Providing Written or Oral Explanations? PDF

The Journal of Experimental Education ISSN: 0022-0973 (Print) 1940-0683 (Online) Journal homepage: https://www.tandfonline.com/loi/vjxe20 Providing Written or Oral Explanations? Differential Effects of the Modality of Explaining on Students' Conceptual Learning and Transfer Andreas Lachner, Kim-Tek Ly & Matthias Nückles To cite this article: Andreas Lachner, Kim-Tek Ly & Matthias Nückles (2018) Providing Written or Oral Explanations? Differential Effects of the Modality of Explaining on Students' Conceptual Learning and Transfer, The Journal of Experimental Education, 86:3, 344-361, DOI: 10.1080/00220973.2017.1363691 To link to this article: https://doi.org/10.1080/00220973.2017.1363691 Published online: 29 Aug 2017. Submit your article to this journal Article views: 985 View related articles View Crossmark data Citing articles: 19 View citing articles Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=vjxe20 THE JOURNAL OF EXPERIMENTAL EDUCATION 2018, VOL. 86, NO. 3, 344–361 https://doi.org/10.1080/00220973.2017.1363691 Providing Written or Oral Explanations? Differential Effects of the Modality of Explaining on Students’ Conceptual Learning and Transfer €cklesa Andreas Lachnera,b,c, Kim-Tek Lya, and Matthias Nu a Department of Educational Science, University of Freiburg, freiburg im breisgau, Germany; bLeibniz-Institut f€ ur ubingen, T€ ubingen, Germany Wissensmedien, T€ubingen, Germany; cDepartment of Psychology, University of T€ ABSTRACT Learning-by-explaining (to ﬁctitious others) has been shown to be an effective instructional method to support students’ generative learning. In this study, we investigated differential effects of the modality of explaining (written versus oral) on students’ quality of explanations and learning. Fortyeight students worked on a hypertext about combustion engines. Afterwards, they were asked to explain the learning content, either orally or in writing. Findings indicated that providing written explanations was more effective than providing oral explanations in supporting students to organize the content of the explanations. The higher levels of organization yielded higher levels of students’ conceptual knowledge. In contrast, generating oral explanations, relative to written explanations, triggered students’ elaborative processes to a more pronounced extent, which was more beneﬁcial to attaining transferable knowledge. Thus, we conclude that the modality of explaining plays a critical role in learning-by-explaining inasmuch as different modes differentially support student learning. KEYWORDS Generative learning; instructional explanations; learning-by-explaining MOST PEOPLE HAVE probably noted from personal experience that the act of explaining something to others helped the explainer to better understand the subject matter. Therefore, it is not surprising that learning-by-explaining is a ubiquitous instructional strategy commonly used in different educational contexts, such as peer tutoring (Chi, Siler, Jeong, Yamauchi, & Hausmann, 2001; Roscoe & Chi, 2008), collaborative learning (Palincsar & Brown, 1984; Rummel & Spada, 2005), or digital learning (Blair, Schwartz, Biswas, & Leelawong, 2006; Chin et al., 2010). Most studies on the effects of explaining either focused either on self-explaining instructional material to one-self (e.g., Ainsworth & Loizou, 2003; Wylie, & Chi, 2014), or explaining the subject-matter to others in collaborative and interactive learning settings (e.g., Palincsar & Brown, 1984; Pl€ otzner, Dillenbourg, Preier, & Traum, 1999; Roscoe, 2014). Recently, various studies also documented that explaining learning content to ﬁctitious other students is also beneﬁcial to learning and even more effective than restudying (e.g., Fiorella & Mayer, 2013, 2014; Hoogerheide, Loyens, & van Gog, 2014). The use of such individual learning-by-explaining practices have particularly increased as new online learning environments such as MOOCs (Margaryan, Bianco, & Littlejohn, 2015) or learning management systems (De Smet, Bourgonjon, De Wever, Schellens, & Valcke, 2012) allow learners to generate online-content by themselves and make it available to potential other unknown students. Generally, it is assumed that learning by explaining to ﬁctitious others can be described by three distinct stages (Bargh & Schul, 1980; Fiorella & Mayer, 2014): First, students are required to plan and CONTACT Andreas Lachner [email protected] Leibniz-Institut f€ur Wissensmedien, Schleichstraße 6, D-72076 T€ ubingen, Germany. This article was originally published with errors, which have now been corrected in the online version. Please see Correction (http://dx.doi.org/10.1080/00220973.2019.1644716) © 2018 Taylor & Francis Group, LLC THE JOURNAL OF EXPERIMENTAL EDUCATION 345 prepare their explanation; second, they have to explain the subject matter to potential others; and third, students answer subsequent questions in interactive settings or receive feedback by an instructor. Several researchers have started to disentangle the contributing effects of the individual phases of explaining to students’ learning. For instance, Fiorella and Mayer (2014) investigated the effects of preparing to explain and actually explaining content by creating a video explanation on students’ learning. Students studied a text about the Doppler effect either with the expectation to be tested or to explain the learning material; half of them afterwards, actually did explain the material by providing a video explanation, whereas the other half simply received additional study time. The authors found that preparing to explain was more effective than preparing for a test with regard to students’ acquisition of conceptual knowledge. More importantly, the authors showed that those students who really explained the learning material outperformed those students who only prepared to explain the material to others and did not actually explain it. Hoogerheide et al. (2014) replicated the ﬁndings by Fiorella and Mayer in a procedural domain. For that purpose, Hoogerheide et al. asked students to read a text on syllogistic reasoning with the intention to complete a test or to explain the content to potential other students. Again, in another condition, students explained the content by creating a video explanation. In accordance with the ﬁndings by Fiorella and Mayer, the authors found that students who created the video explanations showed higher learning and transfer performance compared with students who only prepared to complete a test or to explain the content to others. Together, these ﬁndings suggest that the mere act of explaining plays a critical role in supporting students’ knowledge acquisition, though the expectation of explaining itself may have supplementary effects on students’ learning. Generating explanations What speciﬁcally makes the act of explaining so important? In line with theories of generative learning (Fiorella & Mayer, 2015; Mayer, 2002), it is suggested that the act of explaining may trigger distinct cognitive processes that are germane to students’ knowledge acquisition. For instance, explaining generally requires students to select the relevant information from the learning material and to organize the information in a coherent manner that can be understood by others (Fiorella & Mayer, 2015). Furthermore, students need to elaborate on the information by providing concrete examples or analogies to help the potential recipient relate new concepts to her or his prior knowledge. Following Fiorella and Mayer (2015), these organization and elaboration processes should particularly contribute to the explainer’s generative learning, as these processes enable one to structure the learning material in a coherent manner and integrate new information with existing knowledge, thereby enabling the explainer’s deep understanding. Empirical evidence for the beneﬁts of explaining can be found, for instance, in the study by Roscoe and Chi (2008). The authors analyzed the effects of students’ explanatory processes on their learning. In their study, students learned the basic functions and the structure of the human eye and were asked to explain the previously given learning material. The authors found that primarily students’ higher-order cognitive processes (i.e., elaboration and organization sequences) during explaining contributed to their learning, whereas lower-order cognitive processes, such as paraphrasing, were not related to students’ learning outcomes. Generating written versus oral explanations New learning technologies, such as MOOCs or learning managements systems, however, offer various methods to generate explanations. While explaining, students can resort, for instance, to oral explanation formats that allow students to speak their explanations, by generating videos or creating audio messages. Furthermore, students can use textual explanation formats, for instance, by creating blogs or threads in forums in which they explain subject-matter to potential other students in written format. Despite the mere value of explaining to ﬁctitious others, however, the question arises as to whether such explanation formats, differing in their modality (i.e., written versus oral), differently contribute to the explainers’ generative cognitive processes (i.e., organization and elaboration) and their learning 346 A. LACHNER ET AL. outcomes (i.e., conceptual knowledge and transfer) while explaining. Although both writing and speaking can be regarded as related conversational practices, in order to convey subject matter information to potential others (Sperling, 1996), there is a large debate whether the act of producing written versus oral discourse, and the respective cognitive processes can be regarded as related or different cognitive mechanisms (Akinnaso, 1985; Chafe, 1982; Cleland & Pickering, 2006; Sperling, 1996). Commonly, speaking can be regarded as a rather universal and partially spontaneous communicative act, whereas writing is a cultural practice that requires deliberate engagement (Cleland & Pickering, 2006; Sperling, 1996). Beyond that, the outcome product differs: Whereas speaking results in producing sounds, writing involves producing characters on (virtual) pages on the basis of a cultural symbol system (Cleland & Pickering, 2006). Nevertheless, recent research demonstrated that writing and speaking tend to share similar cognitive processes and the corresponding representations (Akinnaso, 1985; Cleland & Pickering, 2006). For instance, Cleland and Pickering (2006) investigated whether people use the same syntactic strategies while speaking and writing. Therefore, the authors primed the participants with different syntactical structures in written or in oral form and asked participants to complete a number of incomplete sentences either orally or in written form. The authors found that the participants were affected only by the syntactical primes and not by the modality of the primes while completing the given sentences. This ﬁnding suggests that at least for the construction of sentences, people tend to deploy the same cognitive representations regardless of the modalities. Linguistic differences between oral and written discourse Research on pragmatic linguistics has advocated taking the contextual character of oral and written discourse into account, as the affordances of writing and speaking can be regarded as very different (Akinnaso, 1985; Sperling, 1996). Writing, in contrast to speaking, is a nonspontaneous medium (Lakoff, 1982; Sindoni, 2014). The nonspontaneous character of writing allows students to get more opportunities to externalize their ideas and carefully reﬂect upon their thoughts (Klein, Boscolo, Kirkpatrick, & Gelati, 2014). Such externalization processes permit students to reread and further develop their ideas. Thus, rereading and revising written explanations may particularly stimulate students to identify essential concepts, establish links between concepts, and thoroughly organize the subject-matter according to the underlying rhetorical structure (Hoogerheide, Deijkers, Loyens, Heijltjes, & van Gog, 2016; N€ uckles, Wittwer, & Renkl, 2005). In direct oral explanatory situations, however, students need to produce an instant response to a (potential) audience without having the time to carefully organize their explanation. The stronger affordance to realize organizational processes during writing may even be reinforced by the fact that writing is commonly regarded as a more formal type of discourse as compared to speaking. Thus, the formal character of writing may require students to stick to distinct genre-typical rules of organizing their ideas, which may additionally contribute to the overall organization of written explanations (Lachner, Burkhart, & N€ uckles, 2017). Contrarily, oral discourse is typically regarded to evoke a stronger social involvement than written discourse (Chafe, 1982; Chen, Park, & Hand, 2016) as oral discourse might give students stronger impressions to communicate to an “authentic” audience than writing does (Sindoni, 2014). Therefore, students could be more engaged to tailor their explanation to a ﬁctitious audience in oral explanatory situations than in situations in which students are asked to write an explanation. For instance, oral explanations, more so than writing, may trigger the inclusion of more direct references to potential addressees through deictic references (i.e., the use of ﬁrst-person pronouns and second-person pronouns: “I will tell you”; “You probably will know”). Additionally, explaining orally may engage students in providing authentic examples or analogies to illustrate the subject matter for a particular audience. Empirical evidence for the contextual differences of writing and speaking can be found for instance in the seminal study by Chafe (1982). Chafe conducted a descriptive study and compared the oral and written artifacts (e.g., conversations, lectures, letters, and academic manuscripts) of 20 academic scholars. He found that the written artifacts were less fragmented and better structured than the oral artifacts. Additionally, the author found that speaking triggered higher levels of social involvement than writing, expressed by a higher level of person deictic references (higher use of ﬁrst-person and second- THE JOURNAL OF EXPERIMENTAL EDUCATION 347 person pronouns). In a related study Redeker (1984) replicated the main ﬁndings by Chafe under well controlled experimental conditions. For that purpose, the author asked participants to generate explanations or narrations either in oral or written format, while keeping the number of prior planning activities constant across conditions. In line with Chafe (1982), Redeker found that oral discourse was characterized by higher levels of social involvement but a lower organization than written discourse. Effects of written versus oral explanations on students’ learning Redirecting these ﬁndings to learning-by-explaining approaches, it can be assumed that explaining in written form may result in better organized explanations, which should be particularly beneﬁcial for students’ conceptual understanding. Contrarily, the higher level of social involvement while providing oral explanations should encourage students to adapt their explanations toward a potential audience and thereby enrich their explanations by illustrative examples or analogies. Such elaborations in turn should be helpful for linking the learning content to one’s prior knowledge, which should be particularly fruitful for attaining transferable knowledge (Fiorella & Mayer, 2015). However, although there are several studies in which differential effects of oral teaching activities (e.g., peer feedback, peer discussions) and written teaching activities (e.g., learning journals, essay writing) were examined (e.g., Chen et al., 2016; Parr & Wilkinson, 2016; Rivard, 2004; Rivard & Straw, 2000), empirical studies that directly examined the effects of generating written versus oral explanations are scarce. One exception is the study by Hoogerheide et al. (2016). The authors examined the effects of providing explanations in either an oral format (i.e., on video) or a written format on students’ application and transfer of procedural knowledge. Therefore, students received an expository text in which the four central forms of syllogistic reasoning were introduced: conﬁrming the antecedent (if A then B, A therefore B), refuting the antecedent (if A then B, not A therefore not B), conﬁrming the consequent (if A then B, B therefore A), and refuting the consequent (if A then B, not B therefore not A). Afterwards, students were randomly asked either to restudy the learning material or to provide an explanation about these principles in written format with a text-processing program or in video format using a webcam tool. Last, students answered an application test and a transfer test that required the students to transfer the previously learned syllogistic rules to Wason selection tasks. In Wason selection tasks, four cards are presented on a table with a given rule (e.g., “Each card has a number on one side and a letter on the other side. Which cards must be turned over to examine the following rule: If a card shows an even number on one side, then the opposite side is blue?”). Therefore, in the transfer tasks, students had to apply the previously learned syllogistic principles to slightly different tasks. With regard to the application of rules, Hoogerheide et al. (2016) found that explaining on video better supported students’ application of knowledge than simply restudying, whereas there were no signiﬁcant differences between the writing and the video condition. With regard to the transfer of principles, Hoogerheide et al. did not ﬁnd any signiﬁcant differences between experimental conditions. Together, the ﬁndings suggest that for procedural learning such as syllogistic reasoning, generating oral explanations and written explanations can be regarded as comparably effective. However, it has to be acknowledged that the learning material by Hoogerheide et al. can be regarded as rather low-element interactivity material, as students simply had to serially process two single premises (e.g., Premise 1: “When John sees a clown, then he is afraid”; Premise 2: “John does not see a clown”) and combine the two promises to deduce a syllogistic conclusion (e.g., “John is not afraid,” see Hoogerheide et al., 2016). Therefore, the question arises as to whether the previous theoretical assumptions about potential differential effects of generating written versus oral explanations