Yi Shuo School District (46) SPSS Statistical Analysis (56) Validity
【思維導圖】
「分享興趣，傳播快樂，增長見聞，留下美好！大家好，這裡是小編。歡迎大家繼續訪問學苑內容，我們將竭誠為您帶來更多更好的內容分享。
"Share interests, spread happiness, increase knowledge and leave beauty! Hello everyone, this is Xiaobian. Welcome to continue to visit the content of the academy, and we will bring you more and better content sharing wholeheartedly.
在分析問題時，我們常藉助於量表或問卷進行。量表是否能測得所需的測量結果？測量結果的可靠性如何？需要對量表的效度、信度進行分析。
When analyzing problems, we often use scales or questionnaires. Can the scale measure the required measurement results? What is the reliability of the measurement results? It is necessary to analyze the validity and reliability of the scale.
【基本概念】
效度（Vaildity）是指量表是否真實反映了我們希望測量的問題。例如，智商測驗是否真正反映了智商的高低？生存質量調查是否真正反映了人們的生存質量？抑鬱量表調查是否真正反映了人們的抑鬱程度？這些都是關於效度的問題。對於效度，我們沒有絕對準確的答案，儘管不可能證明效度，但是可以用一些指標來評價效度。一般來說有，4個種類的效度，內容效度，實證效度，結構效度和區分效度。內容效度是一種基於概念的評價指標，其他三種效度是基於經驗的評價指標。如果一個量表實際上是有效的，那麼我們希望上述4種指標都比較滿意。一般使用的效度評估方法主要有判斷法和實證法，前者著重於測量特性與質的評估。通常依賴研究者對數據的主觀判斷，實證法則根據具體客觀的量化指標來進行效度的評估，進行實證效度評估的統計方法有：相關分析，多元回歸分析，因此分析，結構方程等。
Validity refers to whether the scale truly reflects the problems we want to measure. For example, do IQ tests really reflect IQ? Does the quality of life survey really reflect people's quality of life? Does the Depression Scale survey really reflect people's depression? These are all questions about validity. For validity, we have no absolutely accurate answer. Although it is impossible to prove validity, some indicators can be used to evaluate validity. Generally speaking, there are four kinds of validity: content validity, empirical validity, structure validity and discrimination validity. Content validity is a concept-based evaluation index, while the other three validity indexes are experience-based evaluation indexes. If a scale is actually valid, then we hope that all the above four indicators are satisfactory.Generally used validity evaluation methods mainly include judgment method and empirical method. The former focuses on measuring characteristics and quality evaluation. Usually relying on the subjective judgment of researchers on data, empirical law evaluates validity according to specific and objective quantitative indicators. Statistical methods for empirical validity evaluation include correlation analysis, multiple regression analysis, hence analysis, structural equation and so on.
信度（Reliability Analysis）是指測量的一致性。例如，準備調查你的年收入，你第1次回答是2萬元，如果能夠消除你對問題和答案的記憶，過一段時間後問你一個問同一個問題（當然也可以通過調查其他人了解你的年收入情況），通過考察你（你們）對同一問題的多次回答，可以判斷答案的一致性如何。答案的波動越大，信度越低；答案的一致性越好，信度越高。信度本身和測量所得結果正確與否無關，它的功能在於檢驗測量本身是否穩定，製作完成一份量表或問卷後，首先應該對該量表進行信度分析，確保其可靠性和穩定性，以免影響問卷內容分析結果的準確性。
Reliability Analysis refers to the consistency of measurement. For example, if you are going to investigate your annual income, your first answer is 20,000 yuan. If you can erase your memory of questions and answers, ask you the same question after a period of time (of course, you can also find out your annual income by investigating other people). By examining your multiple answers to the same question, you can judge how consistent the answers are. The greater the fluctuation of the answer, the lower the reliability; The better the consistency of the answers, the higher the reliability.Reliability itself has nothing to do with the accuracy of the measurement results. Its function is to test whether the measurement itself is stable. After making a scale or questionnaire, we should first analyze the reliability of the scale to ensure its reliability and stability, so as not to affect the accuracy of the questionnaire content analysis results.
根據測試時間和測試內容，信度又可分為內在信度和外在信度，內在信度是指一組問題（也可以稱為題項）是否測量同一個概念，即這些問題的內在一致性如何，能否穩定的衡量這一概念，最常用的檢測方法是Cronbach係數；而外在信度是指對相同測試者在不同時間測得的結果是否一致，重測信度是外在信度最常用的檢驗法。
According to the test time and test content, reliability can be divided into internal reliability and External reliability. Internal reliability refers to whether a group of questions (also known as questions) measure the same concept, that is, how consistent these questions are and whether they can measure this concept stably. The most commonly used detection method is Cronbach coefficient; External reliability refers to whether the results measured by the same tester are consistent at different times, and test-retest reliability is the most commonly used test method of external reliability.
效度與信度的關係是，信度為效度的必要而非充分條件，即有效度一定有信度，但有信度不一定有效度。
The relationship between validity and reliability is that reliability is a necessary but not sufficient condition of validity, that is, validity must have reliability, but reliability does not necessarily have validity.
下期預告：本期，我們學習了信度分析的理論基礎。下一期，我們將會學習內在信度信度分析的基礎知識。
如果您對今天的文章有獨特的想法，歡迎給我們留言，讓我們相約明天，祝您今天過得開心快樂！
參考資料：《SPSS23（中文版）統計分析實用教程》、百度百科
翻譯：訊飛語音
本文由learningyard新學苑原創，部分圖片文字來源於他處。如有侵權，請聯繫刪除。