2009年4月，墨西哥與美國西南地區出現人類感染A／H1N1新型流感病毒，此一新興傳染疾病在短時間內，擴散至全球各地。因媒體是社會大眾獲得健康風險訊息的主要來源之一，媒體呈現的H1N1新型流感樣貌，可能影響閱聽眾對此疾病的認知、態度，甚或行為。故本文研究目的有二，一為探討報紙媒體如何呈現H1N1新型流感新聞，分析新聞版面、性質、篇幅、主題、消息來源與圖表運用，二為檢視H1N1新型流感新聞內容是否聚焦風險不確定的面向，呈現公眾憤怨恐慌的要素，以及是否提供民眾評估風險的資訊。了解台灣報紙媒體如何呈現新流感相關議題，為國內媒體呈現傳染疾病的研究新增不同類型的個案。 本研究蒐集《中國時報》、《聯合報》、《自由時報》和《蘋果日報》，自2009年4月25日至2010年2月25日止，共1,324則H1N1新型流感新聞為樣本，採用內容分析法檢視四報H1N1新型流感新聞的特性與呈現的風險訊息。並依H1N1新型流感疫情發展分為五個階段，比較各階段報導呈現風險訊息之異同。 研究結果發現，主題分布依疫情階段而有顯著差異，雖「疫情現況」為各時期H1N1新型流感報導最常呈現的報導主軸，但各疫情階段報導的主題比例仍有不同，尤其疫情趨緩期的主題比重與其他時期差異較大。在主要消息來源方面，各疫情階段皆以「衛生單位官員」為最多，尤其至疫情趨緩期時，主要消息來源偏向更為明顯。 在報導呈現公眾憤怨恐慌的要素方面，以各時期呈現憤怨恐慌的要素總比例比較之，疫情升溫期呈現憤怨恐慌要素的比例為最多，至疫情高峰期與疫情趨緩期，呈現憤怨恐慌的要素之總比例則明顯減少。四報呈現憤怨恐慌的要素皆以「憤怨恐慌的修辭」為最多，高達近五成新聞出現「憤怨恐慌的修辭」要素。 在報導呈現風險解釋方面，以各疫情階段呈現風險解釋的總比例相比，疫情入台傳散期，報導呈現各項風險解釋的比例為最高，其次為疫情初始期，再次為疫情升溫期。四報呈現的風險解釋以「過程解釋」類訊息為最多，其次為「實用建議」類的訊息，而「定義解釋」、「風險程度」、「風險比較」類訊息則較少被提及。 報導圖表文字內容呈現方面，則發現除《蘋果日報》外，其他三報運用圖表的比重偏低，顯示H1N1新型流感報導運用圖表解釋風險仍不甚普遍。但包含圖表的報導，圖表文字多為「風險解釋」的訊息，常提及「過程解釋」與「實用建議」類資訊。研究也發現，新聞正文與圖表文字呈現相對應的風險解釋類型有顯著相關性，其中以新聞正文與圖表文字呈現「風險比較」的相關程度最高。 透過本研究分析發現，四報有部分報導凸顯風險不確定的面向，憤怨恐慌要素以「恐懼恐慌」、「疫情失控淪陷」、「衝突或矛盾」比例較多；在風險解釋呈現方面，則以「疫情數據資料更新」、「目前疫情傳佈範圍／地區」類的訊息為最多，「風險程度」、「定義」與「風險比較」等訊息則較少被提及。 假定報導提供風險解釋，可減少公眾因對風險的「不確定性」產生的憤怨恐慌，本研究發現提醒媒體實務工作者，記者在危機或風險事件發生，擔負風險溝通的責任之際，應提供閱聽眾評估風險的資訊，以清晰明確的文字或圖表解釋風險過程與民眾自我保護的等訊息，減輕民眾的憤怨不安或恐懼。 / In April 2009, Mexico and the Southwestern United States confirmed human cases of Influenza A/H1N1. In a short time, H1N1 influenza was spreading across the rest of the world. People obtain health risk messages mainly through mass media, so what the media cover about H1N1 influenza may do have influences on their audience’s cognition, attitudes, and even behaviors toward this disease. This study has two research purposes. First, it studies how the newspaper media cover H1N1 influenza by analyzing location of article, article type, news length, content about theme, news source, and usage of graphics. Second, it examines the coverage of H1N1 influenza to see if the media focus the facets of risk and uncertainty, present public outrage factors, and offer people the information to evaluate the risks. From this study, we can understand what related issues of H1N1 influenza were presented by four major newspapers in Taiwan, and to the study of coverage of domestic media on infectious diseases this adds a different case. It sampled 1324 H1N1 influenza news reports from Chinese Times, United Daily News, The Liberty Times, and Apple Daily, covered from April 25, 2009 to February 25, 2010. It adopted content analysis to examine these four newspapers about the characteristics of their H1N1 influenza coverage and their risk communication factors. And according to the development of H1N1 influenza in Taiwan, we divided the period into five phases, comparing the similarities and differences of news content among various phases. Results of content analysis showed that the distribution of news theme varies significantly among the phases, though “epidemiological description” is the main theme. There were differences in news theme proportion among the five phases, especially notable difference of the decreasing phase. And that main news source is health officials in all phases. Source bias was more obvious especially during the decreasing phase. In terms of the coverage of public outrage factors, the