學術產出-學位論文
文章檢視/開啟
書目匯出
-
題名 單例程序和雙例程序在導致人類負向關聯學習上的效果之再探
Further studies on the usefulness of single and double discrimination procedures in inducing human vegative contingency learning作者 廖英廷 貢獻者 顏乃欣
廖英廷日期 2001 上傳時間 18-九月-2009 18:33:49 (UTC+8) 摘要 本研究在人類負向關聯學習（negativecontingencylearning）的主題下，檢視制約抑制派典之效果（conditionedinhibition，簡稱Cl，指P+而後PN-的學習經驗，代表後續事件出現於P事件之後但不出現於PN事件之後）。自Pavlov（1927）以來，使用Cl的動物研究向來皆能出現基本的學習指標，例如負向遷移（negativetransfer，事件之負向關聯性類化到原學習情境之外情境之能力），然而在人類受試研究中似乎有所不同，如Williams（1995）或Shanks,Charles,Darby,& Azmi（1998）便無法觀察到Cl的效果。Williams（1995）因此衍生出所謂的雙例程序（P+而後PN-與N--PN-與N-之呈現無先後之分）以替代原本的Cl（或稱單例程序），而將雙例程序與單例程序之下的所得到的受試表現分別稱為成分抑制（elementalinhibition，出現負向遷移）與整體抑制（configuralinhibition，未出現負向遷移）。 儘管單例程序約有效性受到Williams等學者的質疑，然而相關研究的數量以及相關的理論解釋尚不足以支持此一質疑，因此，本研究再次檢視了單例程序的效果，探討可能造成單例程序之學習現象不一致的原因，亦初步地檢討了雙例程序與負向關聯學習的對應關係。實驗一操弄學習內容的差異，單例組內容為P＋而後PN－，雙例組內容為P＋而後PN－與N－，分別檢視兩類程序的效果，結果發現在兩類程序之下皆出現負向遷移指標。實驗二延續實驗一單例程序的實驗，檢視可能造成Williams（1995）之單例程序實驗失效的原因，共包括四個子實驗。實驗二a操弄單例組的內容，在學習階段以P1，P2＋（P2為陌生事件）取代PT＋，同時在測試階段以C、N，、TC及TN，（幾乎皆為陌生之嘗試組合）取代P、T、TN，及PT，檢視陌生嘗試類別的影響，結果仍獲致負向遷移的現象。實驗二b以實驗二a的事件安排為基礎，操弄嘗試類別之外的因素，以虛構股市交易指導語、表格呈現資料方式以及去除中間指導語的操弄，取代吃食與腹瀉指導語、個別呈現資料方式以及給予中間指導語的操弄，結果負向遷移現象消失，受試出現整體抑制的表現。實驗二c延伸自實驗二b的設計，個別檢視三項因素對於受試表現的影響，結果僅發現去除中間指導語足以使得負向遷移現象消失，但虛構股市交易指導語具有較特別之影響，該因素使得受試的表現既不屬於成分抑制亦不屬於整體抑制。實驗二d延續實驗二。的設計，將測試階段的嘗試修改成受試較熟悉的P、T、TN1及PT，去除了陌生的測試階段嘗試之可能影響，再次檢視去除中間指導語的作用，結果受試的表現不再屬於整體抑制表現，而變成介於成分抑制與整體抑制之間的表現。實驗三延續實驗一的雙例程序實驗，使標的N1事件不具與正向事件配對的學習階段經驗，初步檢視雙例程序與預期違逆的關係（預期違逆指的是在P＋而後PN－的學習中，首先由P＋經驗產生P之後將發生後續事件的預期，但是此一預期卻將在PN－經驗中受到反駁的概念，在此一概念中P＋與PN－皆不可缺乏），結果發現N1事件仍然能夠通過學習測試，暗示雙例程序具有花不符合預期違逆的條件下，導致類似學習表現的可能性。 本研究主要的發現可歸納為兩個部分：第一，在單例程序之下，受試的表現仍受到某些事件安排之外的實驗操弄之影響，無法完全歸因於P＋而後PN－的學習，但是若能盡量減低受試對於實驗操弄的疑惑，單例程序仍可能是相當良好的研究工具；第二，雙例程序之下的學習指標似乎有可能出現在學習條件不符合預期違逆的情況之下，因此雙例程序是否僅是較單例程序更穩定的實驗設計，抑或牽涉到不同的學習運作，仍有待進一步的澄清。
In this research we reexamined the usefulness of conditioned inhibition paradigm (Cl, Cl is a P+ then PN- learning which means outcome appear after P but absent after PN) under the theme of human negative contingency learning. The animal studies of negative contingency learning have used Cl as basic experimental design since Pavlov (1927) and constantly observed indexes of successful learning like negative transfer (The ability that target event transfer its negative contingency to other conditions). Recently human negative contingency learning studies have also started to use Cl, but some of these researches failed to find Cl`s usefulness as animal studies. For example, in Williams (1995) and Shanks, Charles, Darby, & Azmi (1998), subjects didn`t learn that target event was negative contingent with the outcome. Williams (1995) therefore created a new paradigm called double discrimination procedure (P+ then PN- and N-, which means outcome appear after P but absent after PN and N) in order to replace therparadigm (Cl, also called single discrimination procedure). The observations under single and double discrimination procedure were respectively called configural inhibition (negative transfer appear) and elemental inhibition (negative transfer absent) by Williams. Although the usefulness of single discrimination procedure was doubted by researchers like Williams, the quantity of relevant researches and relevant theories are still not strong enough to support this point of view. As a result, in our research we reexamined the usefulness of single discrimination procedure and tried to find out why discrepancy appeared in relevant human researches. In experiment 1, we