Development of a writing skill test for non-Persian learners: Approaches and analysis of errors

In recent years, many advances have been made in teaching Persian to non-Persian learners. However, there are still some shortcomings one of which relates to the assessment of Persian language skills. The writing skill involves the ability to create words and draw ideas from the mind as well as the ability to criticize them so that the learner could decide which one to use. Thus, the assessment of writing skill requires productive and product-based tests. This study aimed to investigate writing skill assessment approaches; written performances such as imitative, intensive, responsive and extensive writing; and a variety of writing questions such as basic, framed, and text-based. Considering the influence of the learners’ systematic errors on the organization of instructional materials and the selection of instruction and assessment methods, 20 syntactic errors which were observed with high frequency in learners’ writing samples were analyzed. These errors included the deletion of grammatical and lexical elements, the addition of unnecessary or incorrect elements, the selection of an incorrect element and using it in place of the correct element, and the incorrect ordering and arrangement of elements. In deciding on the organization of instructional materials and the selection of instruction and assessment methods, these errors need to be considered. 

Expanded Abstract
Considering the many advances that have been made in teaching Persian to non-Persian learners in recent years, there are still some shortcomings one of which relates to the assessment of Persian language skills. Writing skill involves the ability to create words and draw ideas from the mind as well as the ability to criticize them so that the learner could decide which one to use. Writing tests involve productive skills. Therefore, this article attempted to investigate written skill assessment and performance such as imitative, intensive, responsive and extensive performance, and a variety of writing questions such as base prompt, framed prompt and text-based prompt.
The design of a good written test consists of four elements: instruction (frameworks and instructions), stimulus (question), expected response (what we expect to do with examinee), and post-task evaluation (assessment of the effectiveness and success of the task).
In order to design a variety of written test questions, the following points should be noted:
-           the item should generate the desired type of writing.
-           the item should stimulate the examinee to think and solve problems.
-           the item should be accessible, fun and challenging.
-           the item should deal with relevant and stimulating issues.
-           the item should not require prior expert knowledge.
-           the item should be fair and provide equal opportunities for all people to respond.
-           the item should be clear and unambiguous, genuine and limited.
-           the item should specify audience, target and texture
 
The sample of this study included 30 non-Iranian learners (males and females) from the humanities group who had completed a postgraduate course in Persian Language Teaching Center at Imam Khomeini International University. 30 Persian learners were selected randomly as the sample of this study. The age of the learners ranged from 18 to 30 and the duration of Persian learning was 5 to 7 months.
The format of the questions was based on the base prompt and the framed prompt described in the theoretical background. Examinees were required to write two questions. And this test was held in two steps in the morning and afternoon.
The data of this study included 30 writing samples written by advanced Persian learners. Written sheets were evaluated based on analytical scoring. In analytic scores, writing is assessed based on a number of aspects or criteria of writing skills, not based on a single score. According to the test objectives, features such as content, organization, coherence, usage / language, vocabulary, grammar, and punctuation are considered and each feature is assigned a score. There are advantages for analytical scoring. First, unlike the holistic method, this method provides information on the strengths and weaknesses of individuals in writing skills that are very useful for diagnostic feedback. Secondly, this method is very reliable, even if it is done by the less experienced scorer. Finally, it is easier to educate the scorer in this method, because the criteria and scales are obvious and detailed. The main weak point of the analytical method is that it is much more time-consuming than the holistic scoring, because in this method, the scorer must do more than make a decision about the writing (ie, examine different aspects of a single writing sample and give them separate scores)  Moreover, in addition to the necessity of preparing and compiling a set of criteria and detailed specifications for this method, several sessions should be given to training the raters to ensure that the differences between the scores of different scorers are reduced and the validity is increased.
In addition, due to the importance of systematic errors of Persian learners in the organization of educational materials, the choice of appropriate teaching and testing methods, 20 syntactic errors which had a high frequency in Persian language learners’ writings were considered. Examples of these errors included the deletion of grammatical and lexical elements, the addition of unnecessary or false elements, the selection of a false element, and the replacement of it with the correct element and the incorrect ordering and arrangement of elements. Inter-language errors included the removal of a verb, the replacement of a prepositional word with an object, the deletion of the object's symbol, misplacement of the adjectives and nouns, errors in the sequence of the sentence components. The intra-lingual errors included error in compound verb application, mood errors and verb form, error in the application of the tense of the verb, inadequacy of the entity with the subject, categorization error, error in the application of connected and distinct pronouns, and deletion of the words.