on students’ explaining and learning activities will in particular hold when students are required to learn highelement interactivity learning materials. Learning high-element interactivity material generally imposes higher cognitive load during learning, as students have to learn concepts that heavily relate to each other and so cannot be learned in isolation. Hence, the differential effects of written and oral explanation formats could become salient only for high-element interactivity materials; as such materials 348 A. LACHNER ET AL. require students to more thoroughly organize and elaborate the learning material, compared to less complex learning materials (Fiorella & Mayer, 2015). Furthermore, Hoogerheide et al. (2016) operationalized transfer as applying the previously learned information to similar contexts, as they asked novices to transfer the acquired knowledge about syllogisms to Wason selection tasks. Thus, these transfer tasks only slightly differed with regard to their surface features but, at the same time, shared a very similar deep structure compared to the learning tasks. Hence, one may assume that the operationalization of transfer by Hoogerheide et al. (2016) tends to be rather limited, as it neglects the constructive and interpretative character of transfer (Wagner, 2010). In particular, Schwartz, Bransford, and Sears (2005) argued that in many situations successful transfer requires—rather than a simple application of knowledge—a reinterpretation of the previously learned information to be able to categorize and evaluate the new information encountered in the transfer task. For instance, imagine a student who previously was asked to learn the functions and processes about gasoline engines. Afterwards, the student is asked whether the principles of combustion engines would also apply to gas turbines. To answer this question, the student would need to establish a learning representation (representation of the learning domain: gasoline engines) and a transfer representation (representation of the transfer problem: gas turbines) and detect common and dissimilar features between those two representations. Additionally, the student would need to evaluate the functional signiﬁcance of the encountered commonalities and dissimilarities to infer whether the functions of combustion engines can be transferred to the functions of gas turbines. Thus, the student would need to reinterpret the previously acquired knowledge about gasoline engines in order to solve the given transfer task. Hence, given the rather limited operationalization of transfer in the studies by Hoogerheide et al. (2016), the question arises as to whether explaining orally and in written form may cause differential effects on students’ learning when students are confronted with more challenging transfer tasks that call for a reinterpretation of the previously acquired knowledge rather than a simple application of knowledge. Research questions and hypotheses Against this background, we conducted an experimental study to investigate the differential effects of generating written versus oral explanations on students’ explanatory features and on students’ learning (conceptual knowledge, transfer) when studying high-element-interactivity material. In a ﬁrst phase, educational science students read a hypertext describing internal-combustion engines. In a second phase, the students were randomly asked to generate a written or an oral explanation for a ﬁctitious fellow student having no scientiﬁc knowledge about internal-combustion engines. As prior research (Fiorella & Mayer, 2014; Hoogerheide et al., 2014) consistently showed that explaining was superior to restudying, in contrast to the previous studies, we refrained from including a control group of students, who only studied the learning material without explaining it. In a third phase, students completed a conceptual knowledge test and a transfer test to assess their learning outcomes. We assumed that compared to providing an oral explanation, writing an explanation would more likely allow students to adequately select and structure the concepts to be explained that would result in more-organized explanations (Chafe, 1982; N€ uckles, H€ ubner, & Renkl, 2009; Redeker, 1984), contributing to a better organized and therefore more retrievable conceptual knowledge of the learning material. In contrast, as oral explanations more likely trigger students’ social involvement (Lakoff, 1982; Sindoni, 2014), oral explanations should more strongly encourage students to address a potential audience, which would result in more person-deictic references (e.g., “I”, “you”, “me”). More importantly, students’ higher levels of social involvement during oral explaining should result in moreelaborative processes (e.g., providing examples, analogies) than providing written explanations. These elaborations should particularly contribute to more transferable knowledge, as they allow students to contextualize the rather abstract principles and concepts of combustion engines already during explaining, which should help them transfer these principles to other situations in later transfer phases (Fiorella & Mayer, 2015). Based on these assumptions, we tested our hypotheses: THE JOURNAL OF EXPERIMENTAL EDUCATION 349 Learning-outcome hypotheses: With regard to students’ conceptual knowledge, we hypothesized that students who generated a written explanation would outperform students who generated an oral explanation on a conceptual knowledge test (conceptual-knowledge-hypothesis). With regard to students’ transfer of knowledge, in contrast, we expected that students who generated an oral explanation would outperform students who generated a written explanation on a transfer test. Mediation hypotheses: Following our previous theoretical considerations, we additionally tested to what extent the different explanatory features (i.e., level of organization, amount of person deictic references, and level of elaboration) mediated the effect of type of explanation modality (written versus oral) on novices’ learning outcomes. We hypothesized that the level of organization may act as a potential mediator between type of explanation modality and students’ conceptual knowledge, as writing explanations may trigger students to better organize their explanation (N€ uckles et al., 2009), which would result in higher test scores in a conceptual knowledge test (organizationconceptual-knowledge-hypothesis). With regard to students’ transfer, we examined whether the level of students’ social involvement mediated the effect of type of explanation modality on students’ transfer. This hypothesis can be assumed, as oral explanations would likely heighten students’ social involvement, resulting in more illustrative elaborations in the explanations, such as relating the learning material to one’s own prior knowledge (Hoogerheide et al., 2016). These elaborations should help students ﬂexibly apply their knowledge and as such more easily transfer the acquired knowledge to other situations (Involvement-transfer-hypothesis). Method Participants Forty-eight advanced educational science students from a German university participated in the study. Their mean age was 22.29 (SD D 3.16). Nine students were male. The students were, on average, in their third semester (SD D 1.83). The students participated in exchange for course credit. Design We used a between-subjects design with students’ learning outcomes (i.e., conceptual knowledge and transfer) as dependent variable; the semantic features of students’ explanations (i.e., the level of organization, the number of person-deictic references, and the number of elaborations of the explanations) as mediator variables; and the type of explanation modality (oral versus written explanation) as a between-subjects factor. The students were randomly assigned to one of the two experimental conditions. Accordingly, students either generated a written explanation (written explanation group) or an oral explanation (oral explanation group). Additionally, we collected subjective cognitive load ratings after the explanation phase to investigate potential differences with regard to students’ perceived cognitive load during explaining. Materials The entire experiment was computer based and implemented in the open-source LimeSurvey online survey tool (https://www.limesurvey.org). 