Section 2 Designing Test Items and Reviewing

---

I.Thinking

【Case Reflection】

Mr W had taught the graduating class of a senior middle school for many years. Upon the school's request, he was transferred to teach the first-year class. Through careful preparation, he sorted out all the testing items of the College Entrance Examination（CEE）from each unit of the textbook and taught them in class. However, it turned out that his students were not all that engaged in the class, and their results in the unit exams were not all that satisfactory.

Please analyse the reasons.

II.Test Design

Tests and test design are indispensable to both formative assessment and summative assessment. Just as there are the National Curriculum Standards for teaching, there are also testing standards in testing. Testing standards generally include testing objectives, testing formats, test samples and appendices like glossary, subject tables etc.

Testing objectives are usually set according to the purpose of the test, in other words, the type of test. It is not difficult to establish testing objectives, but it is difficult to establish scientific ones. By comparing the examination syllabus of TOEFL and TOFELiBT（new TOFEL）, it is not hard to find that the new syllabus, which was developed about twenty years after the old one, is more scientific. By comparing the examination syllabus of National Matriculation English Test（NMET）in 1949, 1959, 1979, 1989, 1999 and 2009, we can also find that it makes great effort to establish a more scientific test objective.

The testing format should be established after setting the general testing objectives. As mentioned previously, there has been a tendency to apply direct language-use tests in the English Language Proficiency tests in middle and elementary schools. These tests evaluate students' English language proficiency in listening, speaking, reading, and writing by using activities which should contain a real purpose for language use.

Testing standards should provide test samples to clarify the standards in specific cases, and strictly speaking, the National Curriculum Standards should also include the standard and test samples. However, the current English Language Teaching（ELT）curriculum standard of middle and elementary schools in China only includes instructions for evaluating tests but no samples. In contrast, the College Entrance Examination Syllabus, TOFEL and IELTS are all provided with samples.

It is believed that significant tests and newly-established tests（e.g. the oral test supplement in the Senior Middle School Entrance Examination）should be provided with samples to help test takers to make necessary preparation.

Testing standards can be provided with appendices like glossary, subject tables etc. Specific requirements in respect of vocabulary and subjects are not required, but this is dependent on specific forms of the test. Generally speaking, non-curriculum exams, like entrance examinations, can be provided with or without glossary and subject tables.

As the general objective of testing, the ability to use English should be reflected in specific language activities. This requires a test blueprint, also referred to as specification or two-way specification.

Testing items are designed on completion of the blueprint. As the specific activities for realizing testing objectives, item writing should be carried out according to the requirements in the specifications.

A good direction in a test question should be concise and clear, and it should neither mislead the students nor take them too much time to read and understand.

The next section will briefly introduce some common techniques for test design.

(I) Multiple-choice Items Design

A multiple-choice item test is one type of testing pattern in which the best answer is selected from several options. This parallels an authentic language performance activity when we weigh our words in writing and speaking in real life. The weighing of words is widely seen in real language performance when we are in the process of literature creation, art production, and academic writing, in which accuracy of language is highly valued. However, grammatical and lexical multiple-choice items are not suitable for the assessment of English Language Proficiency in middle and elementary schools as there is a huge gap between the language performance objectives for middle and elementary school students and their ability to use language.

1.Testing Point

The testing points in grammatical and lexical multiple-choice items are mainly focused on verb tenses, clauses, and communicative usage, while other language points are less tested because the above three language testing points are both fundamental and sufficiently challenging for students. One multiple-choice item question usually examines one testing point, even if there are two blanks to fill in within the stems. If students have little mastery over the language and there's no necessity to weigh the words, the grammatical and lexical multiple-choice items are not applicable.

2.Stem

The stem should be of authentic context as well as authentic pragmatic purpose. In other words, a certain person should say certain things for a certain purpose in a certain context in a stem. The context in the stem could generate different uses： one part is constructed with one or more blanks, and the other provides context to indicate a basis on which the correct option should be selected. Stem construction can adopt either the embedded pattern or question-answer pattern. For example,

Embedded pattern： The performance __________ nearly three hours, but few people left the theatre early.

Question-answer pattern： Would you like to join me for a quick lunch before class?