news content of the four newspapers showed that it has the most outrage factors during H1N1 influenza increasing phase. In the peak and the decline of H1N1influenza contagion, news content of outrage factors decreased significantly. The most presented outrage factors by the four newspapers is the “outrage rhetoric.” Nearly fifty percent of the coverage has “outrage rhetoric” factor. Regarding the coverage of risk explanations in news content, compared by phases, the spreading phase has the highest percentage of risk explanations of H1N1 influenza news, followed by the initial phase and then the epidemic increasing phase. Among the risk explanations presented by the four newspaper, the news group of “process explanations” was found the most, followed by the news categorized of “ practical recommendation,” whereas the news groups of “definitions”, ”risk level explanations” and ”risk comparisons” were seldom mentioned. On the news with graphics, except for Apply Daily, three other newspapers have low usage of graphics. It appears that for H1N1 influenza news it is not common yet to use graphics to explain risks. For those reports with graphs, they are commonly for “risk explanations” and they often mention of messages of “process explanations” and “practical recommendation.” In addition, this study finds that there is significant correlation between articles and graphs when they both present same risk explanations in the news content. To sum up, through this study we found some reports of the four newspapers highlight the risk uncertainty, among which outrage factors are presented with higher exposure of “fear/ panic”, “epidemic out of control”, and “conflicting reports.” On the coverage of 'risk explanations,” it has the most messages about “update the data of H1N1 influenza” and “current epidemic category and area,” whereas it has less messages about “definitions”,” risk level explanations” and ”risk comparisons”. This study used an exploratory framework that posits risk explanations can mitigate negative public reactions by communicating the hazard with clearer articles or graphics. Result of this study recommends when a crisis or risk event happens, media worker should take responsibility of risk communication, and provide accurate and clear risk information to the audience. Thus, it may alleviate the negative public reactions, like outrage, anger, worry, fear, and panic.

This essay examines the recent electoral success of the Sweden Democrats (SD) in the Swedish municipal election in 2006 and 2010. By using statistical methods it aims to explain which of three contradicting theoretical frameworks best can explain how a populist radical right party could penetrate one of the most stable party systems in the world. The theoretical approaches tested in this essay are: a demand-side, an external supply-side and an internal supply-side approach. By using theoretically anchored proxies to determine the effect of the contradicting theoretical approaches this essay concludes that the internal supply-side explanation measuring the local party organizational ability of the SD had the most substantial effect when it comes to explaining their recent electoral success in the Swedish municipalities, as opposed to a more commonly believed demand-side driven explanation.

