A Cross-National Comparison of Statistics Curricula

The purpose of this study is to compare Turkey to several other countries such as Singapore, Korea, the United States, and New Zealand according to the field of statistics learning in the curriculum. GAISE (Guidelines for Assessment and Instruction in Statistics Education) framework suggests using the components of statistical process (formulate question, collect data, analyze data, interpret results) and developmental levels of statistical process (Levels A, B, and C) in such comparisons. It is found that in both Korea and Turkey, the number of learning outcomes of statistics, as well as the weight of learning outcomes in the curriculum, have been having a weaker profile than in other countries. The field of statistics learning is mostly condensed around Level A in Turkey. On the other hand, other countries in the study included more Level B and C. Considering the importance of statistics instruction, Turkey’s national mathematics curriculum needs improvements in its field of statistical learning. This study suggests designing and developing new standards for teaching statistics that deal with higher developmental levels of statistical process.


Introduction
In a life surrounded by data, the need to make effective decisions and inferences from data makes it a necessity for individuals to have required knowledge and skills related to statistics and to transfer them to their daily life. As a matter of fact, individuals equipped with such skills in statistics are able to make sense of the data complexity they encounter and to make effective decisions. In the standards of National Council of Teachers of Mathematics [NTCM] (2000), it is pointed out to the importance given to statistical information in keeping up with changing world conditions and it is stated that statistics constitute one of the cornerstones of mathematics education. In addition, the ultimate goal of statistics education, it is stressed that raising individuals as statistically literate and to be able to transfer their statistical knowledge to their lives (Franklin et al., 2007). Today, in parallel with the increasing need to raise students with statistical thinking competence, it is also important to make improvements in statistics teaching at all educational levels (Cooper, 2002). Parallel to this importance, statistics topics are taught more widely in all educational levels from primary school to university in many countries, and recommendations that improve statistics teaching are encountered (Ben-Zvi, & Garfield, 2008). This orientation to the field of statistics learning has also affected our country and it is seen that with the regulation made by the Ministry of National Education [MoNE] (2005), subjects and learning outcomes related to this field have been included in the mathematics curriculum since 2005. On the other hand, in order to raise students as statistically literate, it is important that the changes in the mathematics curriculum in our country should reflect the trends towards statistics education worldwide. This situation undoubtedly causes countries to differentiate in terms of statistical literacy, statistical thinking, and statistical reasoning competencies, which are defined as important outcomes of statistics education and the weight they give to the field of statistics learning in the mathematics curriculum. Accordingly, it is necessary to examine the field of statistics learning in mathematics curriculum in different countries and to compare this field n our country with the others. In this study, it is aimed to compare Turkey to several other countries such as Singapore, Korea, USA, and New Zealand according to the field of statistics learning in the curriculum. For this aim, learning outcomes in the field of statistics learning have been compared. Learning outcome has been specified as "content" in Singapore and Korea, "standard" in USA. ,and "goal" in New Zealand. In order to prevent these differences from causing confusion, the expression "learning outcome" is used in the present study.

Theoretical Framework
In this study, the theoretical framework included in the GAISE report (Franklin et al., 2007), which draws attention to the principles determined for statistics instructionwas used in comparing mathematics curriculum of countries in terms of the field of statistics learning. In the report, statistical process is taken as a basis for teaching statistics, and instructional practices that bring statistical process components to the forefront are explained. In addition, the framework presented in this report, the aspects related to the statistical process components were discussed at three hierarchical levels (A, B and C). While analysing the mathematics curriculums of the countries included in this study in terms of the field of statistics learning, it was based on statistical process components and the three hierarchical levels related to these components.

Statistical Process
Statistics allow individuals to direct their lives and to realize different types of thinking about the quality of their decisions. Based on this importance of statistics, it is a substantial need to teach statistics in schools effectively. In the GAISE report, it is underlined that the effectiveness of the student in the learning environment is a significant variable in increasing the quality of statistics instruction (Franklin et al., 2007). It is emphasized that students who have the opportunity to experience statistical process components will be statistically active and productive (NTCM, 2000). Therefore, the statistical process has a key role in making statistics teaching more effective in the GAISE report. Rumsey (2002) has emphasized to raising individuals who can manage the statistical process among the goals he stated for statistics education. Many studies point out the importance of the statistical process (Güven, Öztürk, & Özmen, 2015;Newton, Dietiker, & Horvath, 2011;Özmen, 2015;Rumsey, 2002;Topan, 2019;Watson, 2006). In this sense, it is clearly seen that the statistical process that includes question formulation, data collection, data analysis, and interpretation of results is important in developing students' statistical knowledge and skills (Franklin et al., 2007;Özmen, 2015;Özmen, & Baki, 2017). These steps contribute to gaining research experience, displaying a critical approach and developing psychomotor skills (Wild, & Pfannkuch, 1999). Hence, the statistical process emerges as one of the important factors of effective statistics teaching (Franklin et al., 2007). Taking statistical process as a basis in statistics teaching will also contribute to educators in terms of designing effective learning outcomes (Groth, 2013). In addition, it is thought that planning the statistics teaching according to the statistical process components and designing the activities in this direction will contribute to make leraning environments richer and to provide students more active. Moreover, it is predicted that the teaching designed according to the statistical process will provide the opportunity to train students as a researcher and more effective results in terms of teaching. This functional role of the statistical process has been effective in comparing curriculums in terms of the field of statistics learning. In this sense, the fields of statistics learning in the mathematics curriculums of countries were compared according to statistical process. This comparison is based on the GAISE report (Franklin et al., 2007), in that it explains the expected behaviours at different levels towards the statistical process components and includes recommendations to be taken as the basis for statistics instruction.

