Author name:
هشام جمعة صويح شامخ
Supervisor name:
محمد انور محمود السامرائي
Abstract:
يهدف البحث الحالي الى اجراء مفاضلة بين اساليب انتقاء الفقرات الاختبارية لمقياس التحكم المدرك لدى طلبة الجامعة في الخصائص القياسية وفقا لنظرية المنحنى المميز للفقرة. وقد تم تحقيق هذا من خلال هدفين رئيسين. تمثل الهدف الاول في بناء ثلاث صور من مقياس التحك | The current research aimed to make a differentiation between selection test items for perceived control scale among university students in the standard characteristics according to theory of Characteristic item curve. This has been achieved through two Aims. The first Aim is to build three forms of perceived control scale according to three items selection test methods (method based on theoretical framework, method based on synthesis - logical and method based on factor analysis).The researcher followed scientific steps to achieve these aims, according to the theory of responding to item, starting with the first and second scale (theoretical and synthesis - logical) instructed of these items displayed on expert authorities in Educational and Psychological Sciences, to make sure that these items have appropriate characteristics availability of items scale and alternatives, in form and content, and that it describes the measured characteristic based on definition that was specified for each measure separately, and after making some necessary adjustments, all of two scales promised to measure what was good for in order to measure them. And to ensure the clarity of items and instructions to our research sample, and to determine the time required for performance, the two scales applied to pilot sample of (104 students) from outside statistical analysis sample, and appeared that all instructions and items of each scale were clear. In order to determine the standard characteristics of the two measures and there items, as well as verification of model assumptions, the scale was applied to a sample of (1,100 male and female students) from university students in morning studies in city of Baghdad, which they were selected in a random manner with proportional stratified random sample. The researcher adopted Rating Scale Model originating from a single - parametric model, using a statistical method called (ConstructMap.4.6), resulting in two scales contain (39 and 36) items respectively. And to build a third form for perceived control scale by factorial analysis method, the researcher translated two foreign scales and displayed to expert authorities in Educational and Psychological Sciences to show their opinion about items and alternatives, in terms of form and content, and appropriateness for the local environment, and in terms of classification for measured feature on the basis of the definition that was specified for each unit measured, and with some minor linguistic on two scales on a sample to account time and instructions clarity, and then two scales applied to the same research sample, collected data answers for four scales (theoretical scale and synthesis - logical scale and the two foreign scales) to analysis in a factorial analysis method by using statistical bag of Social Sciences ( SPSS), and to draw the third form of the scale and factorial analysis under controls of using this method, which had produced a scale consists of (60) items which adopted to appropriate assumptions to analysis model. With this the standard features have been identified, to all items and to total scale for the three images to measure perceived control, and using the same statistical method. The second aim was to make differential comparisons between these three forms for measure perceived control in standard characteristics. These procedures showed in accordance to comparisons as follow : Suitable for adjusted estimate for progressive answers in building perceived control scale. And then retained items and individuals that achieved the model assumptions, and overall items and scales had good standard characteristics. For a comparison between three scale forms, it started with (information function and standard error) which showed that all scale forms were close, and results did not showed wide differences between them. So the Tau coefficient of progressive answers difficulty values revealed that three scales forms were close, and there were no significant differences statistically between them after using Kroskal - Alice equation. For property of stability, results showed, that there was a convergence and divergence in this property also, For stability based on information function, and the separation between individuals, there was a convergence in all scale forms, but for the stability that based on calculates variance between expected and observed individuals responses, which calculated through the statistical method, the form of the factorial scale characterized by consistently was higher than other tow form scales. Results from reliability property showed that there was a convergence and divergence of this property, there was a convergence and uniformity in reliability property function represented by accuracy of estimation factor, and there were no statistically significant differences between the three forms of perceived control scale in accuracy of estimation factor based on characteristic test Function. While there were significant differences between forms of scale in sincerity of selection of behavioral scope which objectively interpreter by variance , as it shows there was a statistically significant differences in favor of higher variance ratio for scale of factorial forms of scale, which offers highest ratio of interpreter variance, followed by a theoretical scale and finally synthesis scale.In summary, researcher can judge by convergence of the three forms of scales in standard characteristics, for both items and overall scale. Despite appearance of some statistical differences in factorial method of some of standard characteristics in current research. But cannot resolve the priority by other test methods to prefer some items on another. According to current research results, the researcher reached some conclusions, recommendations and proposals
