Test Development Construction

Test Development Construction - Transcript

Test Development Construction

Test Development Construction
Before real work on test construction to start some broad decisions are taken by the investigator as Outlining the major objectives of test in general terms specifying the population for whom the test is intended Deciding upon possible conditions under which test can be used and its important uses etc

Steps in Test Development Construction
Planning of the test Specification of broad specific objectives of the test in clear terms Deciding upon nature of content items to be included type of instructions to be included method of sampling detailed arrangement of preliminary final administration Length of test probable time limit for completion of test probable statistical methods to be adopted etc

Writing items of the test Item is defined as a single question or task which is not often broken down into any smaller units Item writing is a creative art and lot depends upon test constructor s intuition imagination experience practice ingenuity Item writer must have a thorough knowledge complete mastery of the subject matter and he she must be fully aware of those persons for whom the test is meant

Double barrel double negative items should be avoided Items should be written in simple language so that it to be understood by any in the same way as it was intended After writing down items they must be submitted to a grp of subject experts for their criticisms suggestions which must then be duly modified

Preliminary Administration Experimental try out Purposes of experimental try out are manifold as Finding out major weaknesses omissions ambiguities inadequacies of the items To determine index of difficulty discrimination reasonable time limit of test and no of items to be included in final form Identifying any weakness vagueness in directions or instructions of the test At least 3 preliminary admin of the test

Aim of 1st admin is to detect any gross defects ambiguities and omissions in items instructions and the no of examinees should not be less than 100 is called pre tryout Aim of 2nd admin is to provide data for item analysis for this N be 400 is called tryout proper the sample must be similar to those for whom the test is intended Aim of 3rd preliminary admin is a kind of dress rehearsal and useful for deciding about the time limit

Reliability of final test Final test is administered to a fresh sample to compute reliability coefficient and for this sample size should not be less than 100 Validity of final test Validity should be computed from the data obtained from the samples other than those used in item analysis and size of sample should not be less than 100

Preparation of norms Preparation of manual reproduction of test

How stable are test results
It to depend upon few important factors as Reliability Validity Norms of PT itself Psychological makeup of sample Characteristics of situation where test is being administered as well as that of the investigator who is administering the test Characteristics of the level of measurement used for measuring psychological attributes

Cultural Specificity
Not single test can be universally applicable or equally fair to all cultures Instead of talking of culture free psychological tests we talk of culture common or culture fair or cross cultural tests and example of them are performance test nonlanguage test etc

Uses of Psychometric Tests for HR Work
In the Selection Recruitment Processes During Job Analysis In Identification of Training Needs Culture Building Phase OD Interventions Assessing Job Satisfaction Role Stress Work Motivation etc

Personality Situation Debate

Personality Situation Debate
The issue which is more important in determining what people do the personality or the situation Walter Mischell 1968 Personality and Assessment Personality traits do not predict behavior particularly well Situations are stronger predictor of behaviors People tend to over estimate personality consistency fundamental attribution error Personality assessment is a waste of time

Personality Situation Debate
The definitive test of usefulness of a personality trait is its ability to predict behavior Mischell argued that correlation b w personality and behavior is small typically 0 30 Ricard Nisbett 1980 revised personalitybehavior correlation upward slightly to 0 40 this accounts for approx 16 variance in behavior Situationists argue that this correlation is still small and as such that personality is not important in the prediction of behavior

Personality Situation Debate
Personality behavior correlations do not demonstrate value of situational variables because cause may be an unmeasured personality variable Need to demonstrate a high correlation b w situational variables behaviors rather than simply showing a low correlation b w personality behavior Funder Ozer 1980 reviewed various studies of situational effects and computed effect size correlations ranging from 36 to 42 similar to those found between personality factors and behavior

Personality Situation Debate
Trait is more likely to predict person s behavior if that trait is central to them Allport When people for whom a trait is consistent central are compared with those for whom the trait is not correlation b w that trait behavior is higher 0 50 0 57 respectively b w friendly behaviors conscientiousness friendliness trait Burger 2004 Situational variables are best suited to predict behaviors in specific situations personality traits are more able to predict patterns of behavior that persists across situations time

Personality Situation Debate
The Interactionist Perspective traits and situations interact to influence behavior Different situations affect different people in different ways Some situations allow expressions of personality other situations provoke a narrower range of behavior

Personality Situation Debate
There are individual differences in personalitysituation relationship Self Monitoring Snyder 1997 People vary in how much they adapt their behavior to given situations High self monitors observe the situations and their own behavior matching the two Low self monitors tend to be more consistent irrespective of the situation