GAISE Report
GAISE is a report published by the American Statistical Association (ASA),includes recommendations and instructional principles for statistics education. The principles emphasized by this report also affect the whole world, but it is also important in terms of providing a broad framework for statistics instruction. In GAISE report, the main purpose of statistics instruction is to raise students as statistically literate and there are instructional suggestions in this direction. This report has focused on statistical literacy as the ultimate goal in statistics instruction, and it is emphasized that every high school graduate should be equipped with the statistical reasoning skills to lead a happy and productive life and be able to reflect this skill in daily life (Franklin et al., 2007). In addition, it has been pointed out that raising conscious and productive individuals on behalf of the society can only be possible with a good statistical education. In this sense, GAISE report provides important recommendations to teachers on how to teach statistics. This report, which is shown as a significant resource for statistics education in the world, is frequently used in the design of learning environments. The basic framework that build on the statistical process will make students active in statistics teaching and make statistical concepts more understandable for them. The statistical process is described as the map of a rich statistics teaching (Franklin et al., 2007). Accordingly, the statistical process in the GAISE report comprises of 4 successive components. The first one is the formulation of the problem and it is considered important in terms of starting the statistical process and directing the development of the next components. In this component, it is significant to formulate an appropriate problem and to decide on the type of data to be obtained. The second component refers to the data collection suitable for the research problem. In this component, the selection of the sample representing population, sample size and data collection methods are emphasized, and appropriate measurements are decided for the analysis. The third is the data analysis, and in this section the data can be summarized numerically or visually presented in tables and graphical representations. In the last component, appropriate inferences are made by interpreting the results obtained. In this section, the results are interpreted in their context. Meanwhile, it is thought that new problems may be formulate according to the participation of different individuals and thus it is assumed that the statistical process will have a continuity. Making sense of these components takes place depending on time and instructional practices. This developmental process is presented in the GAISE report at three levels (A, B and C) for each component, ranging from descriptive statistics to inferential statistics. In other words, the statistical process is included in the GAISE report as a framework for acquiring higher level skills based on basic skills. In this direction, at level A, it is stood out to know the basic principles of statistics and the meaning of statistical concepts; at the level B, it is required to associate the basic statistics concepts with each other and to make comparisons over statistical situations. Finally, at level C, it is needed to generalize of the results, to know of advanced statistical concepts and their association (Franklin et al., 2007). For instance, in data collection, while determining the music group to be brought to the end-of-year school concert, A-level-students collect data with a simple experimental process by taking the opinions of their classmates. At level B, realizing that only one class will not represent the whole school, other classes are also included in the sample. Realizing that the sample should be representative of the population, students get to the level C when they start making random choices, assuming that everyone has an equal probability of being selected. Thus, there is a development from choices limited to one group to random selections from a large community.