350 A. LACHNER ET AL. Hypertext environment The hypertext environment consisted of an adapted German Wikipedia article on the four-stroke engine (see https://de.wikipedia.org/wiki/Viertaktmotor), a speciﬁc type of internal combustion engine commonly used in automobiles. The hypertext dealt with the general components and functions of a four-stroke engine, and explicitly explained the four-stroke sequence in an internal combustion engine. The text consisted of 2,091 words and included three static pictures that visualized the components of a four-stroke engine and the four-stroke sequence. Additionally, a subject-matter expert with an MSc in automotive engineering checked the correctness of the information provided in the hypertext. Conceptual knowledge test Based on the information provided in the hypertext, we constructed a conceptual knowledge test, which was used as pre- and posttest to measure students’ conceptual knowledge about internal combustion engines. The test consisted of 12 multiple-choice questions with four answer possibilities and one correct solution (see Appendix A). Five items dealt with the underlying processes of a combustion engine, four items with its components, and three items with the general functions of a combustion engine. Average item difﬁculty of the conceptual knowledge test was low to medium (M D.38; SD D.21), indicating that our test had neither ﬂoor nor ceiling effects. To ensure the content validity of our conceptual-knowledge test, again, the subject-matter expert checked the correctness of the questions and the possible answers. Participants received one point for each correctly solved item, which could yield a high score of 12. For the posttest, to avoid memory effects, both the order of the questions and the order of answer possibilities were randomized per participant. Transfer test Following Schwartz et al. (2005), we conceptualized students’ transfer of knowledge as the successful reinterpretation of the information learned from the hypertext. Thus, the transfer test consisted of two open transfer questions, which required the students to reinterpret the general functions and processes of four-stroke combustion engines and to transfer the functions and mechanisms of a four-stroke internal combustion engine to other similar engines (i.e., two-stroke-engine and a gas turbine, see Appendix B). To this end, students would need to not only detect common patterns between a four-stroke engine and another engine (e.g., gas turbine) but also evaluate the functional signiﬁcance of these commonalities and differences, for example, whether the differences would affect the general functionality of the particular engine. Again, the transfer test was checked by the subject-matter expert to ensure content validity. Thirteen points could be obtained for each question, resulting in a maximum score of 26 points in the transfer test. To facilitate students’ reasoning when answering the transfer questions, we provided them with a schema graphic of the two-stroke engine and a schematic of the gas turbine. Twenty percent of the transfer tasks were rated independently by two trained raters who had no knowledge of the experimental conditions. Interrater reliability was very good, ICC D.98 (Wirtz & Caspar, 2002). Thus, only one rater coded the rest of the explanations. Subjective-cognitive load At the end of the explanation phase, the students rated their perceived subjective cognitive load during explaining. For this purpose, students answered an adapted short questionnaire developed by Berthold and Renkl (2009). Students’ subjective cognitive load was assessed by four items on a 5-point rating scale (1 D easy, 5 D difﬁcult in response to the following: “How easy or difﬁcult was it for you to explain the subject matter of the learning environment?” “How easy or difﬁcult was it for you to convey the central information of the learning environment?”). The reliability of the questionnaire was satisfying, a D.79 (Cronbach’s alpha). THE JOURNAL OF EXPERIMENTAL EDUCATION 351 Procedure The students were tested individually in our laboratory. The students sat in front of a computer. At the beginning of the study, they were informed that they would take part in a study on learningby-explaining in the domain of engineering. Additionally, they were informed that after reading the hypertext, they would explain the subject matter they had learned to a ﬁctitious fellow student (see Appendix C, for the entire instruction) to heighten their explaining expectancy during learning (see Fiorella & Mayer, 2014). Afterwards, we obtained oral consent from the students to participate in the study, and randomly assigned students to the experimental conditions. An experimental session lasted 90 minutes. During the experimental sessions, the students were not allowed to proceed before being signaled by the experimenter (exact time-on-task). First, students answered a short demographic questionnaire (5 minutes) and then they completed the pretest that assessed their conceptual knowledge about internal combustion engines (8 minutes). Afterwards, the students read the hypertext. They were asked to study the hypertext carefully and to understand the information as thoroughly as possible (see Appendix C). The students’ effort and motivation was further supported by telling them that afterwards they would explain the subject matter to a ﬁctitious fellow student (see Hoogerheide et al., 2016). During the learning phase, the students were allowed to take notes on a separate sheet. They were informed that their notes were available only during the learning phase and the explaining phase but would not be available in the posttest phase. The learning phase lasted 30 minutes. Afterwards, students were asked to plan their explanation (7 minutes). They were asked to structure their notes and provide an outline for their explanation. Outlining has been shown to scaffold students’ organization processes (De Smet, Brand-Gruwel, Leijten, & Kirschner, 2014), which should be beneﬁcial to the overall quality of the oral and written explanations. Depending on the experimental condition, they were additionally informed that they would provide either an oral or a written explanation in the subsequent explanation phase (see Appendix C). In the explanation phase, dependent on experimental condition, students either explained the learning content orally or in written format (see Appendix C). In the written explanation condition, students typed their instructional explanation into a computer using a text editor provided in the LimeSurvey software. During writing, students had their notes from the planning phase available. In the oral explanation condition, students were asked to generate an oral explanation. For that purpose, students also sat in front of the computer screen that displayed the notes they generated in the planning phase. When the students were ready for explaining, the experimenter started recording their explanation via a separate sound recorder. The explanation phase lasted 15 minutes. After the explanation phase, students assessed their subjective cognitive load during explaining (2 minutes) and answered the conceptual knowledge test (5 minutes) and the transfer test (15 minutes). We decided to provide students ﬁrst with the conceptual knowledge test followed by the transfer test to more directly align our test procedure with the students’ transfer processes (Schwartz et al., 2005; Wagner, 2010). Analysis and coding Before coding the explanations, a student assistant transcribed the oral explanations so that they would be available in written form. Afterwards, the explanations were rated with regard to their level of organization, their elaboration, and their level of person-deictic references. Level of organization As the level of organization tends to be a holistic feature of texts, it can be more adequately assessed on the molar level of the entire text than on the molecular level of single sentences (Lachner et al., 2017; N€ uckles et al., 2009). For that purpose, we used a rating scheme developed by N€ uckles et al. (2009) and rated the organization of our students’ explanation on four dimensions. Concerning the ﬁrst dimension, we rated whether the students’ explanation was globally structured (i.e., coherent global organization of the entire explanation). Concerning the second dimension, we rated whether the students differentiated between relevant and irrelevant points within their explanation (e.g., presentation of 352 A. LACHNER ET AL. relevant principles of combustion engine, main components and functions of combustion engines). Third, we rated whether the presentation of the concepts was consistent (i.e., logical order of the arguments). Fourth, on a local level we rated whether the arguments were sufﬁciently connected to each other to enhance readability (i.e., local cohesion, see McNamara & Kintsch, 1996). For each dimension, students could receive either one point (dimension fully present), half a point (dimension partially present), or zero points (dimension not present at all), yielding a maximum score of four points, as there were four dimensions. Twenty-two percent of the explanations were rated independently by two trained raters who were blind to the experimental conditions. Interrater reliability was very good, ICC D.84 (Wirtz & Caspar, 2002). Thus, only one rater coded the rest of the explanations. Person-deictic references To analyze students’ use of person-deictic references, as a potential indicator of social involvement, we counted the number of ﬁrst-person pronouns (e.g., I, my, mine, me) and second-person pronouns (e.g., you, your, yours) in the students’ explanations (see Chafe, 1982; Redeker, 1984, for similar approaches). Twenty-two percent of the explanations were rated independently by two trained raters who had no knowledge of the experimental conditions. Interrater reliability was very good, ICC D.99 (Wirtz & Caspar, 2002). Thus, only one rater coded the rest of the explanations. Level of elaboration For the analysis of the elaboration of the explanations, we counted the number of elaborations per explanation. An elaboration was determined as a statement in which a student linked previous information of the hypertext