Andijon davlat chet tillari instituti ingliz tili nazariy aspektlari kafedrasi chet tillarni o

 

Reliability
This term refers to the fact that the research procedures must be consistent, both over time and across 
a variety of people who might use them. It applies to both the data collection and data analysis of classroom 
research. 
One area of obvious concern about reliability is in situations where more than one observer is involved in 
trying to count or code the same things. 
The degree of reliability between observers can in fact be calculated, and of course steps can be taken to train 
observers to improve their "inter-observer agreement" figures. In general, before going ahead with the coding 
of a large corpus of data, classroom researchers strive for at least 85% agreement among observers and raters. 
This figure is called "inter-observer agreement" or "inter-rater reliability". Another matter for concern is the 
percentage of "intra-observer agreement"- that is, the extent to which a single observer or coder, working with 
the same data, codes or categorises the data consistently after a lapse or time. The figures range from 0 to 1.00, 
with decimals nearer to 1.00 representing greater consistency in the observers' use of the categories. 

Validity
As a technical construct in experimental studies, validity takes two important forms. First, there is the notion 
of "internal validity". A study is said to have internal validity if the outcomes of the experiment can be directly 
and unambiguously attributed to the treatment applied to the experimental group, rather than to uncontrolled 
factors. 
Internal validity relates to the extent to which the results of an experimental study can be reliably and 
unambiguously related to the treatment which was implemented. 
Chaudron (1 988b) has discussed 3 types of validation which relate to classroom research: 
-
Construct validation
: the construct is verifiable and can be "captured" through various measurement 
procedures. 
-
Criterion-related validity
: some form of measurement is used to measure a trait along with another 
form. 
-
Treatment validation
: the treatment was in fact implemented and that is identifiably different from 
whatever it was being compared with. 
According to Seliger and Shohamy (1989:190) the information needed for determining the quality of 
data collection procedures is the following: