98361
| Question | Answer |
| What is the critical values of Cronbach's Alpha? | 0.7 acceptable, but anything below 0.8 worth improving |
| What is test-retest reliability? How is it measured? | Test needs to measure same thing over period of time, same respondents resit on same/similar test at later time, correlation between performance at two times gives test-retest reliability |
| What is validity? | Test measures what it intends to measure |
| What is content validity? | Does it adequately sample aspects of topic being tested |
| What is predictive validity? | If it predicts those individuals who possess the attributes which the test intended to measure |
| What is construct validity? | Hypotheses can be derived from predicting relationship between the construct and other variables. Pattern of correlations between test and other measures is predicted |
| What is the first stage of constructing a test? | Outline the content areas of your test |
| What is the second stage of constructing a test? | Construct test items which appear to measure the trait/ability/skill |
| What is the third step of constructing a test? | Check as far as possible that the questions have content validity |
| Once the first draft of a test is done, what is it then important to do? | Test as many people as possible, analysizing for internal reliability of test using Cronbach's Alpha. |
| How do you confirm the validity of a test? | Concurrent validity with exciting measures, predictive validity for whatever the test is designed to measure, construct validity based on predictions of high/medium/low correlations |
| It is important that items can be scored in what way? | An umambigous way |
| What are the problem with rating scales? | People either avoid extremes or stick to extremes |
| What is the problem with having a 'don't know' answer? | Hard to score, people vary in using it, doesn't tell you much |
| What are two characteristics of a poor question? | Correlate only weakly with overall test score, ones most people answer in the same way |
| What are the features of the item pool of the draft of test? | Large, a professional would expect to reject at least half of initial question pool |
| What type of questions are worth using as different markers can easily and reliability mark them, with guessing controlled and is discriminative? | Multiple choice questions |
| What are the important rules in item construction? | Write simple items, all distractors should distract, only one option should be correct, answer of one question should not hint at another, avoid testing triva |
| What are the four types of psychometric tests? | Ability, motivation, neuropsychological, personality |
| What are the characteristics of open format items? | Written response, require detailed and not set responses, qualitative data, time consuming and subjective interpretation |
| What are closed format questions? | Short question or statement followed by number of options |
| What is the aim of closed questions? | It is to be able to quantify respondents answers to give information you want to know. |
| What three sources to use when constructing a test? | Theoretical literature, experts, and colleagues |
| How many items did Kline say a good psychometric should have? | 15-20 |
| What are the necessary properties of psychometric tests? | Reliability and validity, survey seek to accurately describe population, psychometric tests aim to accurately discrimate individuals |
| What is the recommended number of participants per items for testing a draft test? | 5 |
| What embarrassing questions bad? | Ethics, participants may give wrong information or not complete rest of questionaire |
| What is reliability? | Test should measure same thing consistently on different occasions |
| What is internal consistency? | Do all the items measure the same thing, is the test self-consitent |
| Why shouldn't you ask hypothetical questions? (Do not ask participants what they would do in a situation they may not ever find themselves in) | Get a response about situation person has never thought about over real view |
| What is consistency over time? | Same score on future occassions |
| What is a good way to test participants are answering properly? | Reverse wording, to see if participant provides contradictory answers |
| Why do we seek high reliability? | Maximum correlation between two variables depends on reliability of measures |
| How can means be used to identify poor questions? | Particularly high or low shows everyone is answering in same way |
| What is face validity, and how can it be measured? | Measures what it appears to measure, third person rating of individual |
| How much error is associated with each item in a test? | It is random, may increase or decrease score |
| Cronbach's alpha is used to measure... | Internal reliability |
| What does the Cronbach's alpha do? | Estimates average correlation of the tests with other tests of the same length taken from same universe of items |
Want to create your own Flashcards for free with GoConqr? Learn more.