The suggestions and practices of the GAISE report have an important place in determining the level of statistics teaching. It is stated that statistics teaching carried out in line with these recommendations and practices will be effective (Franklin et al., 2007). At this point, the goals and learning outcomes in the field of statistics learning in the mathematics curriculum, which form the basis for designing learning environments, stand out. It is thought that these learning outcomes will give an idea about the statistics teaching. Therefore, it is important for basing on the theoretical framework presented in this report in terms of seeing whether the curriculum constitutes a resource for more effective teaching. From this point of view, the two-dimensional theoretical structure, including the statistical process components and the three hierarchical levels, in the GAISE report (Franklin et al., 2007) is given in Table1. Table 1. Theoretical framework in GAISE report (Franklin et al., 2007, p. 14  In the GAISE report, it is stated that individuals' experiences of the statistical process have an important place to be a qualified statistical literate. At this point, it has been indicated that educators and program developers should design practices that will enable students to experience the statistical process. In this way, the trend of statistics in recent years and the emphasis of GAISE reports on qualified statistics teaching have enabled many countries to innovate in the mathematics curriculum. Within the scope of innovation studies, the fact that countries have included statistics at different densities in their curriculum has paved the way for some countries to come into prominence in statistical education. For this reason, comparative analysis of different countries' curriculum enables countries to see their strengths and weaknesses for educational activities and to integrate effective practices into their own systems. When the literature has been examined, there are many studies on comparing the mathematics curriculums of countries (Altıntaş, & Görgen, 2014;Çelik, Kul, & Çalık-Uzun, 2018;Erbilgin, 2014;Erdoğan, Hamurcu, & Yesiloğlu, 2016;Güzel, Karataş, & Çetinkaya, 2010;İncikabi, & Tuna, 2012;Kaytan, 2007;Kul, & Aksu, 2016;Özkan, 2006;Özreçberoğlu, & Kıvanç-Çağanağa, 2016). In these studies, mathematics curriculums were compared in a way to cover substantially all learning fields by stating the general characteristics and objectives of the curriculums. However, there are also studies to compare the mathematics curriculums of countries in terms of a specific learning field (Arık, 2007;Tezcan, 2016;Uğur-Arslan, 2015). In these studies, the mathematics curriculums were examined for learning fields such as numbers, algebra, and geometry, but it is seen that there have been no studies directly on the statistics learning field. Hence, it is revealed the need to examine different countries in terms of the statistics learning field and to compare them according to similarities and differences. The fact that the GAISE report is an important resource for statistics education and how or to what extent the educational principles and practices pointed out in the GAISE report are based on the curriculums of countries have important roles in making these comparisons. Considering that the curriculum shaped by the report will be the starting point of effective statistics teaching, it is considered that the curriculums of the countries should be examined in terms of similarities and differences in terms of the field of statistics learning in accordance with this report. At this point, comparing the curriculums on the basis of statistics and making the necessary arrangements in the curriculum depending on the results will provide the opportunity to make the necessary changes for students to be statistically equipped. In the present study, it is aimed to examine Turkey with Singapore, Korea, USA and New Zealand mathematics curriculums in terms of the field of statistics learning according to the GAISE report (Franklin et al., 2007) and to compare the tendency of the learning outcomes according to the theoretical framework presented in the report. Depending on this main purpose, the research questions were determined as follows:  How learning outcomes of the field of statistics learning of countries differ in terms of the developmental levels presented in the GAISE report?  How learning outcomes of the field of statistics learning of countries differ in terms of statistical process components presented in the GAISE report?

Method
This study is a comparative research that intends to compare Turkey to Turkey, Singapore, Korea, USAUSA and New Zealand in terms of the field of statistics learning in the mathematics curriculums. Document analysis method, which is of qualitative research patterns, is used for presenting curriculum of the countries in this study. The mathematics curriculums of the five countries have been analysed in accordance with the objective of the study and the theoretical framework defined within GAISE report.

Countries Involved in the Study
Two criteria are used to determine the countries to be included in the study.. The first one is that it is emphasized to take place near the top at exams of Trends in International Mathematics and Science Study (TIMSS) which is implemented as an assessment exam with the aim of monitoring education systems of countries and Programme for International Student Assessment (PISA) which emphasizes on students' ability to use knowledge and skills they learn in school in daily life. Singapore and Korea have been chosen to be involved in the study among the countries that are succeed in the PISA (2015) and TIMSS (2015) exams. Secondly, USA (Lovett, & Lee, 2018;Newton, Dietiker, & Horvath, 2011) and New Zealand (Burgess, 2007;Forbes, 2014) have been preferred among the countries that have come to the fore with their recent studies in the field of statistics.

Data Collection and Analysis
Learning outcomes (content, standard, goal) regarding the field of statistics learning in the curriculum of the countries are analysed in line with levels (A, B, and C) specified for statistical process components of GAISE Report. The curriculums analysed within the study (upgrading year is 2018 for Turkey, 2013 in primary and secondary grade and 2016 in H1, H2 and H2 Forward (H2F is qualified as 13 th grade) grade for Singapore, 2015 for Korea, 2012 for USA and 2012 for New Zealand) include primary, secondary and high school mathematics curriculums in accordance with K-12 specified in GAISE Report. Learning outcomes and explanations in curriculums are evaluated within the context of the aspects regarding A, B and C levels in statistical process components defined in the theoretical framework. As classifying statistics learning outcomes taking place in curriculums according to their related component and levels, in case of the fact a learning outcome has an explanation related to more than one statistical process component related learning outcome takes place in both components. These kinds of learning outcomes are written clearly in the first component that is included (learning outcome number and label) and only the number of the learning outcome is specified in other components. Where a learning outcome is suitable for more than one level, it is written at the highest level that it can go up to prevent such repetitive situations to occur. Analysis process is shown on an example in Table 2. As analysing the Table 2, SO1.2 learning outcome in secondary education level in Singapore includes expressions regarding data collection and data analysis. Due to the fact that the expression of the learning outcome is required to use basic knowledge directly, it is limited with aspects specified in Level A for both components. P4.3 learning outcome of 4 th grade of primary education includes a target behaviour toward Level A due to requiring students to do calculations at a basic level by taking features of line graph for a certain statistical question. In the continuation of the learning outcome, expressions regarding Level B such as relating between bar and line graphs and determining which graph to be used in case of statistical situation by comparing these two graph types take place. Consequently, this learning outcome is evaluated to be included in Level B.
