Chair: Engel Joachim
Chair: Ksenija Dumicic
Resources to Support Innovative Teaching: conceptual maps, data sources, and visualisation tools
Jim Ridgway, James Nicholson, Sonia Teixeira, Pedro Campos
Chair: Péter Kovács
Chair: Laczka Éva
CODAP for ProCivicStat
Using digital tools for teaching statistics about society
Chair: Pedro Campos
Chair: Daniel Frischemeier
Exploring Civic Statstics with CODAP
Jupyter Notebook and JupyterHub - Building Statistics Classrooms For the Cloud
Ioan Cristian Schuszter
Understanding statistics about society: a workshop on task analysis and educational challenges
Jim Ridgway, James Nicholson, Iddo Gal
GameSTAT-I. avagy a fejekbe surranó tudás (wokshop in Hungarian)
Anita Kolnhofer-Derecskei, Erzsébet Kovácsné Bukucs, Viktor Nagy
Visual Methods to Teach Multivariate Statistics with R
Chair: Djordje Kadijevic
Statistical literacy has never been more important that it is today. Humanity faces major challenges – many of our own making – such as famine, migration, poverty, and war. The world is increasingly interconnected, and, in the developed world, almost everyone is exposed to an ocean of data and information. There is optimism that citizens and policy makers will make better decisions, as high-quality evidence collected by agencies such as OECD, the UN, and statistics offices world-wide is made increasingly accessible. There is pessimism that when ‘fake news’ and ‘alternative facts’ go viral on social media, people will be misled and will make poor decisions.
Statistics education is at the heart of these challenges; there is widespread dissatisfaction because curriculums based on statistics invented over 100 years ago are no longer fit for purpose. Here, the fundamental questions of what students and citizens need to know, understand, and be able to do, in order to function effectively in a changing world will be addressed. Dimensions of statistical literacy will be explored. Exciting data resources, and data visualisation tools for teaching statistics will be mapped out. The opportunities and challenges afforded by big data will be discussed – including knowledge about analytic tools, and what citizens need to know about the uses to which personal data is put.
We are facing a battle over epistemology – the nature of evidence, and ways that evidence should be used in decision making is heavily contested. Statistics educators are just some of the combatants; an exploration of how statistics educators can learn from and contribute to the efforts of other agencies directly concerned with the quality of evidence and its interpretation in the public domain will be explored.
There are many resources for learning and teaching statistics. Large scale data, and statistical software are readily available for free on the Internet, and can be used to create lesson plans. However, what is needed to support curriculum planning is an overview of these resources – the substantive content of data sets from different providers, the affordances of different software packages (and information on the technical skills needed by the user), and the key statistical concepts that could be developed.
In this session, we introduce an innovative way to teach via quantitative evidence by using conceptual maps that link ideas, data sources, and visualization tools. Used as a learning and teaching technique, conceptual maps visually illustrate the relationships between concepts and ideas to help students organize and represent knowledge of a subject. The conceptual map supports multiple entry points: starting with a topic, the map suggests a data set, and the best graphics and tools to illustrate an idea; or you can start on a graphing tool (or a statistical measure) and the map suggests a topic and a data set.
A conceptual map is presented relating key social phenomena with authentic large scale data on topics such as migration, quality of life, sustainable development goals, and social inequality. The session includes practical exercises and several lesson plans will be suggested that can be used to help citizens to understand complex, real, social and economic phenomena.
AnswerMiner is a data exploration and visualization tool. Our mission is to help people to understand the word through their data, and help them to make data-driven decisions in a right way. We believe data analysis can be available and understandable to everyone. We melted UX and Data Science together and developed a cloud-based, user-friendly, well-designed web app, with wich data exploration is way faster than it was before.
During the workshop, we will introduce AnswerMiner in a short presentation, then we will start to explore it together.
Step by step participants will go through every feature, including smart data view, automatic charts, correlation matrix/table, relation map, decision tree, and find solutions for the given tasks. After that, there will be the opportunity to ask and discuss their questions and thoughts.
This workshop investigates the dimensions and knowledge-bases (statistical and other) that are needed to understand statistics about important trends in contemporary society, i.e., "civic statistics", and their educational implications. Participants will analyze the demands of real-world tasks taken from diverse sources (e.g., articles found in the media, dynamic visualizations) and items drawn from the GAISE 2016 report, and discuss how well current curricula and assessment systems reflect the skills and knowledge necessary to understand and react to such tasks.