to her or his prior knowledge, for instance, by including examples (e.g., “Motor scooters also have internal-combustion engines), by reporting one’s own experiences (e.g., “From my own experience, diesel engines normally consume considerably less fuel than gasoline”), or by making analogies (e.g., “the combustion can be regarded as the heart of the car”). Again, twenty-two percent of the explanations were rated independently by two trained raters who had no knowledge of the experimental conditions. Interrater reliability was good, ICC D.79 (Wirtz & Caspar, 2002). Thus, only one rater coded the rest of the explanations. Results The analyses of students’ conceptual knowledge were based on the full sample of 48 students. Unfortunately, two students’ explanations in the oral explanation condition and one student’s explanation in the written explanation condition were not stored in our database due to technical reasons. Therefore, we removed these three students for the analyses of our mediation hypotheses. Thus, the mediation analyses were based on 45 students. We used an alpha level of.05 for all statistical analyses. As effect-size measure, we used Cohen’s d qualifying values between 0.20 and 0.50 as small effect, between 0.50 and 0.80 as medium effect, and values greater than 0.80 as large effect (see Cohen, 1988). As an effect size for the indirect effects in our mediation analyses, we used k2 as effect size measure, interpreting k2 <.01 as small effect, k2 between.09 and.25 as medium effect, and k2 >.25 as large effect (see Preacher & Kelley, 2011). Preliminary analyses A series of t tests and x2 tests revealed no signiﬁcant differences between the experimental conditions concerning age, t(46) D 1.00, p D.32; gender, x2(2) D 1.22, p D.54; number of semesters, t(46) D 0.63, p D.53; and prior knowledge, t(46) D 0.68, p D.50. Similarly, students’ perceived cognitive load during the explanation phase was comparable between experimental conditions, t(46) D 0.44, p D.66 (see Table 1). However, as writing commonly takes longer than speaking, the students’ explanations in the oral condition contained more words than the explanations in the writing conditions, t(23.95) D 6.58, THE JOURNAL OF EXPERIMENTAL EDUCATION 353 Table 1. Mean and standard deviations (in parentheses) for the dependent measures. Dependent variable Prior knowledgea Subjective cognitive loadb Features of the explanations Number of words Number of person-deictic referencesc Level of organizationd Number of elaborations Learning outcome Conceptual knowledgea Transfere Oral explanation condition Written explanation condition 4.79 (2.02) 2.91 (0.74) 4.42 (1.77) 2.81 (0.75) 707.61 (326.67) 5.57 (7.22) 1.80 (0.75) 4.52 (2.33) 249.68 (67.37) 0.55 (1.37) 2.68 (1.08) 1.95 (0.90) 8.33 (1.83) 11.06 (3.45) 8.92 (1.14) 9.15 (2.13) a Twelve points were possible on the prior knowledge test. The subjective cognitive load could vary between 1 and 5. c The number of person-deictic references was measured by the sum of occurrences of ﬁrst-person pronouns and second-person pronouns. d The level of organization of the explanations could vary between 0 (low organization) and 4 (high organization). e Twenty-six points were possible on the transfer test. b p <.001 (t test for unequal variances). Thus, in a next step, we examined whether our ﬁndings were confounded by the differences with regard to word length of the explanations. We, therefore, computed correlations between the students’ learning outcomes (conceptual knowledge, transfer) and the word length of students’ explanations. The number of words of students’ explanations was neither related to students’ conceptual knowledge, r D.07, p D.67; nor to students’ transfer, r D ¡.02, p D.89. Thus, we can conclude that students’ learning outcomes were not affected by the mere number of words of students’ explanations. Learning outcome hypotheses To test our conceptual knowledge hypothesis, whether students in the written explanation condition outperformed students in the oral explanation condition with regard to the acquisition of conceptual knowledge, we computed an independent t test with the students’ posttest scores as dependent variable and experimental condition (oral explanation versus written explanation) as independent variable. In contrast to our assumptions, however, there were no signiﬁcant differences among experimental conditions, t(38.45) D ¡1.32, p D.19, d D 0.39 (t test for unequal variances), indicating that writing explanations was comparably effective at as providing oral explanations. To test our transfer hypothesis, we similarly computed an independent t test with the students’ transfer scores as dependent variable and experimental condition (oral explanation versus written explanation) as independent variable. In line with our transfer hypothesis, we found that students who provided an oral explanation outperformed students who produced a written explanation on the transfer test, t(38.38) D 2.32, p D.03, d D 0.67 (t test for unequal variances), indicating that providing oral explanations better helped students in gaining transferable knowledge than writing explanations (see Table 1). Mediation hypotheses Next, we tested our mediation hypotheses. For that purpose, we ﬁrst checked whether the oral explanations and written explanations differed with regard to the level of organization and social involvement (i.e., use of person-deictic references, number of elaborations), as an implementation check (see Table 1 for the descriptive statistics). Thus, we performed independent t tests with explanation modality as independent variable, and the level of organization, the number of person-deictic references, and the number of elaborations as dependent variables. As expected, students in the oral explanation condition generated less organized 354 A. LACHNER ET AL. Table 2. Correlations between the linguistic features of the explanations. Dependent variables 1. Level of organization 2. Number of elaborations 3. Number of person-deictic references 1 2 ¡.19 ¡.17.49* 3 p <.001. explanations, t(37.37) D ¡3.16, p D.003, d D 0.97 (t test for unequal variances), compared with students in the written explanation condition (see Table 1). At the same time, students in the oral explanation included more elaborations in their explanations, t(29.20) D 5.10, p <.001, d D 1.53 (t test for unequal variances), and more person-deictic references, t(23.65) D 3.27, p D.003, d D 0.98 (t test for unequal variances) compared with students in the written explanation condition (see Table 1). Additionally, we looked for relevant correlations among the explanatory features (see Table 2). We found a signiﬁcant correlation between the number of elaborations and the number of person-deictic references. This ﬁnding is in line with the assumption that elaborations and person-deictic references likely result due to students’ social involvement. There were no other signiﬁcant relations among the explanatory features (see Table 2). Next, we tested our mediation hypotheses. First, we tested whether generating written explanations would support students’ conceptual knowledge acquisitions more strongly than generating oral explanations because students who generated written explanations better organized their explanations compared with students who generated oral explanations (organization-conceptual-knowledge hypothesis). Second, we examined whether students in the oral explanation condition would better transfer their knowledge to other situations than students in the written explanation condition, due to students’ higher levels of social involvement, as indicated by the amount of person-deictic references and the number of elaborations in the students’ explanations (involvement-transfer hypothesis). To test our mediation hypotheses, we performed three mediation analyses by applying OLS regression-based path analyses. Type of explanation modality was a dummy-coded predictor (1 D oral, 2 D written). To test the indirect effects of our mediators, we applied the bootstrapping methodology by Hayes (2013) via the PROCESS macro for SPSS. Due to its reliance on resampling and replacement, the bootstrapping methodology has been shown to be a superior alternative to conventional mediation analyses, such as the Sobel test, as the bootstrapping methodology is less vulnerable to potential outliers, which potentially occur in smaller samples (Creedon & Hayes, 2015). Thus, the bootstrapping methodology can be regarded as a robust and valid analysis to test mediation effects for smaller samples. We used 10,000 bootstrap samples to derive a 95%-bias-corrected conﬁdence interval for the indirect effect. The main ﬁndings of our mediation analyses can be found in Figure 1. With regard to our organization-conceptual-knowledge hypothesis, as with previous analyses with regard to students’ conceptual knowledge, there was no total effect of type of explanation modality on students’ conceptual knowledge acquisition (see Figure 1). However, we found an unstandardized indirect effect of.48 (k2 D.17, medium effect), 95% CI [.14, 1.11]. As zero was not included in the conﬁdence interval, the indirect effect was signiﬁcant. Thus, our organization-conceptual-knowledge hypothesis was conﬁrmed: Generating written explanations better supported students in organizing