1 Education-line document submission Author: Chraska, Miroslav Title: Evaluation of noncognitive areas in education Paper presented at the European Conference on Educational Research, Edinburgh, 20-23 September 2000 Abstract: We can evaluate education outputs (with certain simplification) with help three principal indicators: pupil´s knowledge, pupil´s skills (practical even intellectual) and attitudes to learn and education. Evaluation of outputs of education in Czech republic at present is based especially on assessment results in cognitive area, otherwise on assessment of quantity and quality of acquire knowledge and skills. However, that evaluation will not sufficient in future. Owing to living conditions, witch are changing very quickly, school can not give to pupils all knowledge and skills, witch pupils will need during their life. It is desirable, in order to school would educate individuals, which are able and which are helpful adapt to quickly changing living conditions, individuals, witch will be able to learn, obtain new knowledge, skills or new qualifications. In that context will be more important, so that school would form positive attitude to learning, education, school, particular subject etc. (we talk about attitudes to educational reality). We developed in Faculty of education Palacky University in Olomouc (on the basis of semantic differential) special tool ATER (questionnaire for measurement of pupil,s attitudes to educational reality) for measurement that important effects of education. Document type and origin: presented at a European Conference on Educational Research, University of Edinburgh, 20-23 September, 2000, Scotland Suggestied key terms: education outputs, attitudes to educational reality ,semantic differential, factor analysis Version Number: first 2 EVALUATION OF NONCOGNITIVE AREAS IN EDUCATION Miroslav Chraska Palacky University Olomouc, Czech Republic Introduction We can evaluate educational outputs (with certain simplifications) with the help of three principal indicators: pupils’ knowledge, pupils’ skills (practical and intellectual) and attitudes to learning and education. Evaluation of the outputs of education in the Czech Republic at present is based especially on assessment results in the cognitive area, otherwise on the assessment of quantity and quality of acquired knowledge and skills. However, that evaluation will not be sufficient in the future. Owing to living conditions which are changing very quickly, schools cannot give pupils all the knowledge and skills which they will need during their lives. It is desirable for schools to educate individuals who are able to adapt to rapidly-changing living conditions, individuals who will be able to learn and obtain new knowledge, skills and qualifications. In that context it will be more important for schools to form positive attitudes to learning, education, school, a particular subject etc. (We are talking about attitudes to educational reality). We have developed, in the Faculty of Education of Palacky University in Olomouc (on the basis of semantic differential) a special tool known as ATER (a questionnaire for measurement of pupil's attitudes to educational reality) to measure these important effects of education. Attitudes to educational reality We define the attitudes to educational reality as pupils' or students' inclination to react in a stable way to individual aspects and situations of educational reality and towards oneself (comp. Hartl, 1996). The measurement of pupils' attitudes or students to educational reality is rather hard because we are dealing with a very tenuous variable which is hard to cover. Traditional scales of attitudes are not quite suitable for routine diagnostics in practice at schools and, what is more, they provide us with information where attitudes are assessed only in one dimension (more or less expressing assessment of the particular element of educational reality by the person researched). On the 3 basis of the research carried out we argue that it is possible to measure attitudes towards educational reality rather precisely by means of tools created on the principle of semantic differential. Qualities of measurement by means of an ATER questionnaire Measurements' credibility through semantic differential more or less depends on the quality of the scales used for the measurements. We carried out research in 1998 - 1999 in which we appointed ourselves three essential goals: to verify whether the suggested scales of the ATER questionnaire have adequate construct validity, to verify whether measurement by the ATER questionnaire