The analysis of the tasks and the workshop as a whole will be informed by a conceptual framework and materials developed by ProCivicStat, a strategic partnership of the Universities of Durham, Haifa, Ludwigsburg, Paderborn, Porto and Szeged, funded by the ERASMUS+ program of the European Commission. ProCivicStat is designed to improve how teachers and students at the high school and undergraduate university level engage with and understand evidence and statistical messages relevant to the progress of societies. The workshop will involve three parts: a brief introduction, work in small groups on analysis of tasks, and a group discussion on educational implications. Participants will be given ahead of the workshop a description of the conceptual framework and a set of tasks.
CODAP — the Common Online Data Analysis Platform — is a project from the USA to create a web-based tool for data analysis in education at the secondary and post-secondary levels. Although it is not as powerful as a full programming environment, it is designed for beginners and to be relatively easy to use and set up. And it has some very powerful features, borrowing from its ancestors Fathom and TinkerPlots, and extending into new territory, such as in letting users organize data hierarchically through simple gestures. Therefore, it is worth investigating whether it would be a suitable environment for some of ProCivicStat’s work.
Our workshop can be flexible to meet the needs of participants; there are at least two main topics we can anticipate:
Participants can get initial experience with CODAP exploring ProCivicStat data and learning by doing what it is like to do data analysis in this evolving environment. We would also learn how to prepare data for CODAP and read it in.
We can discuss technical details. For example, CODAP is open source, and has an extensive interface for creating plugins that extend the existing tool, both in terms of getting new data and in creating new or specialized analyses or displays. What is involved in making a new plug-in to extend CODAP’s capabilities?
Tim Erickson (USA) has been involved in the international statistics education community for many years, with a particular focus on developing technology and using it in the classroom. He was one of the designers of Fathom, and has been involved in CODAP’s design and in creating plugins for CODAP for two years.
Az Óbudai Egyetem Keleti Károly Gazdasági Karán a névadó halálának 125. évfordulójára időzítve egy új módszert tesztelünk a 2017-es év tavaszi félévében. Az alapképzéseinkben résztvevő hallgatók részéről a tárgy elsajátítása során fellépő nehézségeket oktatói oldalról kihívásként értékelve, a célcsoportra szabott új módszerek alkalmazására készülünk. A Statisztika I. tárgy anyagának (főleg leíró statisztika) elsajátításához a hagyományos eszköztár alkalmazása mellett meghatározott játékszabályok betartásával hallgatóink konkrét példákon keresztül tesztelhetik tudásukat.
A workshopunk célja, hogy bemutassuk ezt az új technikát, mely során a „száraz” statisztika online formában, játékos keretek között (gamification) lopózik be a fejekbe. Több forduló során gyűjtött pontjaikat a játékosok saját döntésüknek megfelelően számíttathatják be a tárgy hagyományos számonkéréseinek eredményeibe. Mivel a program a tavaszi szemeszterben kerül megvalósításra, az őszi workshop időpontjára már annak eredményességéről és a tapasztalatokról is részletesen beszámolhatunk. Reméljük, hogy ezzel egy új hagyomány alapjait teremtjük meg, modernizálva ezen kimondottan hasznos tárgy oktatását és a hallgatóinkban tudatosítva a tárgy valódi gyakorlati jelentőségét.
This workshop, based on both theoretical consideration and practical work, deals with simple data modeling with pivot charts and dashboards, and didactical issues of this kind of modeling. The workshop is based upon the author’s IASE 2016 Roundtable paper [available at http://iase-web.org/Conference_Proceedings.php], as well as his recent experience in teacher professional development concerning this topic.
More details about the workshop will be available at www.mi.sanu.ac.rs/~djkadij/WS17.pdf in May 2017.
This workshop proposes to introduce potential teachers and students to a cloud-based, tool-backed method of teaching modern statistics and visualization classes to students of the modern era.
Teaching relevant statistics efficiently to students and visualizing the results has long been an issue in the classroom, especially in the context of modern-day students that quickly lose focus when they aren’t fully engaged in the lesson. A more attractive and interactive method of presenting the contents of a course is needed.
It would be desirable to have a readily accessible sandbox that one doesn’t need to configure, being able to jump straight into data analysis, using provided examples and datasets. A desirable environment is one in which users can interactively experiment and fiddle with statistics, until achieving the desired solution. Furthermore, an educator might wish to have an easy but flexible method of grading statistics assignments, but not only. All of the above are readily available by using the open-source Jupyter notebook environment, coupled together with the multi-tenant JupyterHub project.
Using these tools, one can swiftly deploy a virtual classroom in the cloud to which students can connect at any time, requiring just an internet connection. All code is run on the server and login can be achieved and integrated with JupyterHub in a variety of ways, according to needs and preferences. Each student that logs in is provided with his own unique environment. However, this environment is provided to each of his colleagues in the same initial form, like a carbon copy. Tampering and experimenting is encouraged from the get-go and the distribution of materials is made easy by creating new containers that provide new materials. Interactive code blocks can be written and executed in the language that one wants (Python or R are popular choices), giving a fast feedback loop to the one that is taught.
During the workshop, potential students, teachers or system administrators will be presented on how to setup and use these tools in order to teach or learn about statistics and data visualization. A test deployment will be setup so that all of the people that wish to participate can connect to the JupyterHub deployment in the public cloud (using Docker containers transparently spawned for each user) and take part in the interactive coding session.
We will go through a bit of descriptive data analysis on an example dataset present in each of the users’ containers, leveraging the power of interactive snippets to visualize our results and finishing with a few plots. Afterwards, an assignment example will be given and the participants will be invited to take it. The autograder feature will be shown, as well as how it can be configured and used in the classroom. The workshop will end with a short Q&A session and subsequent examples, if requested.
Data are abundant, quantitative information about the state of society and the wider world is around us more than ever. In order to re-root public debate to be based on facts instead of emotions and to promote evidence-based policy decisions we as statistics educators are challenged to promote understanding of statistics about society.
Based on the digital data science platform CODAP (http://codap.concord.org) we illustrate how complex multivariate data about issues of high socio-economic relevance (inequality in income distribution, racial bias in sports, human well being etc.) can be explored and visualized. Participants are encouraged to bring their own laptop and are guided into using CODAP
The majority of social science majors dislike statistics and research methods. Things can, and should be, different; (social) statistics and research methods can be interesting and fun. My intention is to do just that, make methodology fun. To achieve this, I currently work on an applied methodology book project (I also teach examples from it in courses in the MA in research methods at my institution). A few years ago, I used to teach a class called “Political Paradoxes”. The central goal of this class was to teach students both methods and substance in a fun way, using examples from the work of other scholars (e.g., the impact of shark attacks on voting behavior in US presidential elections, or how the end result of introducing electronic voting machines in Brazil was a significant drop in the share of underweight newborns!), as well as examples of my own - e.g., explaining why in Ukraine there was a negative correlation (r = -0.8!) between the regional share of the vote of Leonid Kuchma in two consecutive presidential elections. Thus, the book will take a number of unexpected results (mostly from the electoral politics of various countries), and explain them using the logic of social inquiry and basic research methods (such as linear regression and experimental design). The book will be a useful and fun companion to more mainstream research methods textbooks for upper-level introductory methods courses. Moreover, the use of thought-provoking, yet intuitive substantive examples from diverse countries will make the book an equally attractive choice, not just for methods teachers and students from political science, but also from other fields of social science. This book will fill the gap between the typical introductory methods textbooks and that of substantive political/social science (text)books. An even more important goal of the book and the classes I teach (and will be teaching) based on it is to make a contribution to enhance the students’ and the general public comprehension of statistics and methodology, and consequently make a better, more informed citizenry. Sadly, the urgency of this task is well illustrated in recent years by the rise of populist leaders all across both the East and the West; a core ingredient of their success is the public’s inability to distinguish between genuine arguments, based on facts, and pseudo-arguments based on shaky arguments and evidence.
Data is merely raw material of knowledge - the big problem is the ability of humans to use, analyse and make sense of the data. Everybody is talking about Big Data. However, without an understanding of underlying meaningful patterns and insights in vast troves of data, this is just an empty term. For that reason, graduation in statistical courses is as beneficial as it leads to the ability to process and interpret the real data and consequently to understand the contemporary world and to perform “data-driven” decision making.
We may observe two extreme approaches in teaching statistical courses for non-statisticians. Either the teaching is too theoretical - based on a “black board” approach without data processing. Or it is too superficial - represented by a “black box” approach; despite processing the data, this is performed without deeper understanding, just by clicking on recommended buttons with no insight what is behind.
The teaching innovation presented lies in combining both approaches. The lectures are supplemented by computer-aided seminar sessions where any problem based on the real data is solved both theoretically (which leads to deeper understanding “what’s going on”) and practically (processing the data by software itself). Careful attention is paid to verification of underlying assumptions of statistical methods, discussion and detailed interpretation of results including transferring the results into “common language”.