The stages suggested by Forster (1995) are conceived while examining the document. Firstly, a detailed research was carried out to reach first-hand mathematics curriculums used as document. Secondly, the originality of the accessed curriculums was checked by taking into consideration that it is located on the page of the Ministry of Education of the country to which it belongs. Then, the aims and objectives of them for general and mathematics teaching were examined to understand the curriculums. In this direction, the processes of understanding and analysing the statistics learning outcomes have been started. In the fourth stage, the outcomes are analysed in line with the theoretical framework included in the GAISE report. A common analysis procedure related to analysis of learning outcomes regarding in the field of statistics learning of the countries was followed. Firstly, statistical process component and expected behaviours in three levels for every component were analysed in detail in GAISE report. It was discussed on situations differentiating each component and level from another and their scope is revealed. Thus, it was enabled to initiate a common analysis procedure. Following determining the procedure concerning the analysis, statistics learning outcomes in a mathematics curriculum of a country were analysed by all researchers. In the analysis, firstly, the statistical process component to which a learning outcome is related, and then its level were determined.Through process of decision making, it is discussed and discussions are continued until reaching a common idea over theoretical framework related to GAISE report in case of any dispute. Thereafter every researcher have analyzed learning outcomes of a country in accordance with common analysis procedure that is determined. To ensure the reliability of these analyses, attention is paid to ensure that all researchers work in the field of statistics education and are close to the research field. Following the completion of the analysis, descriptive analysis were applied for classifying learning outcomes regarding the field of statistics learning in the mathematics curriculum of a country on the aspect of related statistical process component and level. After the researchers completed the coding process, descriptive analysis were performed. Researchers re-encoded and confirmed whether the results reflect the reality. The data obtained at the last stage of the analysis process have been presented with the help of frequencies and percentages. In the last stage, the analysis results are interpreted in line with the research problems.

Findings
Learning outcomes regarding in the field of statistics learning in the mathematics curriculum of the countries are analysed by using theoretical framework presented by GAISE report. Firstly, learning outcomes included in mathematics curriculum of countries are evaluated according to hierarchical levels presented by theoretical framework. Following this, the distribution of the learning outcomes associated with the levels in terms of statistical process components is included.
Distribution of related levels (A, B and C) of statistics learning outcomes included in mathematics curriculum of the countries are shown in Graph 1.
Graph 1. Distribution of statistics learning outcomes of countries by A, B and C levels (%)* *Calculations have been made taking into account all the learning outcomes of the countries in the field of statistics learning.
As looking into Graph 1, it has been shown that learning outcomes at level B and C, requiring higher-level of statistical knowledge and skills are intensified at Singapore and New Zealand, USA and Korea include learning outcomes at such levels in their mathematics curriculum as well. In Turkey, it is seen that learning outcomes related to the field of statistics learning that is included in mathematics curriculum go up to Level B at most and no learning outcomes are at Level C. Aside from the fact that the ratio of learning outcomes regarding Level A requiring statistics knowledge at more basic level is seen to be more, this ratio points up to be higher than the ratio of other countries. While the ratio of learning outcomes regarding Level B is higher in Singapore, New Zealand, and USA the ratio of learning outcomes regarding Level A and B are equal and are more than C level in Korea.
The weight given to the field of statistics learning in mathematics curriculum of the countries and results (in frequency and percentage) regarding accordance with Level A, B and C of statistical process components associated with learning outcomes in this field are included in Table 3.
