Citing relevant examples, discuss how 'Reliability' and 'Validity' are predicted in a psychological testing.
Vikas Yadav, AIR 776, 2024
As per Anastasi, a psychological test is objective and standardized measure of behavior. ED SBIT.
RELIABILITY OF PSYCHOLOGICAL TESTS
Reliability refers to the Stability or consistency of test Scores which is reflected in its reproducability (Gregory)
There are 3 types of reliability
1. Test-retest reliability : Similar test is conducted over Some sample after some interval (say 15 days)
Fig. Test-retest reliability
for eg: Conducting an IQ test 15 days after on same class.
2. Alternate form reliability: In it an alternate form of test is created with similar items. The scores of both tests are correlated.
It mainly suffers from content-sampling error.
3. Internal Consistency reliability: Split half reliability involves dividing tests into 2 parts and conducting on similar population with time gap
Ins. Split-Half Eg Dividing class Test into 2 parts
Validity
Validity is a unitary concept. It refers to the degree to which a test measures what it PURPORTS to measure ( Gregory)
TYPES OF VALIDITY
1. Content Validity: It involves checking whether content is representative of the knowledge base or not IE Asking experts whether teacher made test has contents related to SBI T 2. FACE VALIDITY — whether on face items appears to be valid or not. 3. CRITERION VALIDITY — It involves correlating the test scores with the realistic, external criteria. IE Correlating Teacher Made IQ test with WAIS Scores 2 type 1. Predictive → correlation with predicted criteria ( future) IE IIT Rank to Predict B.Tech grade. 2. Concurrent — with existing criteria. Eg IQ score with SBIT 4. CONSTRUCT VALIDITY : Whether the test measures the CONSTRUCT / Trait it is designed to measure. 2 types 1. Convergent — Test converges and shows high correlation with different measure of similar construct Eg Personality Test showing high correlation with 16 PF 2. DIVERGENT — Test show low correlation with other construct measures. Eg IQ test showing low correlation with Big-Five Test
In this way, validity & reliability of psychological tests is measured.