System thinking (not sticking to isolated issues) in a complex way with coherences to other themes is strongly supported.
Processing the data is also a part of the final exam which provides feedback to both students and lecturers.
The teaching of statistics is a fundamental part of the education of students studying business and economics. Statistics is therefore present in a considerable number of classes in the curriculum. The final goal is that the students are enabled to use the acquired knowledge and skills in a constructive way during their decision-making processes in the industry later on. Statistics, as an introductory course, is taught during the very first part of the curriculum. Unfortunately, most students are not able to recognize the importance of this subject for their future work and studies at this very early stage. Moreover, subjects related to mathematics are not particularly popular among students. In order to make statistics more attractive to the students and eliminate study problems, the introduction of new teaching methods and approaches, as well as tools is desired in the classroom environment. The computer aided teaching of statistics is one of the possibilities.
As shown by data from the literature, thanks to the development of computers, the possibility of incorporating personal computers into statistics classes was postulated as early as the 1970s. Many researchers would agree, that one of the most important benefits of using computers in the classroom is the speeding up of complicated calculations. This way the students’ minds are virtually liberated from the monotonous calculations so that they can pay more attention to statistical principles introduced in the class.
This presentation aims to show some advantages and disadvantages of teaching of statistics for students in such computer-assisted environments.
At the Budapest Business School, which received the university title last year, we gradually introduced from the first semester of 2013/14. the involvement of computer usage in Statistics I-II. courses both in BSc and MSc groups. The technical changes of the curriculum and education caused the modification of exams as well. Our study demonstrates how accepted and successful computer usage seems to be in different courses and its relations to the extent of drop-outs. After describing the past and present features of education in the university, we introduce BGE’s ideas about the future changes in Statistics courses.
Teaching statistics to the reluctant is hard work. We report here on an innovative approach to learning statistics in applied social research settings, by placing social science undergraduates into the workplace. We have collected data since 2014 based on our students' experiences of undertaking 8-week long paid internships in organizations ranging from university research centres, charities, social enterprises and think-tanks, local and national government departments, polling firms and data consultancies, and media companies, to international statistical organisations (The World Bank). This paper will report on (i) the level of statistical training that forms the foundation for this work (ii) the methods and data analysis skills that our students develop whilst in the workplace (iii) examples of outputs that they produce for their host organisations and (iv) the ways in which we nurture this skill-set for returning students in order to support our undergraduates to become future data analysts.
The paper is based on experiences at The University of Manchester, UK. The internship programme has been developed through our 'Q-Step Centre’ which is funded nationally (along with 14 others) to develop a step-change in quantitative methods teaching in the social sciences. The context for Q-Step is the long-standing quantitative skills deficit for UK social science and humanities graduates, as captured in numerous reports from research bodies and higher education funding councils dating back at least two-decades.
The University of Manchester Q-Step Centre has two strands to our approach to develop quantitative skills in the undergraduate population. First, by increasing the exposure to numerical data for all social science and humanities students from the start of their degrees, and second to train future quantitative social researchers by providing quantitative pathways through their social science degree.
To date 120 students have been hosted in 60 organisations across the public, private and community and voluntary sectors. This paper will share lessons learned from these experiences and highlight how we are now turning out attention to working with employers to ensure the university social science curriculum at Manchester aligns with labour market needs in a time of increasing volumes of data and a growing need for analytical skills to make use and sense of these.
There is a long-term problem of teaching statistics in higher education: Students are not motivated to learn this subject. They regard it as something useless in their future job. Even more: not useful in their daily practice.
In contrast I am strongly convinced that statistics can be an (the most?) important subject to develop analytical and critical thinking.
For this we should set new goals for our courses. It should be shifted from the statistical methods especially from the formulas to the analysis of real life problems. We should teach methods off course. We should solve simple data-analysis exercises - as we did in the past. However we need to set new goals, new focus: How to use the results of the classical statistical education to the analysis of single real life situations. Students may understand why some decisions were wrong - causing bad outcomes, accidents, catastrophes, ...
It can be especially important on the field of medical care. This approach of teaching statistics may form a solid foundation for the proper medical decision making.
I think however, that this approach can be extended to many other fields of higher education where students are not willing to perform 'real' statistical data analysis in their future job - that is at almost all university faculties.
I would like to present how I educate statistics. I educate students of the University of Pécs, Faculty of Humanities, Department of Sociology for 10 years. These students are oriented to Humanities more than to mathematics, that means a challenge for me. I would like to show the structure of the methodological part of the BA program, the focus will be on statistics. I will also illustrate how we use real data during the courses.
