It is said that the advantages of direct testing is that it is intended to test some certain abilities, and preparation for that usually involves persistent practice of certain skills. Nevertheless, the skills tested are deprived from the authentic situation that later may cause difficulties for the students in using them.
Now we can shift to another notion - indirect testing. It differs from direct one in the way that it measures a skill through some other skill. It could mean the incorporation of various skills that are connected with each other, e.g. listening and speaking skills.
Indirect testing, regarding to Hughes, tests the usage of the language in real-life situation. Moreover, it suits all situations; whereas direct testing is bound to certain tasks intended to check a certain skill. Hughes (ibid.) assumes that indirect testing is more effective than direct one, for it covers a broader part of the language. It denotes that the learners are not constrained to one particular skill and a relevant exercise. They are free to elaborate all four skills; what is checked is their ability to operate with those skills and apply them in various, even unpredictable situations. This is the true indicator of the learners real knowledge of the language.
Indirect testing has more advantages that disadvantages, although the only drawback according to Hughes is that such type of testing is difficult to evaluate. It could be frustrating what to check and how to check; whether grammar should be evaluated higher, than composition structure or vice versa. The author of the paper agrees with that, however, basing on her experience at school again, she must claim that it is not so easy to apply indirect testing. This could be rather time-consuming, for it is a well-known fact that the duration of the class is just forty minutes; moreover, it is rather complicated to construct indirect test it demands a lot of work, but our teachers are usually overloaded with a variety of other duties. Thus, we can only hope on the course books that supply us with a variety of activities that involve cooperation of all four skills.
4.2 Discrete point and integrative testing
Having discussed the kinds of testing that deal with general aspects, such as certain skills and variety of skills in cooperation, we can come to the more detailed types as discrete point and integrative testing. According to Longman Dictionary of LTAL (112), discrete point test is a language test that is meant to test a particular language item, e.g. tenses. The basis of that type of tests is that we can test components of the language (grammar, vocabulary, pronunciation, and spelling) and language skills (listening, reading, speaking, and writing) separately. We can declare that discrete point test is a common test used by the teachers in our schools. Having studied a grammar topic or new vocabulary, having practiced it a great deal, the teacher basically gives a test based on the covered material. This test usually includes the items that were studied and will never display anything else from a far different field. The same will concern the language skills; if the teacher aim is to check reading skills; the other skills will be neglected. The author of the paper had used such types of tests herself, especially after a definite grammar topic was studied. She had to construct the tests herself basing on the examples displayed in various grammar books. It was usually gap-filling exercises, multiple choice items or cloze tests. Sometimes a creative work was offered, where the students had to write a story involving a certain grammar theme that was being checked. According to her observance, the students who studied hard were able to complete them successfully, though there were the cases when the students failed. Now having discussed the theory on validity, reliability and types of testing, it is even more difficult to realize who was really to blame for the test failures: either the tests were wrongly designed or there was a problem in teaching. Notwithstanding, this type was and still remains to be the most general and acceptable type in schools of our country, for it is easy to design, it concerns a certain aspect of the language and is easy to score. If we speak about types of tests we can say that this way of testing refers more to a progress test (You can see the examples of such type of test in Appendix 2).
Nevertheless, according to Bynom (2001:8) there is a certain drawback of discrete point testing, for it tests only separated parts, but does not show us the whole language. It is true, if our aim is to incorporate the whole language. Though, if we are to check the exact material the students were supposed to learn, then why not use it.
Discussing further, we have come to integrative tests. According to Longman Dictionary of LTAL, the integrative test intends to check several language skills and language components together or simultaneously. Hughes (1989:15) stipulates that the integrative tests display the learners knowledge of grammar, vocabulary, spelling together, but not as separate skills or items.
Alderson (1996:219) poses that, by and large, most teachers prefer using integrative testing to discrete point type. He explains the fact that basically the teachers either have no enough of spare time to check a certain split item being tested or the purpose of the test is only considered to view the whole material. Moreover, some language skills such as reading do not require the precise investigation of the students abilities whether they can cope with definite fragments of the text or not. We can render the prior statements as the idea that the teachers are mostly concerned with general language knowledge, but not with bits and pieces of it. The separate items usually are not capable of showing the real state of the students knowledge. What concerns the author of the paper, she finds integrative testing very useful, though more habitual one she believes to be discrete point test. She assumes that the teacher should incorporate both types of testing for effective evaluation of the students true language abilities.
4.3 Criterion-referenced and norm referenced testing
The next types of testing to be discussed are criterion-referenced and norm referenced testing. They are not focused directly on the language items, but on the scores the students can get. Again we should concern Longman Dictionary of LTAL (17) that states that criterion-referenced test measures the knowledge of the students according to set standards or criteria. This means that there will be certain criteria according to which the students will be assessed. There will be various criteria for different levels of the students language knowledge. Here the aim of testing is not to compare the results of the students. It is connected with the learners knowledge of the subject. As Hughes (1989:16) puts it the criterion-referenced tests check the actual language abilities of the students. They distinguish the weak and strong points of the students. The students either manage to pass the test or fail it. However, they never feel better or worse than their classmates, for the progress is focused and checked. At this point we can speak about the centralized exams at the end of the twelfth and ninth form. As far as the author of the paper is concerned, the results of the exams are confident, and the learners after passing the exams are conferred with various levels relevant to their language ability. Apart from that, once a year in Latvian schools the students are given tests designed by the officials of the Ministry of Education to check the level of the students and, what is most important, the work of the teacher. They call them diagnostic tests, though according to the material discussed above it is rather arguable. Nevertheless, we can accept the fact that criterion-referenced testing could be used in the form of diagnostic tests.
Advancing further, we have come to norm-referenced test that measures the knowledge of the learner and compares it with the knowledge of another member of his/her group. The learners score is compared with the scores of the other students. According to Hughes (ibid.), this type of test does not show us what exactly the student knows. Therefore, we presume that the best test format for the followi