It is best to select stems from Corpus Linguistics and well-known dictionaries published by English-speaking countries instead of using self-constructed sentences. For instance, if the perfect tense is to be tested, testers can search “have done” in Corpus Linguistics, collect a quantity of sentences in perfect tense, select an appropriate one and then construct it into a stem.

The content of a stem should be mainly about daily topics which students know (though with limited familiarity), so that students' schema-constructive ability can be tested. For example,

The wet weather will continue tomorrow when a cold front __________ to arrive.

Cold front in this stem is a meteorological term, which will not be familiar to those who don't watch weather forecasts often.

3.Option

The options themselves must be grammatically correct to ensure that none of the options will cause grammatical mistakes when placed in the stem. The options will, however, be different in appropriateness from the perspective of context, meaning and pragmatic purpose.

Depending on the requirements of particular multiple-choice items, the four options should be of the same category, form or word class. Alternatively, the form and category of the four options could either all have something in common or nothing in common. All the four options should be of approximately the same length and be as short as possible. Generally speaking, options like A, B and C, All of the above, Any of the above and None of the above should be avoided.

Encourage your children to try new things, but try not to __________ them too hard.

A. draw　B. strike　C. rush　D. push

All the four options are verbs and grammatically correct when placed in the blank.

One of the most important questions they had to consider was __________ of public health.

A. what　B. this　C. that　D. which

All the four options are pronouns and grammatically correct when placed in the blank.

The answer options should be organically arranged, whether in random arrangement according to scattered points, or in artificial disorder.

(II) Cloze test design

A cloze test is a testing pattern based on the theory of Gestalt psychology. The word Gestalt, which is German, means the complete form of separated unity or organization structure. The theory of Gestalt psychology is characterized as the opposition of elemental analysis and emphasis on psychological organic whole. The theory maintains that every psychological phenomenon is a separated whole, a Gestalt, and a complete unity. The organic response of the human brain to the environment provides a kind of organic structure or separated unity. The function of this insight is learning. The founders of the Gestalt psychology theory are M. Wertheimer, K. Koffka and W. Khler. A cloze test applies this theory in examining student ability to choose appropriate words for textual expression in a certain context, i.e. discourse comprehension and accurate language use in a particular context.

1.Testing Point

It is student language competence at the discourse level that is tested in a cloze test. Cloze tests therefore should be designed in such a way that most of the answers can be selected only after discourse comprehension, with only a small number of answers being obtained through sentence comprehension. The testing points should cover all those listed in the examination syllabus, mainly focusing on content words.

If the answer can be reached through comprehension of the sentence itself, the cloze is of low difficulty; if the answer can be reached through comprehension of the sentence and nearby context, the cloze is of moderate difficulty; if the answer can be reached through comprehension of the sentence and distant context, it is of high difficulty. The more distant the source sentences are from the question sentence, the more difficult the question is. Furthermore, some questions whose answers can be reached through schema awareness should also be included.

Cloze tests do not usually examine grammar or phrase collocations. The testing item and the other distractors in a cloze test should generally be high frequency words, though there are occasionally 10% of words of moderate difficulty.

2.Text

A cloze test examines two levels of comprehension competence. On the first level, it examines student schema-construction ability, for a student may know some but not be familiar with all the content of the text. On the second level, it examines student language competence to select appropriate words and expressions to complete the text. Students can select the correct answers only after they comprehend the text and word meaning.

The texts selected for the cloze test are usually narrative writings containing an argued position. Expositions are selected occasionally; however, simple stories are seldom used.

The length of the text should accommodate the amount of blanks. For a cloze with 10 deletions, the text is generally of 100-200 words; for a cloze with 20 deletions, the text is generally of 200-400 words. There should be 1% unfamiliar words for students to comprehend by applying vocabulary strategies; if there are more than this, they should be provided with Chinese annotations.

3.Deletions

No deletions should be made in the first sentence（or in the case of long sentences, the first fifteen words）of the text; the first deletion should be between the 15th and the 20th word. The last deletion should be 10-15 words before the last word. The longest distance between two deletions should be limited to within 30 words, while the shortest distance between two deletions should be more than 5 words. Deletions with more than 20 words in between should occur less than 5 times in a text.

