Table of Specification refers to a test outline that specifies the proportion of the items that deals with each content area and with each objective/ability. [See, Blue Print].
Tailedness refers to specificity of hypothesized outcomes. Some hypothesis allow two directions of outcomes which are called 'tails' (e.g., positive or negative). A good hypothesis allows only one tail.
Tailored Test [See, Multi-stage Tests].
Target (Level) is an indicator of achievement with a time schedule. It is an important factor in the monitoring process. The instruments of evaluation will have to be spearheaded only towards the target in order to find out the actual point of achievement in that direction. Therefore, 'target setting' assumes importance as a tool in planning and performance besides evaluation.
Task refers to a set piece of work or an act or a response required from an individual or a group. Taking up of a test or an examination would itself be referred as a task. If an emphasis is made upon the responses rather than the goal to be reached, it would be the 'task set'.
Taxonomy is an orderly classification, arranged according to their natural relationships. It is an effective guide for identifying and defining instructional objectives. Taxonomies of cognitive, affective and psychomotor educational objectives have been developed by educationists. Bloom's taxonomy is widely known in the field of education.
Technique means the formal aspect of a specialised behaviour requiring expertness. It includes the using of measuring instruments. A little broader usage would be the procedure.
Temperament, a human tendency, is a combination of traits, attitudes and beliefs which would influence the cognitive and affective aspects of the learner's behaviour.
Terminal Behaviour refers to any visible activity displayed by the learner at the end of a course.
Terminal Examination refers to the examination conducted at the end of a term that signifies a learning period. It may also refer to an examination conducted at the end of one of the two or more stages of a course. Quarterly and half yearly examinations, semester and tremester examinations and also the basic, intermediate and advanced level examinations of a structured course are terminal examinations.
Test is an objective and standardised procedure/device for measuring a behavioural sample, such as ability, achievement, proficiency and other traits. It is conducted usually after a prior announcement and is designed to cover a specific unit of instruction as part of the learning process. In a regular teaching programme class tests are usually conducted once in a week or fortnight. [Also see, Examination].
Test Analysis is the process of analyzing the qualities of a test. It helps to find out the statistical quantities pertaining to the test scores, like mean, median, mode, standard deviation, standard error of mean, etc. These quantities provide a basis for proper interpretation of the test results, and also for estimating reliability and validity.
Test of Language Features is one that is developed and administered exclusively for the purpose of assessing the proficiency or skill or expertness of an examinee on a particular language with reference to its features. The domains of such tests may be of subject-verb agreement; verb tenses/aspects/voices; modifiers-comparatives/superlatives; relativisation, embedding; other discrete points, etc.
Test of Language Proficiency is one that is developed and administered exclusively for the purpose of assessing the proficiency or skill or ability of an examinee on a particular language. The domains of such tests may be of listening comprehension, oral production, reading comprehension, composition/essay/precise writing, general integrative tests, viz., cloze, dictation. [See, Proficiency Test].
Test Item is the smallest independent unit of a test. It may be of a statement to be judged, a question to be answered, an incomplete statement to be completed, a blank in a sentence or paragraph to be filled in and so on. [Also see, Item, Question].
Test Profile is a chart on which the ratings/scores for a number of traits are plotted on a common scale in parallel rows. This would help the evaluator to visually perceive the pattern of traits.
Test-Retest (Reliability) is one of the methods of estimating the test reliability. It is obtained by administering the test to the same group of examinees on two different occasions within a reasonable time gap and computing the correlation between the scores. [Syn.: Measures of Stability].
Test Tailoring is the process of constructing a test appropriate to specific measurement needs by matching the difficulty levels of the items administered to the ability levels of the examinees tested. For example, the difficult items not only waste the time of low ability examinees but also impair the scoring (on the easy items). This can be avoided by having a full length test of only easy items that can be matched to their ability. This kind of individualization in testing is termed as 'test tailoring'. This process includes the shaping up of the test information function, constructing multi-stage tests, computer adaptive tests, etc.
Test Wiseness is the skill in test taking. In other words, it refers to one's skill of knowing how to increase scores by evasion of some of the standard requirements. Students who are richly endowed with this skill are expected to score well on any test irrespective of their acquaintance with the subject.
Tetrachoric Correlation is the correlation between two variables, both assumed to be continuous and normally distributed, but each expressed in terms of two classes only. 'Tetra' is the combining form meaning four, i.e., having four parts. It is also one of the indices for calculating the validity of a test.
Theme refers to the topic on which the test is based. This would provide a sense of coherence and credibility. As Carroll and Hall (1985) visualize, sub-themes are often needed to provide a variety of items and also to ensure coverage of the language skills required. They also recommended that different sections chose for a test should have some form of linkage which they call as 'thematization'.
Thorndike Handwriting Scale is an instrument used to compare and grade a given bit of writing on the basis of its general merit. It consists of samples of handwriting arranged in classes. The samples are differing from each other by steps of approximately equal general merit a given bit of writing may be compared and graded by using these forms.