Az előadásban arra keressük a választ, hogy milyen mélységben jelenik meg a statisztika az agrár jellegű szakokon készített szakdolgozatokban és ennek milyen üzenete van a jövőbeni statisztika oktatás számára. A számítógépek és a statisztikai szoftverek fejlődésével megnyílt az út olyan statisztikai módszerek alkalmazása előtt, amelyek korábban nagy számításigényük vagy az algoritmus komplexitása miatt nehezen voltak elérhetők az agrár felsőoktatásban. A tankönyvekben és a kurzusok tematikájában jóval lassabban jelennek meg ezek az új lehetőségek. A Pannon Egyetem Georgikon Karán 1999 és 2016 között készített 4500 szakdolgozat áttekintésével képet kapunk arról, hogy a különböző szakokon készült szakdolgozatok milyen arányban tartalmaznak statisztikai elemzéseket, melyek a leggyakrabban alkalmazott eljárások. Kitérünk arra, hogy a hallgatók mennyire vizsgálták az egyes eljárások alkalmazhatóságát, milyen alternatív módszereket használtak a tananyagban szereplő módszerek mellett. Megnézzük, hogy milyen fejlődésen mentek keresztül a szakdolgozatokban megjelenő adatvizualizációs technikák. Végül ezek alapján javaslatot teszünk, hogyan lehetne az agrárképzésben megjelenő statisztika oktatást hatékonyabbá tenni.
Az előadás egyben egy példa a szövegbányászati eszközök gyakorlati alkalmazására is.
A gyakorlati statisztika oktatás mindig komoly kihívásokkal küzdött. Az elméleti tudás gyakorlatba való átültetése a hallgatók számára nehéz feladatnak bizonyul függetlenül attól, hogy statisztikusként vagy csak érdeklőd hallgatóként szembesülnek a problémával. Bár számos statisztikai elemző program segíti a kutatók, oktatók és hallgatók munkáját, mégis azt tapasztalhatjuk, hogy a szoftver nem helyettesíti az emberi elemzőkészséget. Nem tárja fel helyettünk az összefüggéseket, nem értelmezi az adatokat.
A statisztikai elemzésekre alkalmas szoftverek ugyan segítséget adnak a számítások elvégezésében, de emellett két nagyon fontos kérdéssel kell megküzdenie annak, aki az eredményeket később használni szeretné. Előre meg kell határozni a „Mit szeretnék megtudni?” kérdéskört, hiszen ennek ismerete nélkül a számítások nem sokat érnek majd. Emellett pedig az elemzések lefuttatása után meg kell válaszolnia a „Mit is jelentenek ezek a számok?” kérdést.
Az elmúlt évtizedek oktatási tapasztalatai alapján elmondható, hogy a leggyakoribb problémát nem a különböző statisztikai alkalmazások használatának nehézsége okozza, hanem sokkal inkább az előkészítés és az értelmezés problémája.
A statisztikus és nem statisztikus hallgatók oktatásában egyaránt célravezetőnek bizonyult a valós adatok, gazdasági problémák elemzése, önálló kutatási munkára és a tanultak minél szélesebb körben való alkalmazására való ösztönzés, mely során a hallgató olyan, a mindennapokban fellelhető problémákra keresi a választ, mely akár közvetve vagy közvetlenül az ő életére is hatással van.
Oktatási célunk, hogy hallgatóink korszerű ismeretekre tegyenek szert a statisztika gyakorlati ismereteinek elsajátítása során, amely alkalmassá teszi őket a piaci kihívások gyors és hatékony kezelésére, piacképes tudásuk előnyt jelenthessen a munkaerőpiacon. Fontosnak tartjuk, hogy olyan tudás birtokába kerüljenek, melyet egy kisvállalkozás vagy egy multinacionális cégnél való elhelyezkedés esetén egyaránt tudnak hasznosítani. Kiemelt figyelmet fordítunk arra, hogy segítsük az önálló gondolkodást, kutatási problémák feltárását, összefüggések feltárását, ok-okozati viszony vizsgálatát, hogy gondolkodó nemzedéket neveljünk.
This paper examines new directions and ways to help non-specialist adult users of statistics develop aspects of statistical literacy that pertain to official statistics, with a focus on what official statistics providers and statistics educators can and should do, possibly in collaboration with statistics educators.