4.Options

An option itself must be grammatically correct to ensure that none of the options will cause grammatical mistakes when placed in the stem, i.e. the options are different in semantic meaning alone. The options should usually be words, though phrases can be used as options on rare occasions.

Depending on the requirements of the cloze test, the four options should be of the same category, form and word class. Alternatively, the form and category of the four options could either all have something in common or nothing in common. For example, it is not recommended to construct three options of names of people and one option of the name of a place.

The testing points should mainly be constructed based on text comprehension, which means students can only obtain correct answers through text comprehension. Technically, about 70% of the answers should be obtained through text comprehension, and the rest of the answers should be through sentence comprehension. Options should all be high frequency words.

(III) Listening test design

In real life, listening is mostly related to speaking. The chances of listening to recordings are far less than listening to a spoken utterance. However, due to limitations in assessment, the materials used in the listening assessment of ELT in middle and elementary schools are almost all recorded materials. In examinations, we use unrepeated recorded listening materials. This listening activity has parallels to news-listening and film-watching activities in real life.

1.Testing Point

The testing points in the listening test should be selected according to the listening ability objectives stated in the examination syllabus. The questions in the listening test of middle and elementary schools mainly test for detail. Even the inferential and judgmental questions are mainly testing for detailed information rather than common knowledge or schema. Therefore, many questions of judgment are also to a great extent questions for detailed information.

Questions for short conversations should mainly concentrate on comprehension of detail, supplemented by a small number of inferential and judgmental questions. Questions for long conversations and monologues should mostly focus on detail comprehension, reference and judgment, supplemented by a small number of questions about main idea, attitude and intention.

2.Stem

The stem is usually a complete interrogative sentence rather than an embedded question or an embedded declarative sentence. There should be an explicit pragmatic purpose in the stem, and the question proposed should be about detailed information, the main idea, inference or judgment, attitude or intention of the speaker.

Questions should proportionally cover all the testing points, and each of them should be reasonable. For example, questions with extremely obvious answers such as “Who is Michael Jordan? ” or “Where is Sydney? ” are not appropriate. Questions with inconclusive answers, such as “What day is today? ” should also be avoided.

The following are some common stems：

Questions for main idea：

What are the speakers talking about?

Questions for detail：

What is the weather like?

What do we know about the woman?

What size does the man want?

Who will go to China next month?

Why did the woman go to the New York?

How long does the class last?

Questions for inference and judgment：（If detailed information has been mentioned in the passage, the inferential questions might be turned into questions for detail.）

What can we infer from the conversation?

What can we learn about flights to New Zealand at Christmas time?

Why does the speaker ask the questions?

How long will the man probably stay in New Zealand?

Questions for intention and attitude：

What does the man mean by saying sorry?

Why does Jimmy feel happy?

What does the woman think about the man's idea?

3.Options

Options should be a reasonable response to the proposed question. Irrelevant answers are not recommended, except for certain communicative purposes.

According to the requirements of multiple-choice items, the three options should be of the same category, form and word class. Alternatively, the form and category of the four options could either all have something in common or nothing in common. For example, it is not appropriate to construct two phrases with the preposition “in” and one with the preposition “on”, nor two country names and one holiday name. All the three options should be of approximately the same length.

4.Material

The language materials for a listening test must be authentic oral recordings rather than written language. Dialogues should be authentic dialogues in real communication, and monologues should also be authentic communicative monologues. Story narrations are generally not suitable for the monologue materials, except for story-telling. The subject matter of dialogues should be mainly focused on daily life, with abundant variety in topics.

(IV) Oral test design

An oral test is designed to evaluate oral communicative competence-a frequently used real life skill-and the accompanying testing patterns should be real daily activities. However, this kind of oral test is not widely used currently due to difficulties in practicability.

1.Testing Point

An oral test generally includes the assessment of student linguistic competence in areas of pronunciation, intonation, rhythm, language accuracy, oral discourse competence, expressive competence, social linguistic competence, and strategic competence. These are all communicative abilities which must be acquired.

2.Stem

The stem of an oral test should generally be clear and specific. If the testing pattern is question and answer, the stem should be an explicit question. Common testing patterns include reading-aloud, repetition, question-responding, self-talk, role play, discussion, etc.