Figure 1. Results of the mediation analyses. Numbers represent unstandardized path coefﬁcients. p <.05. THE JOURNAL OF EXPERIMENTAL EDUCATION 355 the information in their explanations than did oral explanations. The higher levels of organization also resulted in higher test scores in the conceptual-knowledge test. At the same time, however, writing triggered additional mediational processes that were detrimental to students’ conceptual learning, which resulted in a nonsigniﬁcant total effect of explanation modality on students’ conceptual learning (Hayes, 2013). With regard to our involvement-transfer hypothesis, there was a total effect of type of explanation modality on students’ transfer, B D ¡1.76, 95% CI [¡3.47, ¡.06], indicating that students in the oral explanation condition outperformed students in the written explanation condition on the transfer test. With regard to students’ use of person-deictic references (use of ﬁrst-person pronouns and secondperson pronouns in the explanations), there was no signiﬁcant indirect effect (B D 0.73, k2 D.13, medium effect, 95% CI [¡10, 1.74]), as zero was included in the conﬁdence interval. On the other hand, with regard to students’ amount of elaborations, we found a signiﬁcant unstandardized indirect effect of ¡1.44 (k2 D.21, medium effect), 95% CI [¡2.93, ¡.05], as zero was not included in the conﬁdence interval. These ﬁndings indicate that the higher levels of perceived social involvement in the oral explanation condition triggered students to use more person deictic references (ﬁrst-person pronouns and second-person pronouns) and more elaborations in their explanations. However, only the amount of students’ elaborations mediated the effect of explanation modality on students’ transfer. Discussion In the present study, we investigated whether learning by providing written or oral explanations differently affected the semantic features of students’ explanations and, subsequently, their learning outcomes. On the level of learning outcomes, our ﬁndings indicated that the modality of explaining (providing written versus oral explanations) had differential effects on students’ acquisition of knowledge. First, contrary to our assumptions, we found that the modality of explaining had no effects of students’ acquisition of conceptual knowledge, as students who generated a written explanation and students who produced an oral explanation comparably beneﬁted from explaining regardless of experimental condition. More importantly, with regard to students’ transfer, providing oral explanations more effectively helped the students to acquire transferable knowledge as compared with providing written explanations. However, it has to be acknowledged that the average transfer performance was rather low, which can be ascribed to the fact that we used rather challenging transfer tasks that asked the students to reinterpret the previously encountered information in order to transfer this information to the transfer tasks (Schwartz et al., 2005). Furthermore, our mediation analyses revealed that the semantic features of students’ explanations mediated the effect of type of explanation modality (written versus oral) on students’ learning outcomes. In line with our mediation hypotheses, ﬁrst, we found that generating written explanations, versus providing oral explanations, supported students more strongly in organizing the information in the explanations. The higher levels of organization in turn helped students acquire deeper levels of conceptual knowledge. At ﬁrst glance, this ﬁnding may be interpreted to be contrary to our nonsigniﬁcant ﬁndings on the total effects of the explanation modality on students’ conceptual knowledge. However, it has to be noted that is legitimate to accomplish mediation, even if the total effect is not signiﬁcant (see Hayes, 2013). Commonly, the total effect can be understood as the sum of the direct effect and all indirect effects. Thus, the absence of a total effect within our study suggests that although writing explanations helped students better organize their thoughts, there must have been additional cognitive processes that were detrimental for students’ conceptual learning and, as such, cancelled out the overall effect of writing explanations on students’ acquisition of conceptual knowledge. Therefore, it is up to further research to investigate which detrimental processes may occur during writing explanations that downsize students’ conceptual learning. With regard to students’ transfer performance, we found that oral explanations fostered students’ transferable knowledge, mainly because students made more elaborations during explaining. Apparently, students’ higher levels of social involvement while explaining orally triggered an increase in the 356 A. LACHNER ET AL. vividness of their explanations accomplished by providing additional illustrative examples, which in turn resulted in more ﬂexible and transferable knowledge. Together, our ﬁndings on students’ learning are particularly interesting because Hoogerheide et al. (2016) recently showed that providing oral explanations and providing written explanations were comparably effective, at least for acquiring direct procedural knowledge, but had no further effects on students’ transfer of such procedures. Apart from this obvious contradiction between the two studies, there is an important difference between the learning materials used in study by Hoogerheide et al. (2016) and in our current study. Whereas Hoogerheide et al. used rather low-element-interactivity learning material about syllogistic reasoning, we used more complex learning materials. Learning the functions and processes of internal combustion engines can be regarded as high-element interactivity material, as it requires students to thoroughly interconnect the presented elements of an internalcombustion engine in order to construct a coherently organized mental model of internal combustion engines. Hence, our ﬁndings suggest that generating oral explanations and generating written explanations have differential effects, predominantly for complex (i.e. high-element-interactivity) learning material. However, this assumption is highly speculative and should be tested in future studies that directly consider the complexity of the learning material. For instance, future research could compare effects of generating written versus oral explanations on students’ learning separately for high-complexity versus low-complexity learning material (see Chen, Kalyuga, & Sweller, 2015, for a similar study on worked examples). In respect to the explanatory features, our ﬁndings extend prior research by directly examining the mediating role of semantic features of explanations on students’ learning. Whereas recent studies (e.g., Fiorella & Mayer, 2014; Hoogerheide et al., 2016) mainly relied on analyzing effects of explaining on students’ learning outcomes, our ﬁndings add to the hitherto scarce evidence regarding the effects of knowledge-building activities on students’ learning (Roscoe & Chi, 2008). Our ﬁndings show that organizing and elaborating one’s thoughts during explaining, as reﬂected in the higher levels of organization and elaborations of the explanations, promoted students’ generative learning (Fiorella & Mayer, 2015). Hence, we can conclude that the effects of knowledgebuilding activities can also be generalized to rather low-interactive explanatory settings, such as providing explanations to ﬁctitious others. Study limitations and future research Despite the promising results, there are some limitations that need to be addressed. One limitation is owing to the fact that we analyzed only linguistic explanatory features as a proxy of the underlying cognitive processes of explaining on students’ learning outcomes. Therefore, we adapted a coding scheme by N€ uckles et al. (2009), who produced fruitful insights into potential cognitive mechanisms of explaining. However, it has to be noted that assessing the students’ products of explaining (i.e., the organization and elaboration of their explanations)—above all for the writing condition—does not provide an “online” measure of the cognitive processes underlying explaining orally versus in written form. For instance, it is possible that, particularly in the writing condition, students could have elicited more elaborations than they actually included in their explanations. This can be expected, as writing explanations typically involves more conscious decisions about including or withholding explanatory information than generating oral explanations (N€ uckles et al., 2009). Thus, future studies should include online measures, such as think-aloud protocols (Ericsson & Simon, 1993) or log-ﬁle data, to more directly assess the underlying cognitive mechanisms of explaining. Another limitation refers to the fact that our design did not allow us to disentangle the effects of students’ expectancy to explain orally versus in written form and the modality of explaining as we informed the students only that they would explain the subject matter to a fellow student (for similar approaches, see also Hoogerheide et al., 2016). Thus, some students could have expected by default that they would explain orally, whereas other students could have expected that they would provide a written explanation. These