The present work was carried out as part of a series aimed at identifying teachers’ ideas on assessment. The two studies presented examine teachers’ ideas about how text, table and graph understanding should be assessed to help students’ learning. First, the psychological models from which teachers’ ideas should be evaluated were established after a careful review of literature on the nature of each competence to be evaluated. Second, 68 secondary school teachers -38 from Spain and 30 from Brazil- were asked to design assessment situations adequate for the mentioned purpose. Results showed convergence of data coming from both countries, enlightened specific strengths and shortcomings of teachers’ ideas on assessment and are useful for designing teachers’ training programs on students’ assessment.

            The main objective of the present work was to identify strengths and shortcomings in social science teachers’ ways of assessing specific understanding of information contained in different social sources and documents –texts, graphs, objects, etc.-.

            There is some evidence that the preferred ways of assessment of many teachers have many shortcomings (Black & William, 1998; Segers, Dochy & Cascallar, 2003). Assessment tasks are often not valid criteria of understanding and cognitive ability, as they rely mainly on root learning (Crooks, 1988; Villa & Alonso-Tapia, 1996). In fact many tasks do not allow teachers to identify the source of students’ difficulties, and so feedback lacks the necessary quality to be of help for overcoming learning problems (Black & William, 1998; Birenbaum, 2003). In a recent paper Birenbaum et al. (2006) point to the shortcomings of current assessment practices and suggest alternatives directions to overcome them. These directions imply that assessment must be multidimensional, formative, integrated into the curriculum, based on authentic problems, context embedded and flexible. However, even if teachers follow these directions and adopt new ways of collecting evidence –for example, performance tasks, learning journals, portfolios- whose positive effects have been well documented (Alonso-Tapia, 1997, 2002; Birenbaum, 2003; Davies & LeMahieu, 2003; Duschl & Gitomer, 1997), feedback may be of low quality if the specific tasks whose solution teachers consider as adequate criteria of understanding and cognitive ability are not adequate, no matter whether these tasks are used in the context of exams, performance tasks or of portfolio assessment. If this was the case, it would be important to identify the specific strengths and shortcomings in teachers’ ideas about what constitutes a valid indicator of understanding or of the cognitive abilities to be assessed in the context of different subject matters in order to design adequate instruction programs for improving teacher preservice and professional development. With this purpose our team has started a research program whose present research focus is the study of social science teachers’ ideas on knowledge assessment.

            The present paper shows the results of two of the studies carried out with this focus. They are centred on teachers’ ideas on how information contained in different kinds of document –text, table and graphs- should be assessed.

            According to explicit objectives of the Spanish & Brazilian curriculum, the study of social sciences should make students able to understand information contained in different kinds of document and to build conceptual and causal models useful for assessing critically present social problems. Nevertheless, the mentioned abilities are not assessed (Villa & Alonso-Tapia, 1996). This might be due to teachers’ practical considerations when designing assessments tasks, but also to the fact that their ideas about what constitutes a valid indicator of each abilities being assessed are not adequate. So we decided to analyse this possibility.

            The first step was to decide the criteria from which teachers’ answers had to be evaluated. According to psychology, the ability to understand and extract information from different kinds of document is shown when assessment tasks meet certain criteria. For example, it is not possible to identify clearly text comprehension at some levels of representation –the level of the propositional and situation models (Kintsch, 1998)- if students are asked only to explain the meaning of words contained in the text, as it often happens. It is necessary to ask the students to make some kind of summary of its meaning and to use text information for solving problems of some kind. That is, assessment tasks need to meet criteria defined by psychological models of mental representations and processes to be valid indicators of ability, understanding and reasoning. This is especially important if assessment is aimed at identifying the kinds of help that should be given to students in order to develop their potential for meaningful learning. So, we reviewed literature un text, table and graph understanding in order to establish the necessary criteria.

       The second step was to identify teachers’ ideas about what constitutes a valid indicator of each abilities being assessed. With this purpose, two similar studies were designed and carried out in Spain and Brazil, using the same sample and methodology.

            In each study, teachers were given different kinds of document -texts, graphs, tables – related to the French Revolution, and asked to design a task for assessing whether students understood it or not.

68 subjects, 38 Spanish Secondary and High School teachers (17 men and 21 women, mean age 40.39) and 30 Brazilian teachers of the same levels (15 men and 15 women with ages lower than 40 in all cases) that volunteered to participate in the study after being asked to do it,  formed the sample of the study. Teachers’ answers were coded using two groups of category. The first group was derived from psychological models of each one of the capacities being assessed. As for the second, it was created inductively in order to describe the characteristics of tasks that did not correspond to any of the categories of the first group. Then, the proportion of teachers suggesting tasks falling in each category was calculated and analysed, showing specific strengths and shortcomings of teachers´ ideas on what can be considered valid indicators of ability, understanding and reasoning. Chi-square analyses of differences between category use and between Spanish and Brazilian teachers were carried out.

            Findings, that will be presented for discussion, showed convergence of data coming from both countries -Brazil and Spain-, and enlightened specific strengths and shortcomings of teachers’ ideas on assessment of text, table and graph understanding.

            Results have theoretical relevance for understanding teachers’ ideas on assessment and also practical relevance for designing teachers’ training on students’ assessment. Each kind of relevance will be discussed.