A test taker's phonetic techniques in pronunciation, intonation, stress, sense group, linked sound, aspiration-weakening, and rhythm can be examined by reading aloud. Due to its ease of implementation, reading-aloud is commonly used in oral tests.

Repetition is an essential ability in language learning and an important testing pattern in oral tests. Student intonation for expressing various emotions and functions can be tested through repetition. The repetition can be long or short sentences, clauses or texts.

Question-responding is a commonly used testing pattern. It refers to the questioning and answering between testers and test takers in such things as greetings and daily conversations, follow-up questioning after role play, or self-presentation etc. It can be used alone or combined with other testing patterns.

Self-talk includes picture-description, presentation according to given information, and self-presentation.

Role-play is frequently applied by assigning tasks of different roles on an information card to students. After reading the information card, students can start a role play. The roles of the students（usually two）can be assigned by the testers on the spot.

Discussion is a higher requirement testing pattern. There are plenty of design methods; for example, testers can assign certain roles to test takers, or provide information in the form of pictures, diagrams and questions for them to discuss. The discussion usually requires three test takers' participation, but it can also be carried out between two test takers.

(V) Reading comprehension test design

Reading comprehension is the most common form of English performance for Chinese students, and is therefore the testing pattern accounting for the largest percentage of scores in various English examinations. Students in real life are not normally required to answer multiple choice questions, but they might reflect on certain issues which would lead to self-questioning while reading. In reading tests, multiple-choice items, blank-filling, essay questions etc. are commonly used to examine students' reading comprehension competence.

1.Testing Point

Test questions should be constructed to examine the reading skills as requested in the examination syllabus. They should mainly focus on detailed information and information-based inferential questions.

2.Discourse

The topics for reading（which are all listed in the National Curriculum Standards）should be related to daily life. Examples include such things as society, history, geography, personal emotions, interpersonal relationships, school life, health, popular science, modern technology, nature, the earth, the environment, weather, literature, art, traveling and transportation.

The genre of reading should mainly be exposition and narration, supplemented by practical writing and argumentation. As for practical writing, advertisements and news report are often adopted. It is not common to use diagrams alone in reading comprehension, except when combined with texts.

There should be a certain number of unfamiliar words in the reading passages in the examination paper.

3.Stems

The stems for reading comprehension should be in agreement with the real purpose of reading. In other words, we read stories to understand the content and plot, as well as get inspiration; we read news to grasp the essential elements; we read argumentations for points of view and attitudes; we read expositions for the knowledge of content and characteristics. We don't normally read advertisements for main ideas, or read stories to compare prices and times.

The stems of reading comprehension should always be stated explicitly. For English tests in middle and elementary schools, stems like NOT TRUE, EXCEPT, etc. are usually not used.

There are embedded stems and question-answering stems, which are used alternatively. Blanks are usually embedded at the end instead of in the middle, but never at the beginning of the stem.

The following are some of the common stems for main idea comprehension：

What is the text mainly about?

Which of the following would be the best title for the text?

What is the best title for the passage?

The passage is mainly about __________.

The best title for the last piece of information would be __________.

Task-based reading generally adopts stem patterns like matching, clozes, essay questions, and blank-fillings. The latter two of these are not provided with options.

The following is a sample of information matching：

The following are some letters from readers of the advertisements. Please match these letters with the corresponding advertisements.

1

May 15, 2006

Gentleman.

The enclosed advertisement for your organization comes from the April issue of Cosmopolitan . I am interested in doing volunteer work for abused children and would like to have your information booklet. My address is as follows：

6 Villa Maurice,

92160 Antony France

Yours truly,

D. Gruber

2

January 15, 2006

Dear Sir or Madam,

I am interested in writing adventure stories for children. Please send information on the course you offer to the following address：

5481 Torrance Blvd,

Torrance, CA 90503

Yours faithfully,

Irene Ponbs（Ms）

3

April 14, 2006

I am interested in finding a part-time job in the morning. Please send information to the following address：

2955 Pacific Coast Highway,

Torrance, CA 90505

Sincerely,

M.E. Nielson（Mr）

4

1068 Baxter Road,

Loveland, Ohio 45140

March 13, 2006

Dear Sir or Madam,

Please send me your catalog and information on the service you offer. I have a small business and have been robbed several times.

