K

Key (Answer Key) refers to the correct (suitable or appropriate) answer among the given alternatives in respect of selection type questions. In supply type questions, it would usually mean the 'clues' that help the examinee to develop his answer.

Knowledge is the primary level of understanding and is concerned with memory. It includes the knowing of new concepts, ideas, information and recalling the same. This is the first among the six levels of scholastic ability classified under the cognitive domain of the Taxonomy of Educational Objectives. Define, recall, recognize, acquire, identity, etc., are some of the action verbs that would represent knowledge. Selection type questions (objective types) will be more useful for the measurement of knowledge. "Knowledge is a structure of concepts and relations built by reflective thought out o information received. Any experience of participation, observation, reading, or thinking can become part of a person's knowledge. It will become part of that knowledge if he or she thinks about it, makes sense of it, understand it" (Ebel, 1979, p.373).

Kuder-Richardson Reliability is one of the methods used for estimating the internal consistency. It involves only a single administration of the test. The following two formulae (KR-20 and KR-21) are employed for this purpose :

1)  KR-20 :

r_tt         =              n   _            __St²  - Σ Si²_

                                                  (n-1)                    St²

where,             r_tt         =          the KR-20 reliability estimate.

                                    n          =          the number of items in the test.

                                    St²        =          the variance of the test scores.

                                    Σ Si²     =          the sum of the variances of all items,

                                                             which may also be expressed as Σ Pq.

2)  KR-21 :

            

r_tt         =              n   _            __x#  - x#²/n_
                           (n-1)                    St²

where,             r_tt         =          the KR-20 reliability estimate.

                                    n          =          the number of items in the test.

                                    x#       =          the mean of scores on the test.

                                    St²        =          the variance of test scores.

Kurtosis is a statistical term used in the test analysis. It describes the extent to which a unimodal frequency curve is peaked. The more peakedness of distribution is called Leptokurtic and the flatter one is called Platykurtic. Both these shapes are extreme deviation from the standard or normal distribution. These concepts are useful in the consideration of the shape of a scoring distribution.