Syftet med denna studie var att undersöka hur artiklar från två välkända svenskakvällstidningar beskriver det våld som män brukar mot kvinnor i nära relationer. Vidare har det undersökts vilka orsaksförklaringar tidningsartiklar använder för att beskriva mäns våld mot kvinnor. En kvalitativ textanalys har genomförts där resultatet har tematiserats med teoretisk ansats genom socioekologiska modellen och Socialstyrelsens definition av våld i nära relationer. Media har visat sig ha en betydande roll för människors förståelse av sociala problem. Det som beskrivs i media blir ofta människors verklighet, därför är det av betydelse att media ger en rättvis och korrekt bild av det stora och utbredda fenomen som våld i nära relationer utgör. Resultatet indikerar att fysiskt våld är den vanligast förekommande typ av våld som beskrivs av tidningsartiklar. Vidare skildrar studien att flest tidningsartiklar beskriver orsaker till mäns våld mot kvinnor med ett strukturellt perspektiv. Endast en av alla inkluderade tidningsartiklar beskriver mäns våld mot kvinnor i nära relationer genom den socioekologiska modellens alla fyra perspektiv. Modellen som i enlighet med Socialstyrelsen är av stor vikt för att få en helhetsbeskrivning av mäns våld mot kvinnor i nära relationer. / The aim of this study was to investigate how articles from two well-known Swedish evening newspapers describe the violence that men use against women in close relationships. Furthermore, the study examined which causal explanations the newspaper articles use in the descriptions of the violence. This is done through a qualitative text analysis of newspaper articles' descriptions of violence. The results of the study have been thematized with a theoretical approach through the socio-ecological model and the National Board of Health and Welfare's definition of violence in close relationships. The media has proven to play a significant role in helping people understand social issues as what is described in the media often becomes people's reality. Therefore, it is important that the media provide a fair and correct picture of the large and widespread phenomenon that violence in close relationships constitutes. The results show that physical violence is the most described violence of newspaper articles. Furthermore, most causes of men's violence against women are described by newspaper articles with a structural perspective. Only one of all included newspaper articles describes men's violence against women in close relationships through all four perspectives of the socioecological model. The model which, in accordance with the National Board of Health and Welfare, is of great importance for obtaining a comprehensive description of men's violence against women in close relationships.

The riveting interplay of industrialization, informalization, and exponential technological growth of recent years has shifted the attention from classical machine learning techniques to more sophisticated deep learning approaches; yet its intrinsic black-box nature has been impeding its widespread adoption in transparency-critical operations. In this rapidly evolving landscape, where the symbiotic relationship between research and practical applications has never been more interwoven, the contribution of this paper is twofold: advancing gradient faithfulness of CAM methods and exploring new frontiers beyond it. In the first part, we theorize three novel gradient-based CAM formulations, aimed at replacing and superseding traditional Grad-CAM-based methods by tackling and addressing the intricately and persistent vanishing and saturating gradient problems. As a consequence, our work introduces novel enhancements to Grad-CAM that reshape the conventional gradient computation by incorporating a customized and adapted technique inspired by the well-established and provably Expected Gradients’ difference-from-reference approach. Our proposed techniques– Expected Grad-CAM, Expected Grad-CAM++and Guided Expected Grad-CAM– as they operate directly on the gradient computation, rather than the recombination of the weighing factors, are designed as a direct and seamless replacement for Grad-CAM and any posterior work built upon it. In the second part, we build on our prior proposition and devise a novel CAM method that produces both high-resolution and class-discriminative explanation without fusing other methods, while addressing the issues of both gradient and CAM methods altogether. Our last and most advanced proposition, Hyper Expected Grad-CAM, challenges the current state and formulation of visual explanation and faithfulness and produces a new type of hybrid saliencies that satisfy the notion of natural encoding and perceived resolution. By rethinking faithfulness and resolution is possible to generate saliencies which are more detailed, localized, and less noisy, but most importantly that are composed of only concepts that are encoded by the layerwise models’ understanding. Both contributions have been quantitatively and qualitatively compared and assessed in a 5 to 10 times larger evaluation study on the ILSVRC2012 dataset against nine of the most recent and performing CAM techniques across six metrics. Expected Grad-CAM outperformed not only the original formulation but also more advanced methods, resulting in the second-best explainer with an Ins-Del score of 0.56. Hyper Expected Grad-CAM provided remarkable results across each quantitative metric, yielding a 0.15 increase in insertion when compared to the highest-scoring explainer PolyCAM, totaling to an Ins-Del score of 0.72.