Yours truly,

J. Barnes

5

Am Frankenheim 21,

5000 Koln 40 Germany

June 1, 2006

Dear Sir,

I would like to receive the list mentioned in your recent advertisement, as well as more information about L.A.Parent . I enclose a self-addressed envelope.

Yours faithfully,

H. Wolf

4.Options

Four options of the reading comprehension should be of the same category, form and word class. Alternatively, the form and category of the four options could either all have something in common or nothing in common t. For example, it is not recommended to construct three options of person name and one option of place name.

If the stem is an interrogative sentence, the option items should be response sentences, adhering to the communication principle. Punctuation marks cannot be neglected.

(VI) Written expression test design

Writing is one of the most authentic language performances. However, due to the limitations of student English language competence and the cost constraints of paper grading, it has been very difficult to design real-life writing activities. In the assessment of English language teaching in middle and elementary schools, the assessment of writing ability generally uses a semi-writing activity.

1.Testing Point

The highest standard of written expression competence in elementary and middle schools specified by the CEE syllabus states that students should be able to “use grammar and vocabulary accurately, and use a certain number of sentence patterns and vocabulary to express ideas clearly and coherently”.

The highest written expression scoring of CEE is：

The writing task is fully accomplished as specified by the examination with all the information content covered; a wide range of grammatical structures and vocabulary are applied; there are a few lexical and vocabulary mistakes, which are caused by the intent of using complex sentence structures and advanced vocabulary; the cohesive devices are used effectively to make the whole structure compact; and the expected writing goal is fully achieved.

This description can be interpreted as： The information is intact without any missing content; the language structures are correct and of variety, with complex grammatical structures and advanced vocabulary; the discourse is coherent with logical connections among sentences.

The examinations at other levels should be inferred in line with this standard.

2.Stem

The writing stem usually consists of material and requirements.

The materials should be properly provided, with neither too much nor too little detailed background information. The English material should be limited to within 150 words, while the Chinese material should be limited to within 100 words. Pictures, diagrams and tables can also be used as background materials.

The requirements for writing should be clear with an emphasis on real writing context which illustrates the pragmatic purpose and audience. Enough space should be given to students for creative writing.

The writing context can be designed as writing letters to foreigners; going abroad for study and traveling; providing relevant information of one's own to foreigners; introducing one's life in school, family or community; writing for English newspapers; inquiring about information; or briefing about certain situations etc.

III.Review Teaching

Examinations can be highly significant measures of achievement and as such require a high level of preparation. CEE is an example which has direct influence on whether a student can go on with his or her further study in a university. In preparation for this kind of exam it is very important that the teacher has developed a good approach to review teaching.

Scientific and rational review teaching is helpful in raising students' overall language proficiency. Ignoring this approach will have a very negative impact on the learning and the exam results. At present, a large number of teachers of English in middle and elementary school have been misled by unscientific review teaching, and as a result student ability to use the language cannot be fully developed. To address this, teachers need to fully understand the requirements of the test and apply scientific review teaching methods to their teaching.

Generally speaking, different types of testing patterns place an emphasis on different abilities. It is fundamental to review teaching to analyse the requirements for student language competence in test questions.

(I) The competence-oriented analysis of listening tests and teaching suggestions

Listening is an active process by which students receive information, and is one of the most common forms of language use in real life.

There are differences between listening in testing and listening in real life. Listeners, i.e. participants, can give immediate feedback while listening in real life, whereas in an exam all they can do is listen. Although people just listen for information on some occasions in real life, such as when they are watching TV or watching movies, in many other situations they are likely to give some sort of response. However, due to limitations in exams, a listening test can only examine students' listening comprehension ability.

Most listening test questions are ones that are asked in real-life communication. For example, authentic listening activities can be： listening to others to clarify the weather; listening for arrangements; or listening to identify the relationship between two speakers.

Listening tests should also examine student ability to respond after listening in such contexts as note-taking in a lecture, asking questions after a lesson, debating after hearing opinions. These are abilities needed in real life and learning.

Based on the above analysis, the following are required in listening review teaching：

1.The teaching of listening should be combined with speaking. The content of a listening test at the elementary and middle school stage is mostly connected with spoken English, which means the contents are dialogues for the most part. Therefore, the teaching in preparation for the listening test is supposed to be teaching of both listening and speaking, with a large number of dialogues to improve students' listening competence.