Tablo 3. Distribution of the learning outcomes of the countries in the statistical process components in terms of A, B and C levels.  *Related ratio is acquired by dividing the number of learning outcomes regarding related component and levels of countries into total number of learning outcomes related to the field of statistics learning in mathematics curriculum. (Exp: Having 33 statistics learning outcomes in total and 1 learning outcome belonged to Level A at the component of formulate question in Turkey, the ratio is calculated as 3%.) **Related ratio is acquired by dividing total number of learning outcomes that every country has included regarding related statistical process component into total number of learning outcomes of all countries for the component. (Exp: By having 15 learning outcomes in total regarding the component of formulate question included in curriculum of all countries and Turkey having 2 learning outcomes belonged to this component, the ratio is calculated as 13%.) ***Related ratio is acquired by dividing total number of learning outcomes regarding a certain statistical process component in general into total number of learning outcomes included in all components. (Exp: By having total number of 549 learning outcomes regarding statistical process component and 15 of these learning outcomes belonging to the component of formulate question, the ratio is calculated as 6%.) ****Related ratio is acquired by dividing total number of learning outcomes related to the field of statistics learning each country into total number of learning outcomes included in mathematics curriculum of each country. (Exp: By having 574 learning outcomes in mathematics curriculum of Turkey and 33 of these learning outcomes belonging to the field of statistics learning, the ratio is calculated as 6%.) Table 3, the number of learning outcomes related to the field of statistics learning in mathematics curriculum of New Zealand is shown to be higher among all countries. Besides, comparing the ratios of statistics learning outcomes, New Zealand is noted to have the highest ratio (30%). Despite Singapore being the second on the aspect of total number of statistics learning outcomes, it is shown to come the third (10%) following USA (12%) on the aspect of proportion of statistics learning outcomes. Furthermore, Korea has the least learning outcomes related to the field of statistics learning. Aside from Turkey having more learning outcomes in number than Korea. Korea (5%) and Turkey (6%) have almost the same percentage on the aspect of the weight of learning outcomes in curriculum. Consequently, it is seen that Korea and Turkey have weaker profiles in comparison of other countries on the aspect of both the number and the weight of statistics learning outcomes in curriculum.

As looking into
When analysed on the aspect of statistical process components and related levels of learning outcomes in Table 3, there are more learning outcomes at Level A and B at every component in general. Moreover, the higher amount of learning outcomes regarding the component of data analysis is pointed out. The number of learning outcomes regarding this component in curriculum of all countries to be spared is noted to be higher. The number of learning outcome for this component is 66% of total number of learning outcomes in all components. It has been shown that data analysis has the highest learning outcomes reflecting each level in all other components. It is noted that all countries have more learning outcomes at Level B in data analysis and Singapore is prominent on this component. Singapore includes more learning outcomes regarding this component with the ratio of 32% in comparison of other countries and 83% of them belong to the component of data analysis. Furthermore, 46% of them are at Level B, while Singapore has the highest percentage (17%) in terms of learning outcomes at Level C as well. For example, in Singapore, understandings related to inferential statistics subjects such as sampling methods, sampling distribution, confidence interval, hypothesis testing, correlation, regression and least squares method are prominent regarding Level C in data analysis in H1, H2 and H2F mathematics curriculum which is presented as elective course during the period following the secondary education and pre-university. The subjects included in the curriculum of the countries, grade levels that these subjects and a total number of learning outcomes within these grade levels are given in Appendix 2. Besides, learning outcomes such as H4.4 (meaning of correlation), HF1.6 (confidence interval), H3.3 (hypothesis testing) are regarding subject and concept of inferential statistics (look at Appendix 1- Table 6). Turkey is shown to be prominent on the aspect of including learning outcomes related to Level A in data analysis. In fact, learning outcomes related to this component and level in curriculum of Turkey are given much more weight in comparison of other countries (44%). However, it is detected that subjects of inferential statistics or any learning outcomes that can be a basis for these subjects are not included in Turkey. Thus, it is shown that no learning outcome at Level C is available in curriculum of Turkey. Despite of the fact that Korea, USA and New Zealand have more learning outcomes in number regarding the component of data analysis, learning outcomes related to Level C emphasizing on inferential statistics such as normal distribution, correlation, regression, hypothesis testing and confidence interval are noted to be included finitely in these countries. For example, in Korea, subjects of inferential statistics are limited with understanding of normal distribution by including only L.9.5 (meaning of normal distribution) regarding inferential statistics at level of high school (look at Appendix 1 - Table 9). Besides, learning outcomes related to question formulation which is the most important component for initiating procedure of statistical research and planning correctly are found to be included in less ratio in all countries. Thus, it is noticed that the ratio of learning outcomes regarding this component (6%) is the least in all components. As Korea doesn't have any learning outcomes regarding this component, few learning outcomes in Level A and B are included in Singapore, USA and Turkey. Having limited learning outcomes regarding this component in these countries and not including learning outcomes concerning Level C which questions further understanding is thought as limitedness for the component of question formulation to be the first step of statistical process. Thus, as New Zealand includes more space for the component of question formulation with a ratio of 67%, learning outcomes reflecting Level C are found to be included in this country. Many learning outcomes regarding question formulation such as formulating the problem by using statistical process, formulating research questions based on relating and comparing and searching different and independent problems are found in New Zealand. Additionally, learning outcomes regarding component of the interpretation of the results which enables students to infer and think critically are fewer in all countries and this component is seen to follow the component of formulating question. It is noted that the ratio of the number of learning outcomes regarding this component in all components is 13% and these learning outcomes