respectively examined the usefulness of single and double discrimination procedure. The subjects in single discrimination group received P + then PN- as learning material, and the subjects in double discrimination group received P+ then PN- and N-----. Results indicated that both conditions were sufficient for negative transfer. Experiment 2 included 4 single discrimination sub-experiments based on the design of experiment 1 and intended to find out what made Williams (1995) fail to observe negative transfer under single discrimination procedure. In experiment 2a, we replaced PT+ in the learning stage by P1P2 + (P2 was a novel event) and replaced P, T, TN1 and PT in the test stage by C, N1, TC and TN1 (all four trails were novel trails), but negative transfer remained. In experiment 2b, the learning material kept the same as experiment 2a but we changed three parts of the former design as manipulations: first, we used fictional stock market version instruction instead of the former eating and diarrhea version; second, we showed all possible events on a table and indicated which event is relevant instead of only plotting the relevant events on the screen; third, we gave no more instruction for test stage like before. In this experiment, the negative transfer index disappeared as reported in Williams (1995). In other words, this experiment showed configural inhibition. Later, a relevant examination continued in the experiment 2c as we respectively checked the three possible causes indicated by experiment 2b. And the results showed that only the withhold-instruction-for-test-stage manipulation was sufficient to abolish negative transfer but the fictional stock market version instruction or table presenting were not. However, the fictional stock market version instruction acted in a special fashion that subjects under this manipulation performed in between configural and elemental inhibition. The independent function of withholding instruction for test stage was then checked in experiment 2d by changing those test stage trials back to familiar ones(P, T, TN) and PT). This time subjects didn`t act that extremely but performed in between configural and elemental inhibition. Experiment 3 was a double discrimination experiment based on the design of experiment 1, but we let the target event N1 lack of the compound experience with positive contingent event in order to understand the relationship between double discrimination procedure and expectancy violation (In the P+ then PN-learning, subjects first learned that P will be followed by outcome, but this expectancy should be violated in the later PN- experience. The point of expectancy violation is that both P+ and PN- should be necessary for learning). Even so the target event N1 still passed the test. Such result rose the possibility that the learning under double discrimination might not necessarily correspond to negative contingency learning. In conclusion, our research had two main implications: First, subjects under single discrimination procedure are easily affected by some experimental manipulations. But if subjects` doubts about experiment manipulations could be reasonably reduced, single discrimination procedure would still be good enough for human researches; Second, the learning under double discrimination procedure might not necessarily correspond to negative contingency learning. As a result, its usefulness compared to single discrimination procedure should be reconsidered.