is sufficiently reliable and accurate, to propose an acceptable method for the presentation and interpretation of the results attained. a) Construct validity of scales of the ATER questionnaire We verified the validity of scales by means of a sample of 523 fifteen-year-old pupils, who were in their last year of compulsory school attendance. The research sample included 205 pupils who were finishing their compulsory school attendance in the ninth class of ”basic school” (a school for the 6-15 age group) and 218 pupils of the same age who were finishing their compulsory school attendance at a ”gymnazium”, which more or less corresponds to the traditional grammar school. The pupils were presented with 10 term indicators (Table 1) with 10 rating scales evaluating from one to seven. Table 2 shows an example of the term indicator HOME STUDY and the scales which was used. Table 1: Term indicators for measurement of attitudes to educational reality HOME STUDY OUR SCHOOL STUDY IN SECONDARY SCHOOL LEARNING POPULAR TEACHER UNPOPULAR TEACHER SCHOOL REPORT EDUCATION EXAMINATION 4 MARKING Table 2: Recording sheet for the ATER questionnaire HOME STUDY 1. good bad 2. demanding undemanding 3. unpleasant pleasant 4. light dark 5. strict lenient 6. simple complicated 7. beautiful ugly 8. problematic easy going 9. sweet sour 10. difficult easy For the analysis of the quality of the designed scales the results of all respondents in all term indicators were used, i.e. 5230 ways of usage were analysed altogether. First the correlation among the results in individual scales was calculated and then these results were subjected to traditional factor analysis. We discovered that the results of the measurements are influenced the most by the first two common factors, which together clarify nearly 70% of the discovered variability. With regard to this result of the factor analysis we decided to extract only two common factors. Table 3 shows rotating factor matrix (rotation VARIMAX) with factor loadings (factor weights) and communalities of individual scales. The first factor has dominant factor loadings in the scales DEMANDING – undemanding, STRICT – lenient, DIFFICULT – easy and PROBLEMATIC – easy going (dominant factor loadings are showed in table 3 in bold). These scales imply to what extent the respondents understand the term indicators as ”something” connected with labour, difficulties, changes or activity. That is why we indicate this factor as energy factor. The scales GOOD – bad, PLEASANT - unpleasant, LIGHT – dark, BEAUTIFUL – ugly, SWEET – sour express to what extent the respondents positively 5 evaluate the given term indicator (to what extent it is ”good” or ”bad” for them). Factor delimited by the stated scales is referred to, in accordance with C. Osgood, as assessment factor. Table 3: Rotating factor matrix Scale number 1 2 3 4 5 6 7 8 9 10 Scale good - bad demanding – undemanding pleasant – unpleasant light – dark strict – lenient complicated - simple beautiful - ugly problematic – easy going sweet - sour difficult - easy Energy factor 0,37296 0,83483 0,36803 0,18782 0,77844 0,83082 0,25093 0,64433 0,21974 0,78108 Assessment Communality factor h2 0,74131 0,68864 0,14148 0,71696 0,74483 0,69022 0,79144 0,66165 0,27109 0,67947 0,30029 0,78043 0,83103 0,75358 0,37107 0,55285 0,76247 0,62964 0,32978 0,71884 The factor loadings of the scales are represented in picture 1 as dots in a two-dimensional graph. It is clear from the picture that 5 of the scales have dominant factor loadings in the energy factor while the remaining 5 scales have dominant loadings in the evaluation factor. By means of confirmative factor analysis we verified the hypothesis that it is possible to describe pupils' attitudes to educational reality through two common factors between which some correlation does exist (we talk about so-called oblique factor rotation). The probability of this hypothesis comes out very high (P=0,98) and the average residual correlation is very low (0,029). b) Reliability of measurement by means of the ATER questionnaire We were looking for an answer to the important question of to what extent the results obtained are influenced by accidental factors and consequently what the reliability of our measurement is. We evaluated reliability by means of methods based on the analysis of variance. All the models of calculations used get very similar results. It shows a rather high rate of reliability and accuracy of performed measurement (calculated coefficients of reliability attain universally higher values than 0,85). c) Presentation and interpretation of attained results of measurement In order to be able to express pupils’ attitudes to the whole of their educational reality, we defined for both factors summary indexes. 