are centered upon in Level B in general. In comparison of other countries, USA includes more learning outcomes regarding component of interpretating results with a ratio of 42%. USA gives more emphasis on learning outcomes regarding this component on the aspect of Level A (11%) and Level B (20%) as compared to other countries. In USA, these two levels in interpretation of results are presented with subjects such as summarization of data, interpretation of table and graphs, and basic inferring based on correlation coefficients. For example, HSS.ID.C.8 (interpretion correlation coefficient) learning outcome is shown to be related to Level B on the aspect of interpreting the strength of relation between specified variables and putting forth situation of variables according to each other; HSS.IC.B.6 (interpretation of data) learning outcome requires interpretations at basic level to be expressed by giving obvious data set and it is in accordance with aspects of Level A (look at Appendix 1- Table 7). Whereas no learning outcome concerning this component is included in Korea, interpretation of results is limited with Level A in Turkey.. As USA does not have any learning outcomes at Level C concerning this component, New Zealand spares 7% of its curriculum for Level C. For example, as different from USA, L5.3.1.1 (comparison of sampling distribution) learning outcome includes objectional behaviour concerning Level C regarding sampling distribution. Inferential statistics such as estimating situations that may occur within the separate data set by using these data set (interpolation) and making forward looking inference outside of given data set (extrapolation) are included more in detail. As a result, learning outcomes related to Level B and C such as making prediction, detecting sampling error and estimating for average by central limit theorem are more in number. Another example, as L7.2.1.2.b (estimation of population parameter) learning outcome relates to Level C requiring beyond data literacy and generalisation, it includes L7.2.2.1.b (detection of sampling error) learning outcome associated with Level B in which comparisons among groups are made dominantly in making sampling to reflect a certain statistical situation as well (look at Appendix 1- Table  8). On the other hand, the number of learning outcomes at all three levels is almost equally included at the component of data collection.. In this sense, distribution of learning outcomes to all levels is the most balanced at the component of data collection. Learning outcomes belonged to this component have 14% percentage within the statistical process components. Whereas no learning outcome reflecting Level A in this component is found in USA, any learning outcomes related to Level C regarding this component are not found in Turkey and Korea. Singapore represents Level C with only one learning outcome H2.1 (choosing random sampling) included in H2 mathematics curriculum and in which the relation of population and sampling is evaluated in the focus of concept of randomness (look at Appendix 1- Table 6). However, New Zealand gives more emphasis on the component of data collection with the ratio of 53% in comparison of other countries. It is prominent on the aspect of reflecting this component with learning outcomes at Level A and B as well. It is shown that the importance given to both these levels within the component (B: 9% and C: 8%) is more than other countries. New Zealand gives subjects and concepts that are to set an example to Level B and C such as data set with many variables, random choices, representability, limitedness of choices of sampling and experimental designs. For example, 8.3.1.1 (producing data sets) learning outcome is addressed as related to Level B due to this learning outcome requiring new data sets to be produced as result of relating many variables; L7.2.1.1.a (experimental designs based on random sampling) learning outcome reflects Level C on the aspect of making sense of randomness concept (look at Appendix 1- Table 8). Tables belonging to each country presenting distribution of learning outcomes given as frequency and percentage in Table 3 on the aspect of contents are given in Appendix 1. The number of learning outcomes is indicated in the tables and access addresses of websites covering their contents are given.

Discussion and Conclusion
The components, starting with formulating a problem, continuing with collecting data, analysingrepresentation data and inferring from data forms the basis of statistical literacy (Watson, 2006). It is also stressed that handling the components of the statistical process with a holistic approach is significant in students' experiencing more effective statistics teaching (Newton, Dietiker, & Horvath, 2011). In this way, handling the statistical process, that helps us to provide active participation and more effective management of the process, with a holistic approach (following each component hierarchically) has an important role. It is not enough to integrate our statistics teaching activities related statistical process, but it is also important to design learning environment with these activities in a way that will provide a basis for effective and permanent statistics teaching (Franklin et al., 2007). Besides, analysis based on the theoretical framework pointed out in GAISE report has an important role on determining the tendency of the curriculums that forms the statistics teaching. In this way, the statistics learning outcomes in mathematics curriculum of countries were analysed through the reflection of the statistical process which is seen as an important for GAISE report.The research results have shown that number of the statistics learning outcomes in math curriculum and the level of these learning outcomes in Turkey was lower than other countries. Although the learning outcomes are only related to Level A and B in our country, other countries have more learning outcomes related to Level B and C. Especially Singapore and New Zealand come to the forefront in this sense. It is also thought that this result could reflect the fact the achievement of PISA and TIMSS exams for these countries. Nevertheless, it is an important fact that Turkish students have lower success compared to other countries for the problems requiring determining the literacy, reasoning levels related to the field of statistics learning (PISA, 2015;TIMSS, 2015). One of the reasons for this failure could be raised to give more emphasis on basic statistics knowledge rather than statistical literacy at statistics subjects and learning outcomes in our curriculum. It is suggested that more emphasis on Level B and C learning outcomes and inferential statistics subjects should be place in our mathematics curriculum like other countries. Although there is an increase about the research on statistical literacy, reasoning and thinking in our country, a common result of the research is about the insufficiencies of the students, teachers or individuals in terms of the statistical literacy (Koparan, 2012;Özmen, 2015;Reston 2005). In this way, these results will only make sense if there is an attempt to enhance the curriculum and later statistics teaching. Otherwise, research would be limited to picture the current situation. In this way, studies about incorporating the learning outcomes related to the higher level emphasized in the GAISE report would be an important attempt.
