问题描述:
英语翻译
For the measures of reliability,test-retest is hard to do,so surveyors often use a measure of internal consistency instead.However,this requires multiple questions that seek to measure the same idea,making the survey longer and somewhat repetitive.When this approach is used,the Cronbach’s alpha statistic measuring inter-item correlation is often the basis for measured reliability.Groups of questions that yield high alpha values,such as 0.7 or higher are preferred.
Validity measures are trickier.Criterion validity can be measured using an independent comparison.For example,a survey instrument intended to assess a firm’s quality practice using Baldrige criteria could be completed by a sample of Baldrige participants and winners and the scores from your instrument compared to their actual Baldrige scores.Content validity is often determined best by experts.The quickest way to do this is to review the relevant literature to make sure everything is covered correctly.Construct validity is measured using a statistical test called factor analysis,which uses the actual survey results to assess how well different measured concepts are differentiated.For example,factor analysis of a Baldrige pilot survey should group together by responses all of the questions related to leadership and not mix them with questions related to process measurement.
Given all of these requirements,survey instrument development usually starts with many more questions than expected in the final version.Pilot sample studies narrow the questions down to those that produce the best reliability and validity.
For the measures of reliability,test-retest is hard to do,so surveyors often use a measure of internal consistency instead.However,this requires multiple questions that seek to measure the same idea,making the survey longer and somewhat repetitive.When this approach is used,the Cronbach’s alpha statistic measuring inter-item correlation is often the basis for measured reliability.Groups of questions that yield high alpha values,such as 0.7 or higher are preferred.
Validity measures are trickier.Criterion validity can be measured using an independent comparison.For example,a survey instrument intended to assess a firm’s quality practice using Baldrige criteria could be completed by a sample of Baldrige participants and winners and the scores from your instrument compared to their actual Baldrige scores.Content validity is often determined best by experts.The quickest way to do this is to review the relevant literature to make sure everything is covered correctly.Construct validity is measured using a statistical test called factor analysis,which uses the actual survey results to assess how well different measured concepts are differentiated.For example,factor analysis of a Baldrige pilot survey should group together by responses all of the questions related to leadership and not mix them with questions related to process measurement.
Given all of these requirements,survey instrument development usually starts with many more questions than expected in the final version.Pilot sample studies narrow the questions down to those that produce the best reliability and validity.
问题解答:
我来补答展开全文阅读