different explanation expectancies could have impacted how the students approached processing the hypertext while learning. THE JOURNAL OF EXPERIMENTAL EDUCATION 357 Therefore, future research could use an expanded study design in which the explanation modality, the explanation expectancy, and the actual act of explaining are treated as independent factors. That said, future research should use larger sample sizes to increase test power and to test the generalizability of our ﬁndings. In the same vein, we have to acknowledge that we did not assess students’ reading and writing skills prior to the study. Students’ reading abilities and their writing abilities could have largely impacted the effects of explanation modality on students’ learning. We, therefore, sought to carefully select a rather homogeneous sample of advanced educational science students who likely were comparable with regard to their reading and writing skills. Nevertheless, future research should include students’ reading and writing skills as potential control variables in their studies. A ﬁnal caveat relates to the fact that we used only one learning task (learning the functions of an internal-combustion engine), which may limit the generalizability of our ﬁndings. Although internalcombustion engines represent a common and relevant topic in engineering, further research should include different learning materials varying in complexity to map a more comprehensive picture of the effects of generating written versus oral explanations. That said, future research should also examine long-term effects of the explanation modality on students’ learning. Thus, future research should implement delayed tests to examine whether explaining might help not only to better understand the material but also to construct sustainable knowledge that can be used in future situations. Our ﬁndings suggest that explaining to (ﬁctitious) others can be regarded as a valuable scaffold for students’ learning. Due to the different affordances of written and oral discourse, they suggest that writing and speaking explanations might be appropriate if the primary learning goal is the acquisition of conceptual knowledge. However, if the main goal is to enable students to construct ﬂexible applicable knowledge, generating oral explanations might be the better alternative. Ethical statement The authors declare that they have no conﬂict of interest. All procedures performed in this study were in accordance with the 1964 Helsinki declaration, and the German Psychological Society’s (DGPS) ethical guidelines. According to the DGPS guidelines, experimental studies only need approval from an institutional review board if participants are exposed to risks that are related to high emotional or physical stress or when participants are not informed about the goals and procedures included in the study. As none of these conditions applied to the current study, we did not seek approval from an institutional review board. Informed consent was obtained from all individual participants included in the study. Acknowledgments We would like to thank Iris Backﬁsch for helping us with coding the data. We thank Wesley Dopkins for proofreading the manuscript. References Ainsworth, S., & Loizou, A. T. (2003). The effects of self-explaining when learning with text or diagrams. Cognitive Science, 27(4), 669–681. doi:10.1016/S0364-0213(03)00033-8 Akinnaso, F. N. (1985). On the similarities between spoken and written language. Language and Speech, 28(4), 323–359. doi:10.1177/002383098502800401 Bargh, J. A., & Schul, Y. (1980). On the cognitive beneﬁts of teaching. Journal of Educational Psychology, 72(5), 593. doi:10.1037/0022-0663.72.5.593 Berthold, K., & Renkl, A. (2009). Instructional aids to support a conceptual understanding of multiple representations. Journal of Educational Psychology, 101(1), 70–87. doi:10.1037/a0013247 Chafe, W. (1982). Integration and involvement in speaking, writing, and oral literature. In D. Tannen (Ed.), Spoken and written language: Exploring orality and literacy (pp. 35–53). Norwood, NJ: Ablex. Chen, O., Kalyuga, S., & Sweller, J. (2015). The worked example effect, the generation effect, and element interactivity. Journal of Educational Psychology, 107(3), 689–704. doi:10.1037/edu0000018 Chen, Y. C., Park, S., & Hand, B. (2016). Examining the use of talk and writing for students’ development of scientiﬁc conceptual knowledge through constructing and critiquing arguments. Cognition and Instruction, 34(2), 100–147. 358 A. LACHNER ET AL. Chi, M. T. H., Siler, S. A., Jeong, H., Yamauchi, T., & Hausmann, T. (2001). Learning from human tutoring. Cognitive Science, 25(4), 471–533. doi:10.1016/S0364-0213(01)00044-1 Chin, D. B., Dohmen, I. M., Cheng, B. H., Oppezzo, M. A., Chase, C. C., & Schwartz, D. L. (2010). Preparing students for future learning with teachable agents. Educational Technology Research and Development, 58(6), 649–669. doi:10.1007/s11423-010-9154-5 Cleland, A. A., & Pickering, M. J. (2006). Do writing and speaking employ the same syntactic representations? Journal of Memory and Language, 54(2), 185–198. doi:10.1016/j.jml.2005.10.003 Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum. Creedon, P. J., & Hayes, A. F. (2015, May). Small sample mediation analysis: How far can you push the bootstrap? Presented at the annual conference of the Association for Psychological Science, New York, NY. De Smet, C., Bourgonjon, J., De Wever, B., Schellens, T., & Valcke, M. (2012). Researching instructional use and the technology acceptation of learning management systems by secondary school teachers. Computers and Education, 58(2), 688–696. doi:10.1016/j.compedu.2011.09.013 De Smet, M. J. R., Brand-Gruwel, S., Leijten, M., & Kirschner, P. A. (2014). Electronic outlining as a writing strategy: Effects on students’ writing products, mental effort and writing process. Computers and Education, 78, 352–366. doi:10.1016/j.compedu.2014.06.010 Ericsson, K. A., & Simon, H. A. (1993). Protocol analysis: Verbal reports as data (Revised ed.). Cambridge, MA: MIT Press. Fiorella, L., & Mayer, R. E. (2013). The relative beneﬁts of learning by teaching and teaching expectancy. Contemporary Educational Psychology, 38(4), 281–288. doi:10.1016/j.cedpsych.2013.06.001 Fiorella, L., & Mayer, R. E. (2014). Role of expectations and explanations in learning by teaching. Contemporary Educational Psychology, 39(2), 75–85. doi:10.1016/j.cedpsych.2014.01.001 Fiorella, L., & Mayer, R. E. (2015). Eight ways to promote generative learning. Educational Psychology Review, 1–25. doi:10.1007/s10648-015-9348-9 Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York, NY: Guilford Press. Hoogerheide, V., Deijkers, L., Loyens, S. M., Heijltjes, A., & van Gog, T. (2016). Gaining from explaining: Learning improves from explaining to ﬁctitious others on video, not from writing to them. Contemporary Educational Psychology, 44, 95–106. doi:10.1016/j.cedpsych.2016.02.005 Hoogerheide, V., Loyens, S. M., & van Gog, T. (2014). Effects of creating video-based modeling examples on learning and transfer. Learning and Instruction, 33, 108–119. doi:10.1016/j.learninstruc.2014.04.005 Klein, P., Boscolo, P., Kirkpatrick, L., & Gelati, C. (Eds.). (2014). Writing as a learning activity. Leiden, Netherlands: Brill. Lachner, A., Burkhart, C., & N€ uckles, M. (2017). Mind the gap! Automated concept map feedback supports students in writing cohesive explanations. Journal of Experimental Psychology: Applied, 23(1), 29–46. doi:10.1037/xap0000111 Lakoff, R. T. (1982). Some of my favorite writers are literate: The mingling of oral and literate strategies in written communication. In D. Tannen (Ed.), Spoken and written language: Exploring orality and literacy (pp. 239–260). Norwood, NJ: Ablex. Margaryan, A., Bianco, M., & Littlejohn, A. (2015). Instructional quality of massive open online courses (MOOCs). Computers and Education, 80, 77–83. doi:10.1016/j.compedu.2014.08.005 Mayer, R. E. (2002). Rote versus meaningful learning. Theory Into Practice, 41(4), 226–232. doi:10.1207/ s15430421tip4104_4 McNamara, D. S., & Kintsch, W. (1996). Learning from text: Effects of prior knowledge and text coherence. Discourse Processes, 22(3), 247–287. doi:10.1080/01638539609544975 N€ uckles, M., H€ ubner, S., & Renkl, A. (2009). Enhancing self-regulated learning by writing learning protocols. Learning and Instruction, 19(3), 259–271. doi:10.1016/j.learninstruc.2008.05.002 N€ uckles, M., Wittwer, J., & Renkl, A. (2005). Information about a layperson’s knowledge supports experts in giving effective and efﬁcient online advice to laypersons. Journal of Experimental Psychology: Applied, 11(4), 219. doi:10.1037/ 1076-898X.11.4.219 Palincsar, A. S., & Brown, A. L. (1984). Reciprocal teaching of comprehension-fostering and comprehension-monitoring activities. Cognition and Instruction, 1(2), 117–175. doi:10.1207/s1532690xci0102_1 Parr, J. M., & Wilkinson, I. (2016). Widening the theoretical lens on talk and writing pedagogy. International Journal of Educational Research, 80, 217–225. doi:10.1016/j.ijer.2016.08.011 Pl€otzner, R., Dillenbourg, P., Preier, M., & Traum, D. (1999). Learning by explaining to oneself and to others. Collaborative learning: Cognitive and computational approaches, 1, 103–121. Preacher, K. J., & Kelley, K. (2011). Effect size