0.05), we apply the Mann-Whitney U test to compare two independent samples. Therefore, we first categorized the results of the questionnaires into the two groups (patterners and dramatists) as grouping variable, which we determined qualitatively from the interviews. Then, using the Mann-Whitney U test, we calculated the differences of these groups on the Math Preferences and Story Telling Preferences questions. Significance level was 0.05.
Results in Table 8 show that there is a statistically significant difference at all items on asking math preferences between the qualitatively identified groups of patterner and dramatists.
Results in Table 9 show that there is no statistically significant difference at item SP4 ("I learn many interesting things in story writing") on asking story telling preferences between the qualitatively identified groups of patterner and dramatists. All remaining items on asking story telling preferences between the qualitatively identified groups of patterner and dramatists are statistically significant different.

Differentiation of preferences
To compare the results from the qualitative and quantitative surveys, we calculated the difference between the two preferences. In Fig. 5 the difference between the math preferences and storytelling preferences is calculated and triangulated with the interview results. Subjects with low preferences for storytelling and high preferences for math have a negative score, pupils with balanced preferences show a score around zero, and pupils with high preferences for storytelling and low preferences for math show a high score. The results are color-coded, therefore pupils who showed patterner preferences in the interviews are visualized in green, with dramatists' preferences in blue and mixed preferences in orange.

Discussion
This paper investigated the question how to engage pupils in a formal and technical subject such as informatics in school lessons. We were interested, in particular, in the perspective of different preferences of patterners and dramatists [11] and how they can be taken into account. The aim of this contribution was to find out whether these two types of preferences can be identified in school classes in informatics lessons with the leading research question: Can the preferences of patterners and dramatists or a mix of preferences be identified in the context of informatics education?
Using a mixed method approach, the following results were obtained: Based on interviews, tendencies towards patterners and dramatists' preferences can be found in K12 education. The amplitude of the preferences varied in its expression; this means some pupils showed strong emotions towards either patterner or dramatist preferences while others had weaker tendencies or rather balanced views. It is noticeable that some pupils reported low motivation in, for example, mathematics, but also stated to have high achievement in this subject. This indicates that it would not be sufficient to compare school grades to gain insight into preferences. Additionally, learning problems could distort the result of the survey. In some cases, pupils with reading-spelling disability rated story writing very low in the survey but showed dramatist preferences when interviewed. In order to further investigate the connection between the quantitative and qualitative methods results are triangulated. The Mann-Withney U-Test shows significant differences (p ¡ 0.05) for all variables, except for one item (SP4). This further indicates that the applied scales on pupils' math and storytelling preferences identify the same preferences as the qualitative interviews do. Besides surveys and interviews, we used an approach in which pupils arranged figures on sheets of paper. This approach will not be continued with pupils in secondary school because the overall result was confirmed only for boys in the 5th grade. However, the figure arrangement could be evaluated again in the context of primary school or kindergarten.

Conclusion and future work
Digital literacy is a key skill for future generations, and no one should be left behind. An important aspect of motivating as many pupils as possible could be the adaption of science lessons to pupils' preferences or thinking styles [10,12]. This could be done by enriching learning material socially, but it is stressed in [10] that this will not be the "instructional panacea", as pupil benefit differently. Rather, it is suggested that inclinations should be taken into account in the design of learning materials and that adapted learning paths be offered. This assumption is supported by the results of the interviews, where some pupils mentioned that they feel hindered and distracted by contextualization and collaboration, while other pupils stressed, that they could hardly learn efficiently without contextualization or collaboration.
Preferences for contextual tasks and storytelling are associated with girls' preferences in e.g. [7,8], and considering professional preferences, women seem to be more likely to be interested in people and chose social professions [19]. Our results are consistent with this, as girls showed rather dramatists than patterners preferences. Nevertheless, some girls showed patterner preferences and indicated that they disliked social or contextual tasks. Almost as many boys as girls showed dramatist preferences, and stated, that they would prefer a social and contextualized learning style.
Preferences are highly individual and depend on various aspects, such as prior experiences. However, this study shows that clusters can be found in the preferences of pupils in secondary education. A triangulation of the surveys and interviews proposes that preferences for mathematics and storytelling could be good indicators to forecast preferences according to the definition of patterners and dramatists. Additionally, the interviews allow the assumption that pupils in secondary school education can judge their learning preferences well when interviewed in person. However, more research is needed to validate the presented methods. Further research will focus on addressing the preferences of patterners and dramatists in teaching informatics and evaluating the impact on motivation and learning success. Preliminary studies showed promising results when patterners and dramatists were separated and assigned different tasks. Also, in a preliminary study, dramatists were more motivated to perform programming, after the tasks were enhanced by collaboration. Both findings will be examined in more detail in future research.
In order to evaluate which kind of enhancement in learning material motivates dramatists, the following approaches will be researched: Enrichment of learning through a focus on context, storytelling, or social interaction. For instance, pupils could be motivated to add storytelling to their programming projects or to work in teams. In terms of robot programming, initial research has shown that dramatists tend not to be interested in the robot as an object, but in using it in a context or in a game played in teams [21]. In this case, a robot is just a tool that is used in a social game. Robots can also be used in narrative or creative contexts. For example, art students used robots to retell scenes from movies for an exhibition, while in a workshop, rides, and scenery of an amusement park were built based on robots [11,12].

Limitation
The study was conducted with a limited number of pupils in a private school that emphasizes teamwork and social interaction in its instructional design, which could lead to a higher number of pupils with dramatist preferences. Also, the quality of the answers depends highly on the motivation and self-reflection of the pupils. The proposed method showed overall good results but could be enhanced by more detailed approaches, like observing the pupils in selecting and conducting tasks or interviewing educators.