Reliability:

Refers  to  the accuracy  and consistency  of a measuring tool. A measure is reliable when  an  individual remains nearly  the  same  in repeated measurements. Reliability  is measured by reliability coefficient  and  the estimation  of rkliability  can be  done statistically. How to estimate  the  degree  of reliability?

a)  Coefficient of  Stability  shows consistency of performance  on  a test  over  a period  of  time. It is based  on  the correlation between performances on initial test and retest after  a distinct interval. That  is,  it  involves repetition of  the same test. A  test may  be  reliable even though  it may  not be valid. A  valid  test  is usually reliable. 

b)  Coefficient of Equivalence  shows consistency of performance on different forms  of a  test. It  is based  on the correlation between performances on the different .forms administered at nearly the  same time. That &,  a second equivalent  form of  the test is  administered. 

c)  Coeflcient of Internal Consistency  shows consistency  of performance  on different  parts or items  of  the test taken at  a  single  sitting. This  coefficient is usually computed  by  "split-half'  correlation (e.g. Sperman-Brown formula), or by  using Kuder-Richardson formula. That  is,  the analysis involves subdivision of the  test  into  two  or more  equivalent fractions.