When the distribution of the learning outcomes related to the field of statistics learning according to the statistical process components across countries is examined, it is seen that there is less emphasis on the components of formulating the problem and the interpretation of the results, and more emphasis on the component of data analysis. The nature of the data analysis requires to summarize the data set by the help of the different measurements and representations. This could be effective on this result. Similarly, research on statistical process, data analysis and representation component are forefront whileformulating the problem and interpreting the data components are generally ignored (Koparan, & Güven, 2013;Money, 2002). In this way, carrying out research on these components and making necessary improvements on curriculums depending on this research results should be an important attempt. Representation of the data with the proper way and analysing data following the appropriate methods are important aspects of statistical literacy (Steen, 1999). But it is also of great importance to formulate an appropriate problem that will start the process for resorting to such representations and analyses and to critically evaluate and interpret the results. On the other hand, formulating the problem is the first component of the statistical process and has a crucial importance to continue the following steps correctly. Therefore, having not enough knowledge about this component or being unsuccessful at formulating an effective problem would affect continuing components. In this way, learning outcomes related to formulating the problem and interpretation of what kind of results would emerge on referring problems fit or not with the purpose should integrate the curriculums. Otherwise, interpretation of the results is an important component for evaluating the obtained results, the failure about this component could be a barrier to discuss on different ideas or interpretations, and this would also negatively affect the cyclical structure of the statistical process. The fact that the number of learning outcomes included in the mathematics curriculums of countries is not considered to mainly reflect each component of the statistical process may cause obstacles to the full understanding of these components by students. It will also raise problem about the experiencing the statistical process, is one of the important aspects of the statistical literacy. Therefore, giving place learning outcomes for each component of statistical process in our mathematics curriculum is an initial attempt to overcome these problems.
Differences about the number and content of the learning outcomes across countries are seen one of the important factors for determining the statistical knowledge and skills of the students. For example, statistics learning outcomes in Singapore mathematics curriculum are more than and based on the structure that is starting with basic concepts and extending inferential statistics, on the other hand the number of the statistics learning outcomes in our mathematics curriculum is less and related with lower levels. It is thought that these differences are effective on determining the success of the countries for the statistics field. Following a structure from basic concepts to inferential statistics and having enough learning outcomes for this structure should be an important factor to obtain important outcomes and to emerge different achievement level between countries. Thus, it is also thought that students, raising as following this kind of structure, have higher skills as reasoning, argument, critical thinking. For example, raising students as having the enough knowledge at the concept of the relationship between variables at school level, it could be helpful to interpret and to make sense of the relationship between variables related to correlation and regression subjects at university level. When students have enough knowledge about the big ideas (significant difference) as a basis for further concepts such as confidence interval and hypothesis testing, would be helpful at making the meaningful understanding related formulating the hypothesis, testing this hypothesis with a confidence level and justifying about the hypothesis. Therefore, it is important for individuals to take decisions related their daily and professional lives through statistical process rather than leaving these decisions to chance. It is also stated that students who can handle their statistical knowledge as embedded in the components of the statistical process will also be trained as a good problem solvers (Neumann, Hood, & Neumann, 2013). In this sense, giving place on the learning outcomes as a basis for inferential statistics in the field of statistics learning in the mathematics curriculums would be an important attempt for statistics education.