2.The listening content should be authentic and specific. Since the listening test questions are related to real life, the preparation materials should be so as well.

3.The listening training should pay special attention to the training of students' practical listening competence. English belongs to a completely different language family compared to Chinese. For example, English pronunciation is distinct from that of Chinese, and the pronunciation of an English word usually has no direct connection with its meaning. As a result, it is important to enhance students' ability to distinguish different pronunciations by listening and to establish direct connections between the pronunciation and the meaning.

4.Listening skills training should improve students' listening competence. At present, the listening materials and test questions in CEE are not that difficult nor highly discriminative, with a relatively low requirement for listening. Therefore, the teaching and training of listening should focus more on the connection between the pronunciation and the meaning rather than listening skills, because the former is the practical difficulty that Chinese students are faced with when listening to English materials.

(II) The competence-oriented analysis of multiple choice and teaching suggestions

Choosing the most appropriate word or phrase from several possible options to complete a sentence is a practical language activity in real-life writing and speaking. Word or phrase selection widely exists in practical language activities such as literary, artistic and academic creation, and daily life. A very famous example of weighing words is the “knock or push” choice in a poem by Jia Dao, a poet of the Tang Dynasty. If we want to address a person, we choose the most appropriate title as well, so before visiting a scholar who is a professor, doctor or dean, the visitor will consider whether to address the scholar as Professor Zhang, Doctor Zhang, Dean Zhang or Mr Zhang. This is also a multiple choice question.

It is more appropriate to examine this language use ability at a relatively advanced level rather than a primary or secondary level. Therefore, this type of question should be avoided as much as possible at the elementary and middle school stage. Even if it were used, it would mainly examine a student's ability to use accurate sentence structures, rather than an ability to choose suitable words.

The following two aspects should be noted in the teaching of ability based on multiple choice in grammar and vocabulary tests：

1.Teaching reading is an important way to help students prepare for multiple choice questions in grammar and vocabulary tests because students' sense of language, which is the key to this type of questions, can be cultivated through reading.

2.Students should be trained to express themselves with accurate presentation skills; basic sentence structures and words; and real semantic choice in a real language situation for real pragmatic purposes. It is neither necessary nor appropriate to teach difficult grammatical items, words, structures or the usage of seldom-used words etc.

3.Classroom teaching should clearly raise students' schema construction skills, instruct students to read extensively, and construct schema in the process of reading.

(III) The competence-oriented analysis of cloze tests and teaching suggestions

Cloze tests, based on Gestalt psychology, can examine student ability to understand and construct the whole discourse by filling in blanks in the passage. It is also a real reading activity people may come across in real life. For example, guessing the missing words / characters in a slogan in books or newspapers. This is a kind of cloze without choices, namely, non-option blank filling.

Multiple choice is a common type of cloze which includes blank filling with the initial letter, blank filling with the words given, and free blank filling. Multiple choice is the main type applied in the national paper in CEE, although non-option blank-filling is the true gestalt of cloze. Due to the subjectivity of non-option blank filling, its results can't be machine graded. Therefore, only with an increase in funding for artificial grading can non-option blank filling gradually be implemented.

The purpose of a cloze is to evaluate student ability to complete a whole passage using appropriate words in context. There are two aspects of reading comprehension that a cloze examines： Schema activation and construction, as well as the ability to complete a passage by selecting appropriate words. First of all, there might be some content that students are not familiar with in the text, which requires schema construction during reading. Moreover, students have to understand the passage and all the options to make the right choice, which requires generating schema of the passage and understanding of relevant vocabulary.

All the options in the cloze are frequently-used words. The purpose of a cloze is not to examine student vocabulary, or ability to memorize word meaning, but the ability to use words in a specific context.

In the process of teaching, many teachers pay special attention to the differentiation of word meaning and ask students to analyse words from perspectives which are possibly even unfamiliar to English native speakers, such as the differences between take, bring, fetch and catch. In some tests, take is placed with mind, change, and save as options; while bring is with leave, give, and show . Apparently students are not tested for their understanding of the difference between take, bring, fetch and catch, but for their ability to use take and bring in authentic contexts. It's obvious that the teaching of word differentiation sometimes goes far beyond the requirements of the test.