measures for mediation models: Quantitative strategies for communicating indirect effects. Psychological Methods, 16(2), 93–115. doi:10.1037/a002265810.1037/a0022658.supp Redeker, G. (1984). On differences between spoken and written language. Discourse Processes, 7(1), 43–55. doi:10.1080/ 01638538409544580 Rivard, L. P. (2004). Are language based activities in science effective for all students, including low achievers? Science Education, 88(3), 420–442. doi:10.1002/sce.10114 Rivard, L. P., & Straw, S. B. (2000). The effect of talk and writing on learning science: An exploratory study. Science Education, 84(5), 566–593. doi:10.1002/1098-237X THE JOURNAL OF EXPERIMENTAL EDUCATION 359 Roscoe, R. D. (2014). Self-monitoring and knowledge-building in learning by teaching. Instructional Science, 42(3), 327– 351. doi:10.1007/s11251-013-9283-4 Roscoe, R. D., & Chi, M. T. (2008). Tutor learning: The role of explaining and responding to questions. Instructional Science, 36(4), 321–350. doi:10.1007/s11251-007-9034-5 Rummel, N., & Spada, H. (2005). Learning to collaborate: An instructional approach to promoting collaborative problem solving in computer-mediated settings. Journal of the Learning Sciences, 14(2), 201–241. doi:10.1207/ s15327809jls1402_2 Schwartz, D. L., Bransford, J. D., & Sears, D. L. (2005). Efﬁciency and innovation in transfer. In J. Mestre (Ed.), Transfer of learning from a modern multidisciplinary perspective (pp. 1–51). Greenwich, CT: Information Age. Sindoni, M. G. (2014). Spoken and written discourse in online interactions: A multimodal approach. New York, NY: Routledge. Sperling, M. (1996). Revisiting the writing-speaking connection: Challenges for research on writing and writing instruction. Review of Educational Research, 66(1), 53–86. doi:10.3102/00346543066001053 Wagner, J. F. (2010). A transfer-in-pieces consideration of the perception of structure in the transfer of learning. Journal of the Learning Sciences, 19(4), 443–479. doi:10.1080/10508406.2010.505138 Wirtz, M. A., & Caspar, F. (2002). Beurteiler€ ubereinstimmung und Beurteilerreliabilit€at: Methoden zur Bestimmung und Verbesserung der Zuverl€assigkeit von Einsch€atzungen mittels Kategoriensystemen und Ratingskalen [Interrater agreement und interrater reliability: Methods for calculating and improving the reliability of ratings by category systems and rating scales]. G€ ottingen, Germany: Hogrefe. Wylie, R., & Chi, M. T. (2014). The self-explanation principle in multimedia learning. The Cambridge handbook of multimedia learning (2nd ed., pp. 413–432). New York, NY: Cambridge University Press. Appendix A Multiple-choice items used in the pretest and in the conceptual-knowledge test 1 [x] 2 [x] 3 [x] 4 [x] 5 [x] 6 [x] 7 [x] 8 What is the correct process of a four-stroke engine? Induction, compression, power, exhaust Compression, induction, exhaust, power, Power, induction, compression, exhaust Induction, exhaust, power, compression What is the cubic capacity? The volume between inlet valve and outlet valve The volume between top dead center and stock The volume between top dead center and down dead center The volume between plug and stock Which feature of an engine determines its performance? Inlet valve Engine speed Conrod length Cubic capacity What is no element of a gasoline engine? Plug Stock Injection system Combustion chamber What is a necessary element of a diesel engine? An injection plug A hot environment in the cylinder A vacuum in the cylinder A spark for ignition Which are the underlying processes of an Otto engine? Two strokes and four strokes Three strokes and ﬁve strokes Three strokes and six strokes Four strokes and six strokes When does the inlet valve and outlet valve open? Inlet valve opens during the third stroke; outlet valve opens during the fourth stroke Inlet valve opens during the ﬁrst stroke; outlet valve opens during the fourth stroke Inlet valve opens during the third stroke; outlet valve opens during the third stroke Inlet valve opens during the second stroke; outlet valve opens during the fourth stroke What are the reasons for opening and closing inlet valves and outlet valves? (Continued on next page) 360 A. LACHNER ET AL. Appendix A (Continued). [x] 9 [x] 10 [x] 11 [x] 12 [x] To start the control valve To generate a compression in the cylinder To enable a processing of cooling water To bleed the stroke How does an ignition work in a diesel motor? Cold air reacts with the fuel Hot air reacts with the fuel The spark ignites the fuel The fuel ignites due to the cylinder friction Which function does the conrod have? Transfer of energy of the piston to the spark Control of the inlet valve Transfer of energy of the piston to the crank Control of the outlet valve What are the advantages of an Otto engine compared to a diesel engine? Higher performance Lower weight Higher efﬁciency Lower levels of wear Where does the energy of gas exhaustion come from? From the oscillating weight of the crank From one cylinder in the induction phase From the ignition of the gas From the control valve Note. Correct responses are marked by an x. Appendix B Open transfer questions used in the transfer test Statement Points 1. Can the processes of a four-stroke engine be transferred to a two-stroke engine? Please explain! First stroke of a two-stroke engine corresponds to ﬁrst stroke of a four-stroke engine: Compress Piston moves from lower dead end to the upper dead end Gasoline is compressed First stroke of a two-stroke engine corresponds to second stroke of a four-stroke engine: Gasoline is supplied Due to increased temperature, compression is increasing Second stroke of a two-stroke engine corresponds to third stroke of a four-stroke engine: Power Gas is ignited Piston produces mechanical energy due to movement from upper dead end to the lower dead end Second stroke of a two-stroke engine corresponds to fourth stroke of a four-stroke engine: Exhaust Piston closes intake duct and compresses fresh gas Gas exhausts Due to exhaustion new gas is supplied 1 1 1 1 1 1 1 1 1 1 1 1 1 2. Can the processes of a four-stroke engine be transferred to a gas turbine? Please explain! Compressor corresponds to ﬁrst stroke of a four-stroke engine Air streams in the compressor Air is compressed Compressed air is transferred to the combustion chamber Combustion chamber corresponds to second stroke of a four-stroke engine Fuel is supplied 1 1 1 1 1 1 (Continued on next page) THE JOURNAL OF EXPERIMENTAL EDUCATION Statement 361 Points Fuel-air mixture is ignited Due to increased temperature, compression is increasing Turbine corresponds to third stroke of a four-stroke engine Turbine transfers chemical energy into mechanical energy Exhaust corresponds to fourth stroke of a four-stroke engine Mechanical energy produces shear force Gas exhausts 1 1 1 1 1 1 1 Note. Participants received 1 point if they mentioned the particular statement in their answer. A maximum of 13 points could be obtained per answer. Appendix C Translated instructions used in the experiment Instruction in the introductory phase In this study, we examine the inﬂuence of explaining on learning. For this purpose, you will receive a web-text about combustion engines. After the learning phase you will explain the previously learned subject matter to a potential fellow student who could not manage to attend the study. Instruction for the learning phase Please carefully read the following web-text about combustion engines. You are allowed to take notes on a separate sheet. Your notes will be available only during the learning phase and the explaining phase and will be removed for the posttest phase. Instruction for the planning phase (oral condition) Dear participant As a fellow student could not attend the study, but is very interested in the principles of combustion engines, you are asked to provide an oral explanation about combustion engines for her/him. Please outline how you would structure your explanation in bullet points. Instruction for the planning phase (writing condition) Dear participant As a fellow student could not attend the study, but is very interested in the principles of combustion engines, you are asked to provide a written explanation about combustion engines for her or him. Please outline how you would structure your explanation in bullet points. Instruction for the explanation phase (oral condition) Now you are asked to provide your explanation to your fellow student. Take care that your explanation is understandable for the fellow student without any further learning material. That said, you should include the main components and functions of combustion engines in your explanation. When you are ready to speak your explanation, sign to start to the experimenter to start the audio-recording device. Instruction for the explanation phase (writing condition) Now you are asked to provide your explanation to your fellow student. Take care that your explanation is understandable for the fellow student without any further learning material. You should include the main components and functions of combustion engines in your explanation. When you are ready to write your explanation, sign to start to the experimenter to start the word processing editor.

Providing Written or Oral Explanations? PDF

Document Details

Tags

Related

Summary

Full Transcript

Upgrade to continue