We first point to a gap in the literature in terms of the lack of an agreed conceptual basis needed to develop educational materials that can help adults at large, educators, and even regular statisticians, understand key aspects of official statistics and be critically aware of social and economic phenomena. We then present a new model encompassing six elements about which non-specialists and adults at large should possess knowledge in order to make sense of official statistics, including knowledge about: (1) the system of official statistics and its work principles; (2) the nature of statistics about society (as identified by the ProCivicStat project); (3) indicators; (4) selected key statistical techniques and concepts; (5) research methods and data sources; as well as (6) awareness and skills needed for citizen access to statistical reports and products from statistics providers.
Based on this typology we discuss directions that official statistics providers and statistics educators could take. We argue for the need to advance or validate the conceptualization of skills needed to understand official statistics as outlined in our model, and accordingly to expand educational activities and services to improve public capacity for understanding of official statistics. In particular, we focus on two needs that require that statistics providers and statistics educators combine forces, in terms of developing a digital textbook about official statistics literacy that is based on the proposed model, as well as developing a modular online course, that would be both geared for and tailored for non-specialists and citizens at large. Further implications for statistics education and for future research are also highlighted.
The higher the lifelong learning development level or the participation rate in education and training of adults in a country, the more persons are competitive on the labour market, and the greater the chance for a national economy to be more competitive. All these lead to higher development level in a country and to higher well-being level of citizens.
Here the lifelong learning is defined as a participation rate in education and training of people aged from 25 to 64 years (Eurostat). In order to better understand this participation rate level, this phenomenon is observed from different views: by the gender, by the employment status, by the educational attainment level (primary, secondary, tertiary), and by the urbanisation degree (city, rural areas).
Descriptive statistics analysis results revealed great disproportions in the participation rates in education and training between the European countries observed. In 33 European countries included, in average 11.4% of total population aged from 25 to 64 years participated in education and training in 2014. Only in Switzerland and Denmark more than 30% of total population aged from 25 to 64 years participated in education and training in 2014, whereas the lowest participation rate there is Romania, Bulgaria, and Croatia. The average percentage of total population aged from 25 to 64 years which participated in education and training increase from 9.95% in 2006 to 11.41% in 2014.
Because of missing data problem, two non-hierarchical cluster analysis approaches were conducted. In the first approach, countries for which data is missing are omitted whereas in the second approach variables with missing data are excluded. In both applied clustering approaches, the conducted cluster analyses resulted in three clusters of countries. The both clustering approaches have resulted with clusters with almost the same countries as the elements. In both clustering approaches there was no doubt which group of countries has the lowest economic development level. The conducted statistical tests shown that this group of countries also had the lowest lifelong learning development level. Despite of clear differences in lifelong learning development levels, countries in other two clusters have statistically the same economic development level. Consequently, the research hypothesis that the European countries with higher average lifelong learning development level have also higher average economic development level can be only partially accepted.
As statistical literacy arose to be very crucial for better citizenship of individuals, it is imposed that the lifelong learning activities may include educational courses in that direction to the higher extent. If people should become more statistically literate, they needed to be taught to use statistics as evidence in the arguments encountered in their daily life as citizens, workers, consumers, patients, voters, etc., compare to Shield (2001). Therefore, the role of statistics educators becomes very responsible. The lifelong learning for improving statistical literacy is especially important for journalists and all those who influence the public opinion. All this seems to be quite important tasks and challenge for statistics educators, as well for those that organize such educational programs for all society segments.
The Hungarian Central Statistical Office recruits people with university degree achieved in different institutions in lack of special training for official statistics. The statistical knowledge of people entering/working in the statistical office is very heterogeneous, unfortunately there is no a solid standard basic knowledge in statistics, on which to build specific internal trainings for them.
The key areas where we miss general knowledge: basic statistical concepts, handling large datasets (more then what fit on the screen), use of at least one statistical software, understanding mathematical background of basic concepts and methods and finally to understand the relation of statistical characteristics and phenomena in the real world. Beyond special knowledge needed in statistical office there is an emerging need for a certain group to learn basic math and statistics, normally supposed to be known.
The statisticians need sophisticated thinking in understanding and translating real social and economic phenomena into statistical concepts, need a general overview of the statistical business process model, and within this, special knowledge in the future working area, process phase (like sampling, seasonal adjustment) and/or subject matter area. A system of courses is offered.
Teaching is burdensome. In some special or newly emerging areas there may be even lack of knowledge. These are the reason why EMOS, ESTP courses and new flexible teaching methods are welcomed.