Students are more likely to meet the requirements for cloze tests if the language teaching includes the following：

1.Pay more attention to the teaching of basic vocabulary and semantics rather than the meaning differentiation of difficult, complex, seldom-used and out-dated words. It is the student's ability to use basic vocabulary and basic semantic meanings that the cloze test examines, and therefore, more attention should be paid to this area.

2.Guide students to develop the schema construction on the basis of previous schema. A cloze test is designed to test student ability to construct schema on the basis of schema they have already formed. Therefore, student comprehension ability and schema-construction should be cultivated in order to meet the requirements of the cloze test.

(IV)The competence-oriented analysis of reading comprehension tests and teaching suggestions

Reading is the most prominent form of English language use for Chinese people, thus it accounts for the highest percentage of scores among various testing patterns. Students in real life are not normally required to answer multiple choice questions, but they might reflect on certain issues which would lead to self-questioning while reading. In reading tests, multiple-choice items and blank-filling etc. are usually used to examine student reading comprehension competence because a wide range of examinations can be carried out at limited cost.

The reading test requires test takers to understand short essays about general topics in notices, instructions, advertisements, books, newspapers, and magazines; and to have acquired basic reading skills. It mainly assesses the two basic skills of understanding detailed information and making simple judgments and inferences. Comprehension abilities at high levels, such as grasping the writer's intention, points of view and attitude, are less assessed.

Of course reading comprehension competence cannot be developed through the training of reading skills alone; it should also be trained from the perspective of schema, which is the key to the development of reading comprehension competence.

Based on the requirements of reading competence, the review teaching for reading tests should be as follows：

1.Develop student basic reading abilities rather than higher level reading skills, as the comprehensive competence assessed at the CEE level is not of high difficulty.

2.Put an emphasis on the training of schema activation, which has long been neglected. Reading is an activity of schema activation and construction, and teachers should help students to develop this ability in reading. The more schema students acquire, the faster their comprehension speed is, and the higher the level of comprehension they can achieve. Therefore, students should be encouraged to read extensively, with the purpose of constructing a wider range of schema.

(V) The competence-oriented analysis of writing tests and teaching suggestions

Due to the limitations in student English language competence and the cost constraint of test paper grading, it has been very difficult to design real-life writing activities. In the assessment of English language teaching in middle and elementary schools, the assessment of writing ability generally adopts a guided-writing approach. Just as the name implies, written expression requires students to accomplish the presentation of information in written language. Obtaining information is not the emphasis of the assessment.

The general requirement for written expression competence in English tests is that students should be able to “use grammar and vocabulary accurately and express ideas clearly and coherently with a certain amount of sentence patterns and vocabulary”.

The description of the highest written expression scoring is： the writing task is fully accomplished as specified by the examination, with all the information contents covered; a wide range of grammatical structures and vocabulary are applied; there are a few lexical and vocabulary mistakes, which are caused by the intent of using complex sentence structures and advanced vocabulary; the cohesive devices are used effectively to make the whole structure compact; the expected writing goal is fully achieved. It can be seen that the requirement for the syntax structure is not demanding.

Therefore, review teaching for writing tests should focus on the following：

1.Teachers should focus on the training of the students' ability to express themselves accurately and comprehensively by applying common vocabulary and sentence structures. This is not only a requirement in writing tests but also the basic competence requirement of writing in English after graduation. More emphasis should be put on students' ability to express themselves using learnt vocabulary.

2.There should be different writing requirements for students of different competence levels. Teachers should not only pay attention to the writing competence of average students, but also encourage the top ones to write articles of higher quality.

With the continuous implementation of English language courses as well as the wide application of task-based listening tests, reading tests and writing tests, the assessment for students' language competence is becoming increasingly direct. The English language teaching in middle and elementary schools should put more direct emphasis on the cultivation of comprehensive language competence

1.understand the definitions of formative assessment and summative assessment; and

2.learn the methods of formative assessment.

Please analyse the limitations of grammatical and lexical multiple-choice item tests in relation to the current practice of ELT in middle and elementary schools in China.

Please analyse the functions and limitations of the listening materials used in listening test design.

Please choose some examples of writing tasks in CEE and analyse the context and purpose of language use.