Spinal cord injury, whether traumatic or nontraumatic, can partially or completely damage sensorimotor pathways between the brain and the body, leading to heterogeneous gait abnormalities. Mobility impairments also depend on other factors such as age, weight, time since injury, pain, and walking aids used. The ASIA Impairment Scale is recommended to classify injury severity, but is not designed to characterize individual ambulatory capacity. Other standardized tests based on subjective or timing/distance assessments also have only limited ability to determine an individual's capacity. Data-driven techniques have demonstrated effectiveness in analysing complexity in many domains and may provide additional perspectives on the complexity of gait performance in persons with spinal cord injury. The studies in this thesis aimed to address the complexity of gait and functional abilities after spinal cord injury using data-driven approaches. The aim of the first manuscript was to characterize the heterogeneous gait patterns in persons with incomplete spinal cord injury. Dissimilarities among gait patterns in the study population were quantified with multivariate dynamic time warping. Gait patterns were classified into six distinct clusters using hierarchical agglomerative clustering. Through random forest classifiers with explainable AI, peak ankle plantarflexion during swing was identified as the feature that most often distinguished most clusters from the controls. By combining clinical evaluation with the proposed methods, it was possible to provide comprehensive analyses of the six gait clusters.     The aim of the second manuscript was to quantify sensorimotor effects on walking performance in persons with spinal cord injury. The relationships between 11 input features and 2 walking outcome measures - distance walked in 6 minutes and net energy cost of transport - were captured using 2 Gaussian process regression models. Explainable AI revealed the importance of muscle strength on both outcome measures. Use of walking aids also influenced distance walked, and  cardiovascular capacity influenced energy cost. Analyses for each person also gave useful insights into individual performance.     The findings from these studies demonstrate the large potential of advanced machine learning and explainable AI to address the complexity of gait function in persons with spinal cord injury. / Skador på ryggmärgen, oavsett om de är traumatiska eller icke-traumatiska, kan helt eller delvis skada sensoriska och motoriska banor mellan hjärnan och kroppen, vilket påverkar gången i varierande grad. Rörelsenedsättningen beror också på andra faktorer såsom ålder, vikt, tid sedan skadan uppstod, smärta och gånghjälpmedel. ASIA-skalan används för att klassificera ryggmärgsskadans svårighetsgrad, men är inte utformad för att karaktärisera individens gångförmåga. Andra standardiserade tester baserade på subjektiva eller tids och avståndsbedömningar har också begränsad möjlighet att beskriva individuell kapacitet. Datadrivna metoder är kraftfulla och kan ge ytterligare perspektiv på gångens komplexitet och prestation. Studierna i denna avhandling syftar till att analysera komplexa relationer mellan gång, motoriska samt sensoriska funktion efter ryggmärgsskada med hjälp av datadrivna metoder. Syftet med den första studien är att karaktärisera de heterogena gångmönster hos personer med inkomplett ryggmärgsskada. Multivariat dynamisk tidsförvrägning (eng: Multivariate dynamic time warping) användes för att kvantifiera gångskillnader i studiepopulationen. Hierarkisk agglomerativ klusteranalys (eng: hierarchical agglomerative clustering) delade upp gång i sex distinkta kluster, varav fyra hade lägre hastighet än kontroller. Med hjälp av förklarbara AI (eng: explainable AI) identifierades det att fotledsvinkeln i svingfasen hade störst påverkan om vilken kluster som gångmönstret hamnat i. Genom att kombinera klinisk undersökning med datadrivna metoder kunde vi beskriva en omfattande bild av de sex gångklustren. Syftet med den andra manuskriptet är att kvantifiera sensoriska och motoriska faktorerans påverkan på gångförmåga efter ryggmärgsskada. Med hjälp av två Gaussian process-regressionsmodeller identiferades sambanden mellan 11 beskrivande faktorer och 2 gång prestationsmått, nämligen gångavstånd på 6 minuter samt metabola energiåtgång. Med hjälp av förklarbar AI påvisades det stora påverkan av muskelstyrka på både gångsträckan och energiåtgång. Gånghjälpmedlet samt kardiovaskulär kapaciteten hade också betydande påverkan på gångprestation. Enskilda analyser gav insiktsfull information om varje individ. Resultaten från dessa studier visar på potentiella tillämpningar av avancerad maskininlärning och AI metoder för att analysera komplexa relationer mellan funktion och motorisk prestation efter ryggmärgsskada. / <p>QC 20240221</p>