On the other side experts from HCSO take part in teaching statistics at universities. It is important that students having a question/task on a real phenomena should have to find the most adequate data, understand them with metadata, choose and apply method to analyse and finally answer the original question. Our experience is that the main problem is to translate the real problem in a statistical question, and at the end interpret the results, i.e. translate the parameters in phenomena of the real world.
One can define statistical literacy as the capacity to challenge statistics encountered in everyday life. Above all, it enables us to “consume and critically digest the wealth of information being produced in today’s society” (Rumsey, 2002). In other words, statistical literacy helps one ask better questions and improve judgment, as well as decision making.
Arguably, statistical literacy is a key ability in a society where data, variation and chance are ubiquitous. Since most adults are consumers and not producers of statistical information, the skills associated with statistical literacy could be activate in various contexts. For instance, it helps people to be aware of trends and social phenomena evolution: crime rates, spread of diseases, environment changes, employment trends, political or social polls results. What is more, it supports people improving their decisions when confronted with chance-based situations such as buying lottery tickets, understanding risks associated with certain diseases or diagnostics, evaluating financial risks. Consequently, it makes people informed consumers of data. That is way it becomes vital to have statistically literate citizens.
Yet, despite the importance of statistical literacy and statistical reasoning in various life contexts, the subject is underrated in curriculum frameworks both in secondary / high school, as well as in undergraduate courses. The focus in courses associated with Introductory Statistics is mainly on the mathematical context underpinning the statistical concepts and regrettably low emphasize is put on the practical aspects of these concepts.
This paper aims at filling in this gap by suggesting teaching activities that would provide students opportunities to develop skills, attitudes and abilities of a statistical literate citizen. For this we propose a series of case studies, interactive graphical representations and exercises developed based on real survey database. Respondents in this survey are young people, aged between 15 and 29, questioned about their beliefs, attitudes towards life and society. For the presented activities, we introduce the objectives aimed and skills students should acquire after following the specific tasks. Main themes covered are: summarizing and presenting data (various graphical representations), main descriptive indicators and different contexts where they can be used, correlation (both parametric and non-parametric).
The proposed course activities target the undergraduate students in social sciences, thus no prior knowledge in statistics or probabilities is required, basic algebra being the only prerequisite.
Classical lectures are rather inefficient for the members of the Z-generation. They do not feel any motivation in it and it is usually not helpful in the learning procedure. They need much more chance to interact in the classroom and this interaction must be forced. We also need to make them realize that during their university studies basically they will not be taught but they have to learn on their own to reach deep, advanced level knowledge and skills by the end. The balance of the responsibility of their studies must be pushed toward them, so they will be able to handle their future life responsibilities as well. If the students are motivated to study before the “lectures”, there is a chance in the classroom for a real discussion of the topics. The aim of this presentation is to introduce some techniques, methods that can be used in a “flipped classroom”, together with the experiences of the previous semester in a 2+2 type statistics course.
This talk proposes different technology tools to increase interactivity in teaching statistics courses. Especially it is focused on Audience Response Systems (ARS) that usually uses a hand-held control unit („clicker”) and/or a mobile device, and how to get started using it efficiently. We survey some free and some commercial services and solutions and show their infrastructural requirements, educational advantages and disadvantages. We exhibit a result of an ARS readiness questionnaire of students and lecturers to outline the main expectation of the end users. Finally we share how classes were selected for a pilot test, and demonstrate some of our real life experiences using one of the systems.
In teaching demography (in particular demographic statistics) it is crucial to use information obtained from different resources located on various websites of producers of official statistics. New technologies and software programs are essential to handle with lots of data and to take full advantage of a specific data resources. Decision on using specific IT tool depends on the following issues: adaptation to the various levels of knowledge and ability of students, possibility of testing different solution to a given question, and opportunity to play an active role in the learning process. We present and describe some educational software for teaching demography. We consider software dedicated specific to demography as also “nondemographic” software. Presented examples provide a comprehensive assessment of the usefulness of IT tools, their advantages and limitations.
Assessment and examinations are major parts of a course and also a great workload for instructors. Creating new exercises for every exam is a time consuming task. In this talk we introduce an application of the exams package in R and the Moodle environment to create personalized exams, meaning every student has a (slightly) different exam. The approach has a lot of advantages for both sides: automatic assessment of midterms/finals, less chance for cheating and less continuous workload during the semester for teachers. In our experience students like the immediate feedback of the system and it is easy for them to use it. Of course there is an initial “cost” of creating the dynamic examples. The instructor can use categories to store the exercises which makes keeping track of them easier.
The paper considers some examples of incorrect applications of measures of asymmetry and shape of one dimensional distributions. It points to some misconceptions of establishing skewness using relative positions of mean, median and mod that frequently occur in introductory textbooks. The examples of violations of the rule that if the distribution is skewed to the right, then the median is greater than the mode and vice versa for the skewness to the left are given for discrete and continuous distributions. Also, the example which disproves assertion that for skewed distributions the mean lies toward the direction of skew relative to the median is presented.
The coefficient of asymmetry or skewness is also often misinterpreted. The fact that it is zero does not imply that the distribution is symmetric both for discrete and continuous distributions.
Furthermore, it is wrongly thought that the coefficient of kurtosis measures the peakness of a distribution and that leptocurtic curves are more sharply peaked and platykutic curves more flat-topped than the normal curve. The examples in the paper show that it is not necessarily true. Since there exist distributions with same kurtosis but with different shape, it is wrong to assume any major dependence between the coefficient of kurtosis and the shape of a distribution.
Researchers are often faced with the major dilemma of identifying causation relationships between chunks of data that predict each-other. Even though this usually is the result of poor research design or methodological errors, investigative research projects may intrinsically contain such data given their exploratory nature of looking into the unknown. But is there anything that can be done when flawed study design does provide unclear results, with no possibility of identifying which factors predicts which? Converting quantitative investigations into mixed research models containing qualitative components may prove very useful for accurate interpretation of statistical analysis results, thus saving an otherwise doomed endeavor.
One of the main goals of civic-education is Mündigkeit (The German word Mündigkeit is difficult to translate, because it refers to a cluster of ideas around empowerment, emancipation, self-determination and taking responsibility). Mündigkeit is an indispensable prerequisite for democracy. The functioning of a democratic society is dependent on mündige citizens who contribute to the development and maintenance of a democratic culture. Mündigkeit is a prerequisite for participation, further development and the stabilization of the democratic system. Mündigkeit means having an orientation in the confusion of the modern information jungle and the deluge of quantitative information and statistics. It can be seen as a bundle of qualities that enable the citizen to obtain the necessary information, to make sense of data, to ask critical questions and to understand them in order to be able to make a decision.
In the information age, the requirements for understanding and evaluating information about societal developments have changed. Statistical skills are becoming increasingly important for an evidence-based judgment in today's society. They entail understanding data-related arguments or representations, questioning possible conclusions as well as uncovering opinions and pre-made decisions. Therefore, statistical knowledge is indispensable for Mündigkeit.
Critical statistical thinking and a sustainable knowledge in civic statistics is inevitably to become concerned citizen. This kind of thinking and knowledge can already be enhanced in secondary school. For the implementation of civic statistics in mathematics classrooms in secondary school, teachers themselves have to be well educated in the field of civic statistics. For this purpose we have designed and realized a university course about civic statistics in the winter term 2016/2017 at the University of Paderborn, where preservice teachers worked on projects and activities in regard to civic statistical contexts. For instance they have analysed official open data of the German Statistical Office on the German Gender Pay Gap with digital tools, explored the distribution of net assets in German households and investigated the unemployment situation in different countries in the European Union via Google Public data files. In this talk we will present some activities and have a closer look at the outcomes of the participants working on these activities. Finally we derive first implications for re-designing of the activities.
Not only statistical method’s usage but also their applications on real data have key importance in statistics classes, which can help to understand social- and economic phenomenon. IT tools, real data and visualization tools can mean the link among the statistical methods (for instance correlation, descriptive statistics, regression models) and application fields (for instance ageing society, poverty, income inequalities). Besides applications, statistical literacy should be also improved in the classes.
The mail goal of the presentation is to show statistical course materials which are using real statistical data and newer interactive visualization tools (for instance population pyramid, Gapminder, heat map, online tools). The ProCivicStat project is the framework of our work.
The lecture aims to exemplify how statistical concepts, which generally handled as separate knowledge, to bring to a common platform. This , in our view, makes the thinking and teaching process more effective and economic.
Five number of instances illustrate the effort above. The ‘identity’ of average and median is demonstrated as well as the use of scalar product in three different statistical situations (average, Steiner-equation, regression). The well-known base-relation for the normal distribution is applied in five major directions of mathematical statistics. Methods of visualization is grouped in a simple, but useful way, and finally, the regression task in its most elementary form is discussed in a partially new, less popular form.
The higher number of illustrations allow only an outline form of explanations.