6 Assessment of attitudes to educational reality index I assAER indicate, to what extent the pupils perceive term indicators as "good" or "bad" for them. I assAER 1 n assi n i 1 Energy of attitudes to educational reality index I enAER indicate to what extent the pupils perceive term indicators as "something" connected with labour and energy. I enAER 1 n eni n i 1 In the above formulae n is the number of term indicators, assi are assessment factors for particular term indicators and eni are energy factors for particular term indicators. Some results obtained by means of the ATER questionnaire By means of the ATER questionnaire we obtained an amount of important and interesting data about the examined population of fifteen-year-old pupils. For example, one interesting finding is that significant differences exist between attitudes to educational reality at pupils in the ninth class of ”basic school” and students in the “gymnazium”. If we compare the results obtained in the ”basic school” and in the ”gymnazium”, we find out that there is higher assessment factor at almost term indicators in the basic school (differences are statistically significant). This result can be interpreted that pupils in ”basic school” (who are finishing their obligatory school attendance) have more positive attitudes to school and all educational reality than students of the same age in grammar school. If we compare values of energy factor in the ”basic school” and in the ”gymnazium” we find out that are higher energy factors in the latter. Picture 2 and 3 display the results of our research in a graphic way. While we expected such results in the energy factor, results in the assessment factor are very surprising. We cannot overestimate the results of our research (mainly owing to our small sample size) but we can raise some alarming questions e.g. "does the Czech ”gymnazium” create suitable conditions for the development of pupils at present?", "What is the real level of educational work in this type of school " ? etc. 7 8 Conclusion On the basis of our research we came to the conclusion that the measurement of attitudes to educational reality through semantic differential in the form in which it is described by C. Osgood (e.g. Osgood at all., 1957) and by other authors does not suit the sociocultural conditions in Czech schools and provides us with a rather distorted picture of reality. The differences between Osgood's findings and ours are mainly in the interpretation of the discovered common factors. We came to the conclusion that it is reasonable and profitable to extract only two common factors. The tool which measures pupils' attitudes to educational reality by means of two common factors is what we describe as the ATER questionnaire. Bibliography BARCHARD, K. A. - HAKSTIAN, A. R. The Effects of Sampling Model on Inference With Coefficient Alpha. Educational and Psychological Measurement, roc. 57, 1997, c.6, , s. 893 - 905. BINTIG, A.: The Efficiency of Various Estimations of Reliability of Rating Scales. Educational and Psychological Measurement, roc. 40, 1981, c.3, , s. 619 - 642. CRONBACH, L. J.: Coefficient alpha and internal structure of tests. Psychometrika, 16, 1951, s. 297 - 334. HARTL, P. Psychologicky slovnik. Praha: Nakladatelstvi Tercie, 1996. CHRASKA, M. Jake jsou postoje zaku a studentu ke skole a edukacni realite ? Pedagogika, 48, 1998, c. 1, s. 54 - 66. CHRASKA, M. Postoje studentu ucitelstvi a ucitelu v praxi k ucitelske profesi. (Zaverecna zprava vnitrniho grantu Univerzity Palackeho.) Olomouc: Pedagogicka fakulta UP, 1996. CHRASKA, M. Odhady reliability mereni semantickym diferencialem. In. Sbornik prispevku z 6. konference Ceske asociace pedagogickeho vyzkumu. Ceske Budejovice: Pedagogicka fakulta Jihoceske univerzity, 1998, s. 46 – 53. KERLINGER, F. N. Zaklady vyzkumu chovani. Praha: Academia 1972. MARES, J. Jak zjisťovat reliabilitu pozorovani ? Pedagogika, 33, 1983, c. 2, s. 169 189. McDONALD, R .P. Faktorova analyza a pribuzne metody v psychologii. Praha: Academia, 1991. OSGOOD, C., SUCI, G., TANNENBAUM, P. The Measurement of Meaning. Urbana, III, University of Illinois Press, 1957. 9 PAPICA, J. Zaklady psychometrie. Olomouc: Filozoficka fakulta Univerzity Palackeho, 1984. PRUCHA, J. Moderni pedagogika. Veda o edukacnich procesech. Praha: Portal, 1997. Contact address: Assoc Prof Miroslav Chraska Palacky University in Olomouc Czech Republic phone 42068/5635154 E-mail: capv@risc.upol.cz