<p dir="ltr">Tree-based machine learning models, such as decision trees and random forests, are one of the most widely used machine learning models primarily because of their predictive power in supervised learning tasks and ease of interpretation. Despite their popularity and power, these models have been found to produce unexpected or discriminatory behavior. Given their overwhelming success for most tasks, it is of interest to identify root causes of the unexpected and discriminatory behavior of tree-based models. However, there has not been much work on understanding and debugging tree-based classifiers in the context of fairness. We introduce FairDebugger, a system that utilizes recent advances in machine unlearning research to determine training data subsets responsible for model unfairness. Given a tree-based model learned on a training dataset, FairDebugger identifies the top-k training data subsets responsible for model unfairness, or bias, by measuring the change in model parameters when parts of the underlying training data are removed. We describe the architecture of FairDebugger and walk through real-world use cases to demonstrate how FairDebugger detects these patterns and their explanations.</p>

Trust is important for the adoption of autonomous vehicles. Providing voiced explanations explaining a vehicle’s behavior has been found to improve trust, but it is unclear how explanations should be presented. In this experimental study, it was investigated whether manipulating the temporal sequence of “How” explanations, that describe what the vehicle does, and “Why” explanations, that describe why the vehicle does something, influence users’ trust, mental workload, situational awareness, and preferences. The research questions are: 1. What is the optimal sequence of presenting explanations, measured by the least amount of mental workload, increased trust, and increased situational awareness? 2. Which explanation type do people prefer the most? 3. Does explanation type affect attentional task performance? These questions were assessed by using questionnaires and qualitative data in a within group design. The sample consisted of 42 participants between the ages of 21 and 37. The data was analyzed with a repeated measures analysis of variance (ANOVA) and a binomial test. The study found no effect of different explanation types on trust and mental workload. However, the scenarios had a significant effect on the results. Presenting the “How + Why” explanation before an event caused the best situational awareness. Although no significant difference in preferences was found, the “Why” explanation had the highest mean preference. This study shows that when evaluating different explanation types, it is important to consider situational factors that can affect measured variables such as vehicle’s behavior and surrounding environment. Future studies should focus on exploring situational factors further. / Förtroende är viktigt för införandet av autonoma fordon. Att ge auditiva förklaringar om bilens handlingar har setts förbättra förtroendet, men det är oklart hur förklaringar bör presenteras. I denna experimentella studie undersöktes det om manipulering av den tidsmässiga sekvensen av "Hur"-förklaringar, som beskriver vad fordonet gör, och "Varför"-förklaringar, som beskriver varför fordonet gör något, påverkade användarens förtroende, mentala arbetsbelastning, situationsmedvetenhet, och preferenser. Forskningsfrågorna var: 1. Vilken är den optimala sekvensen för att presentera förklaringar, mätt med minsta möjliga mentala arbetsbelastning, ökat förtroende och ökad situationsmedvetenhet? 2. Vilken typ av förklaring föredrar folk mest? 3. Påverkar typ av förklaring uppmärksamhetsuppgift? Dessa frågeställningar undersöktes med hjälp av frågeformulär och kvalitativa data i en inomgruppsdesign. Urvalet bestod av 42 deltagare mellan 21–37 år. Data analyserades med en variansanalys med upprepade mått (ANOVA) och ett binomialtest. Studien fann ingen effekt av olika förklaringstyper på förtroende och mental arbetsbelastning. Scenarierna hade dock signifikant effekt på resultaten. Att presentera "Hur + Varför"-förklaringen innan en händelse gav den bästa situationsmedvetenheten. Även om ingen signifikant skillnad i preferens hittades, hade "Varför"-förklaringen det högsta medelvärdet. Denna studie visade att när man utvärderar olika förklaringstyper är det viktigt att ta hänsyn till situationsfaktorer som kan påverka uppmätta variabler som fordonets beteende och omgivande miljö. Framtida studier bör fokusera på att utforska situationsfaktorer ytterligare.

<p dir="ltr">Anomaly detection is an active research field that widely applies to commercial applications to detect unusual patterns or outliers. Time series anomaly detection provides valuable insights into mission and safety-critical applications using ever-growing temporal data, including continuous streaming time series data from the Internet of Things (IoT), sensor networks, healthcare, stock prices, computer metrics, and application monitoring. While Generative Adversarial Networks (GANs) demonstrate promising results in time series anomaly detection, the opaque nature of generative deep learning models lacks explainability and hinders broader adoption. Understanding the rationale behind model predictions and providing human-interpretable explanations are vital for increasing confidence and trust in machine learning (ML) frameworks such as GANs. This study conducted a structured and comprehensive assessment of post-hoc local explainability in GAN-based time series anomaly detection using SHapley Additive exPlanations (SHAP). Using publicly available benchmarking datasets approved by Purdue’s Institutional Review Board (IRB), this study evaluated state-of-the-art GAN frameworks identifying their advantages and limitations for time series anomaly detection. This study demonstrated a systematic approach in quantifying the extent of GAN-based time series anomaly explainability, providing insights for businesses when considering adopting generative deep learning models. The presented results show that GANs capture complex time series temporal distribution and are applicable for anomaly detection. The analysis from this study shows SHAP can identify the significance of contributing features within time series data and derive post-hoc explanations to quantify GAN-detected time series anomalies.</p>