In our country, dealing with statistics in a narrower scope in the mathematics curriculum is seen as the biggest precursor of the difficulty for students, to be able to statistically evaluate the situations that they will encounter in their life. Statistics, which is one of the leading roles of today's world surrounded by data involves much more than basic concepts such as average, standard deviation, graphs (bar, line, pie chart etc.). In this way, it is also important to design statistics teaching as building on statistical process and concentrating on statistical literacy, thinking, and reasoning skills to provide more qualified and effective teaching related to the field of statistics learning. However, number and content of the learning outcomes are not enough potential to meet the requests regarding the field of statistics learning in our mathematics curriculum. Thus, it is stated that it will be difficult to raise students as productive consumer of the statistical data by the help of the current learning outcomes in our curriculum, mostly Level A and B (Koparan, 2012;Yolcu, 2012). In order to cope with these difficulties, it is suggested to increase the number of the learning outcomes and to be arranged their content that will allow to the development of skills such as reasoning, argument and critical thinking as referring to basic knowledge and ideas on statistics. Besides, research on statistics education also stressed that the quality of statistics teaching should be enhanced, and skills considered important should be developed for students (Akoğlu, 2018;Chick, & Pierce, 2008;Garfield, & delMas, 2010). It is obvious that the importance of including learning outcomes for Level C in the development of necessary knowledge and skills expected from students. It is not possible to expect to raise students as equipped with the statistical literacy and thinking skills required by our age by the help of the current mathematics curriculum which has limited with Level A and B. In this way, to achieve the goal of raising students as statistically literate it is suggested to enrich the curriculum with the learning outcomes build on each component and level. It will be an important attempt to arrange our curriculum in this direction for the goal of raising more qualified individuals after statistics teaching. To raise students as equipped with the statistical literacy, thinking and reasoning skills, the importance of the proper statistics teaching is mentioned (Callingham, & Watson, 2017;Groth, 2017;Özmen, & Baki, 2017). Besides curriculum plays an important role on designing the mathematics teaching, classroom practices. As a matter of fact, although curriculum involves the higher-level topics and learning outcomes, it is the teacher who will apply it. This also points out that attention should be drawn to statistics teaching. Besides, limited topics and practices related to the field of statistics learning, it is also stated that teachers are inadequate or feel inadequate themselves for statistics teaching. In this sense, firstly, it is recommended to review and enhance the field of statistics learning in the mathematics curriculum in our country. Later, in line with the renewed curriculum, studies can be carried out to improve statistics teaching and, accordingly, to examine the outputs in line with the ultimate goals for statistics education.
There are some limitations in this study, which aims to compare mathematics curriculums of different countries in terms of the field of statistics learning. Present study, learning outcomes in the field of statistics learning in mathematics curriculums of countries were analysed through the theoretical framework presented in GAISE report based on the statement and their explanations mentioned in the curriculum. Due to this limitation, it is not mainly understood that how the statistics teaching design in these countries. In this context, it is suggested to carry out research on examining the textbooks or observing statistics teaching in learning environment for these countries and to make more detailed comparisons. Moreover, reviewing our curriculum based on these research results will also contribute to obtain more effective learning outcomes.      L.9.5.

INTERPRET RESULTS
---APPENDIX 2  As examining Table 10, concepts regarding measures of central tendency and spread and table and graphs are included in the countries, yet Turkey is detected to be at more inadequate level on the aspect of concepts given at both categories in comparison of other countries. Particularly, learning outcomes that meet variance, quartiles and percentiles as measures of spread and dot, box plots, stem and leaf graph, scatter plot in table and graphs are not seen in Turkey. However, learning outcomes reflecting these concepts are seen in other countries. For example; in the process lasting from primary school to high school, learning outcomes reflecting Level A, B and C based on forming and interpreting table and graphs are included; for Singapore and USA regarding dot plot, for Singapore, USA, and New Zealand regarding box plot and scatter plot and for Singapore, USA, Korea and New Zealand regarding stem and leaf graphs. These kinds of graphs that are included in the least two of the countries are not found in Turkey and pictogram, bar, line, pie and histogram graphs are commonly used for presenting data visually. Tally and frequency tables are often applied for representation of data in comparison of other countries. As in the table, mean for measures of central tendency and spread and bar graph for table and graphs according to other subjects and concepts are more commonly used. Besides, grade levels including these concepts are shown to be varied. As concepts of median and mode are started to be given in secondary grades in other countries, these concepts are started to be given from 9 th grade in Korea. Similarly, concept of histogram is started to be given in secondary school in other countries as this concept is among learning outcomes at the level of high school in Turkey. As subjects and concepts regarding inferential statistics such as parametric and nonparametric tests, distributions, correlation and regression, hypothesis test and confidence interval are put more emphasis in Singapore, only normal distribution from these concepts are addressed in Korea, these concepts are shown to be limited on aspect of number of learning outcomes and content of the subject in New Zealand and USA Inferential statistics which is touched as only one subject or with limited learning outcomes in most of the countries is not included in any way in Turkey. In other words, providing subjects and concepts of inferential statistics to students is not included in objectional behaviour of mathematics curriculum in Turkey. Besides, when taking mathematics curriculum depending on grade levels through the countries into consideration, concepts such as frequency table, pictogram and bar graph, mean, mode and median are touched on in low levels and they are switched towards the concepts of standard deviation, line and pie graph, histogram, parametric and nonparametric tests, distributions, correlation and regression, hypothesis test and confidence interval in higher levels. In this sense, principle of simple to complex applied in mathematics curriculum in all countries is seen to be common, yet addressing level of the subjects depending on grade levels for every country can be varied.