Three-Parameter Logistic Model of item analysis is an application of the latent trait theory. It is also referred to as the Fred Lord's Item Response Theory. As the name suggests, this model incorporates three parameters, viz., Item difficulty, Item discrimination and Guessing. The guessing parameter is sometimes called lower asymptote parameter as it also measures things other than guessing. The advantage of this model is that every item of the test is characterized by a set of three values of the three parameters. Similarly, for every candidate taking the test, his ability is estimated and is observed that for the same person same ability is obtained every time. [Also see, Latent Trait Measurement Theory, One-Parameter (Rasch) Model, Two-Parameter Model].
Tools of Evaluation are the instruments/devices/techniques used for the purpose of obtaining information/data or evidence regarding the progress (behavioural changes) of the examinees under various domains. Some of the tools - both objective and subjective - which are widely used are given below :
Cognitive (Scholastic - knowledge and Intellectual Ability) :
Oral and written tests/examinations consisting the following items/ questions to be administered under the formal and informal situations.
Objective Methods :
n Completion - Cloze types
n Constant alternatives
n Matching types
n Multiple choice
n Multiple facet
n Re-arrangement
n Simple questions - Word levelSubjective Methods :
n Completion (Sentence level)
n Long answer
n Problem solving
n Simple questions - Sentence level
n Short answer questions
Affective (Non-scholastic : Emotional State and Personality) :Oral and written inventories of the following types to be used/assigned under formal and informal situations.
Objective Methods :
n Check-list
n Direct question (simple)
n Q-technique
n Rating scale
n Situational testSubjective Methods :
n Anecdotal record
n Assignment
n Autobiography
n Drawing tests
n Group discussion
n Ink blot
n Interview
n Inventory questions
n Observation
n Picture projection
n Presentation of verbal stimuli
n Project and Fieldwork, Role play, etc.Psychomotor (Skills: Perception and Performance):
Practical assignments/inventories of the following types assigned under formal and informal situations.
Objective Methods :
n Checklist
n Rating Scale, etc.Subjective Methods :
n Field assignments
n Laboratory practicals
n Tests of acting (non-verval means)
n Tests of drawing
n Tests of firing
n Tests of lifting (weights)
n Tests of running (movements)
n Tests of singing
n Test of speaking
n Tests of Throwing
n Tests of typing and so on
Traits refers to a physical, mental or behavioural characteristic that distinguishes one person from another [Sattler, 1982].Translational Tests require the examinees to translate a set of words, phrases, paragraphs or passages, etc., from one language being learnt. Basically translation requires a high profile of vocabulary. If the translator cannot recall the appropriate equivalent of a particular word, he cannot demonstrate his ability to handle the given task. Valette (1972) suggests a variety of translation tests, viz., wordlists, words in context, grammatical structures and so on. Translation tests may be made more objectively by isolating those forms that present particular difficulty. They may be scored for accuracy or for literacy expression depending on the levels of examinees.
True-False Item consists of a statement that requires the examinee to judge unequivocally true or false. The statement usually provides only one central idea in it. It is an objective test item which provides a simple and direct means of measuring the essential outcome of formal learning. It is usually referred to as 'constant alternative item', classified under 'selection type'. It, however, has the disadvantage of 50% guessing probability.
True Score is a hypothetical score expected to be obtained by an examinee on a test when it is not affected by any measurement error. In other words, it is an ideal score obtained in an ideal test condition. For operational purposes, it is defined as the mean (average) of the number of scores that are obtained by an examinee on similar tests (or equivalent forms of a test) under similar conditions. For example, the computation of true score of an examinee who has obtained the following scores on a series of 10 similar tests under similar conditions is done as given below :
-------------------------------------------------------------------------------------------------- Test Score
--------------------------------------------------------------------------------------------------1 ::: 34
2 ::: 33
3 ::: 29
4 ::: 32
5 ::: 32
6 ::: 35
7 ::: 30
8 ::: 34
9 ::: 33
10 ::: 33__ 325
Mean = 325/10 = 32.5 (True score)
--------------------------------------------------------------------------------------------------
True Value [See, True Score].Truncation as defined in Dictionary of Statistical Terms (p.156) refers to a truncated distribution formed from another distribution by cutting off and ignoring the part lying to the right or left of a fixed variate value. In the same manner a truncated sample is obtained by ignoring all values greater than or less than a fixed value. It usually results in artificially low coefficients of correlation, reliability or validity.
T-Score is one of the derived scores (standard scores). It shows the difference (or distance) between a stated group's mean and any specified raw score value. T-score is designed to have a mean of 50 and a standard deviation of 10. It is equal to 10 times the respective Z-score, plus 50. That is,
T = 10Z + 50
Where, Z = x - x##
s.d.T-Test is a statistical procedure applied for determining whether the difference between two means is significant. It has the advantage of simplicity in computation. However, it does not permit simultaneous comparison of more than two groups.
Two-Parameter Model of item analysis is an improvement over the 'One-Parameter (Rasch) Model' based on the Latent Trait (Measurement) Theory. In addition to the first parameter (scale of person ability and item difficulty), this model consists of a second parameter called 'continuous estimate of discriminability'. [Also see, Latent Trait (Measurement) Theory, One-Parameter (Rasch) Model, Three-Parameter (Logistic)].
Two-Tailed Test is a test for the statistical stability (or statistical significance) of a difference when it is assumed that the difference may be in either direction.