Includes bibliographical references (leaves 106-115) / The purpose of this study is to critically analyse selected scholarly views on the visionary experiences during Jesus’ baptism. Modern scholars have different opinions about the reports in the New Testament on Jesus’ baptismal visionary phenomena. Some scholars interpreted the events as Jesus’ actual seeing or vision and others accept it as literary creations by the authors, to make sense of the reports on seeing and hearing that are hard to understand. Reports like a Spirit descending in the form of a dove are extraordinary for most people and pose interpretive challenges. The two distinct trends identified in the study of visionary experiences are those who take the text on the visions literally and those who see them as literary creations. There is a new trend in biblical scholarship, which is comparative and invokes insights from cross-cultural research in order to understand the accounts of the visions as altered states of consciousness (ASC). These views are also presented, compared and evaluated selecting three major views from modern New Testament scholars. Among the scholars identified, who take the baptism visions literally, are Dunn, Meier, Marcus, Hurtado, Borg and Webb. The scholars selected, who consider the baptism visions as literary creations, are Sanders, Crossan, Miller and Strijdom. The scholars, who contributed to the new development in interpreting the texts on visionary experiences as ASC, are Pilch, Davies and DeMaris. Pilch uses the theoretical model of ASC and understands it differently from those used by Davies and DeMaris on which they base their interpretations. A scientific explanation of ASC is built from theories about how the brain and culture, together, create certain states of consciousness. All these views are analysed based on the scholarly interpretations from the three definitive trends in the visions research, comparing the caretaker versus critical at a meta-analysis level. / Biblical and Ancient Studies / M.A. (Biblical archaeology)

Forskning har visat på att etablerandet av en dialogisk klassrumskultur ger rikare lärandetillfällen för eleverna eftersom de får rikare möjlighet att utbyta tankar och uppfattningar. Studien är en kvalitativ intervjustudie av hur lärare, verksamma i förskoleklass samt årskurserna 1–3, arbetar för att skapa dialog i klassrummet. Syftet med intervjuanalysen är att belysa olika strategier lärare använder för att främja en samtalande klassrumskultur. Datan från intervjuerna kodades och analyserades i relation till tidigare forskning. Forskningen som presenteras i studien grundas i den konstruktivistiska traditionen och lägger mycket tyngd på normskapande, specifikt sociomatematiska normer som är knutna till matematiskt lärande. Forskningen, som till stor del består av gedigna forskningsöversikter, lyfter att förmågor som att resonera och argumentera endast utvecklas i en social kontext och att lärares handlingar och strukturering av undervisning därför har stor betydelse. Studiens resultat ger en bild av vilka förväntningar de intervjuade lärarna har på sina elever, det vill säga vilka sociala normer som råder. Dessa är: vara aktivt deltagande, att verbalisera samt interagera. Resultatet visar också på de specifika sociomatematiska normer som att: använda korrekta matematiska termer, använda sig av olika matematiska representationer samt att ge godtagbara matematiska förklaringar. Analysen visar att lärarna använder sig av ett stort spann av strategier för att etablera normer för ett samtalande klassrumsklimat. Exempel på strategier för att främja ett samtalande klassrumsklimat är att etablera fasta samtals-strukturer och svarssituationer som eleverna är införstådda med, positiv förstärkning och respons, att låta eleverna lyckas med sitt aktiva deltagande, modellering och revoicing. Syftet är att skapa mönster och normer för matematisk kommunikation. Genom lärarnas strategiska val av uppgifter främjas deltagande, interaktion och kommunikations-mönster, det vill säga att eleverna får en normativ förståelse för att ”alla ska vara med”, ”hur vi talar om matematik” och ”vad vi ska tala om” i genuint matematiska samtal. Lärarna upplevs ha en ansats att samtalande ska bli utforskande till sin karaktär. När lärare kan se sin undervisning ur ett normperspektiv kan det påverka deras strukturering av matematik-undervisningen mot ett samtalande klassrumsklimat och ett fördjupat deltagande i elevers interaktion. Interaktion mellan elever och lärare bidrar till att eleverna får tillgång till andras matematiska uppfattningar vilket kan berika undervisningen signifikant. Enligt forskning möjliggör sådana strukturer förhandlingar av sociomatematiska normer vilket kan bidra till att utveckla högre kognitiva förmågor hos eleverna. De specifika sociomatematiska normerna som förhandlas fram i den sociala kontexten har potential att med lärarens stöd och hjälp ge elever rikare möjligheter att reflektera över, fördjupa och förfina sina förklaringar.