負向關聯學習-----01 條件抑制-----03 阻斷效果與預期違逆-----04 區辨學習、負向遷移與CI下的學習機制-----06 人類負向關聯學習-----10 人類的預期違逆表現-----11 人類面對CI情境有不同的反應傾向?-----13 關於單例程序失效的理論探討-----17 對於雙例程序之學習內容的檢討-----23 測試方法的考量-----25 研究目的-----29 實驗一-----30 實驗二-----37 實驗二 a-----39 實驗二 b-----43 實驗二 C-----48 實驗二 d-----53 實驗二 綜合討論-----56 實驗三-----61 綜合討論-----65 人類受試的處理策略-----66 雙例程序的暗示-----71 本研究結果的可能應用-----72 參考文獻-----74 表一、Chapman與Robbins（1990）實驗二的實驗設計-----11 表二、Wi11jams（1995）實驗二b的實驗設計-----24 表三、實驗一的材料安排-----32 表四、Wi11iams（1995）實驗二b單例組與本研究實驗一單例組主要的不同-----37 表五、實驗二a的材料安排-----41 表六、實驗二d的材料安排-----54 表七、可能影響受試表現為成分處理或整體處理的操弄-----58 表八、實驗三的材料安排-----62 圖一、N之負向調節直接作用於Us表徵上的CI運作機制-----06 圖二、N之負向調節作用於P與US聯結關係上的CI運作機制-----08 圖三、在PearCe（1987）的理論架構下，負向關聯學習可能的運作方式-----21 圖四、實驗一單例組與雙例組中兩類關鍵判斷的次數-----33 圖五、實驗二a中兩類關鍵判斷的次數-----42 圖六、實驗二b中兩類關鍵判斷的次數-----46 圖七、實驗二c中各組中兩類關鍵判斷的次數-----50 圖八、實驗二d中兩類關鍵判斷的次數-----54 圖九、本研究所有單例程序實驗申兩類關鍵判斷的次數-----56 圖十、實驗三與實驗一雙例組兩類關鍵判斷的次數對照-----63描述 碩士
國立政治大學
心理學研究所
90資料來源 http://thesis.lib.nccu.edu.tw/record/#G91NCCU1102012 資料類型 thesis dc.contributor.advisor 顏乃欣 zh_TW dc.contributor.author (作者) 廖英廷 zh_TW dc.creator (作者) 廖英廷 zh_TW dc.date (日期) 2001 en_US dc.date.accessioned 18-九月-2009 18:33:49 (UTC+8) - dc.date.available 18-九月-2009 18:33:49 (UTC+8) - dc.date.issued (上傳時間) 18-九月-2009 18:33:49 (UTC+8) - dc.identifier (其他 識別碼) G91NCCU1102012 en_US dc.identifier.uri (URI) https://nccur.lib.nccu.edu.tw/handle/140.119/36442 - dc.description (描述) 碩士 zh_TW dc.description (描述) 國立政治大學 zh_TW dc.description (描述) 心理學研究所 zh_TW dc.description (描述) 90 zh_TW dc.description.abstract (摘要) 本研究在人類負向關聯學習（negativecontingencylearning）的主題下，檢視制約抑制派典之效果（conditionedinhibition，簡稱Cl，指P+而後PN-的學習經驗，代表後續事件出現於P事件之後但不出現於PN事件之後）。自Pavlov（1927）以來，使用Cl的動物研究向來皆能出現基本的學習指標，例如負向遷移（negativetransfer，事件之負向關聯性類化到原學習情境之外情境之能力），然而在人類受試研究中似乎有所不同，如Williams（1995）或Shanks,Charles,Darby,& Azmi（1998）便無法觀察到Cl的效果。Williams（1995）因此衍生出所謂的雙例程序（P+而後PN-與N--PN-與N-之呈現無先後之分）以替代原本的Cl（或稱單例程序），而將雙例程序與單例程序之下的所得到的受試表現分別稱為成分抑制（elementalinhibition，出現負向遷移）與整體抑制（configuralinhibition，未出現負向遷移）。 儘管單例程序約有效性受到Williams等學者的質疑，然而相關研究的數量以及相關的理論解釋尚不足以支持此一質疑，因此，本研究再次檢視了單例程序的效果，探討可能造成單例程序之學習現象不一致的原因，亦初步地檢討了雙例程序與負向關聯學習的對應關係。實驗一操弄學習內容的差異，單例組內容為P＋而後PN－，雙例組內容為P＋而後PN－與N－，分別檢視兩類程序的效果，結果發現在兩類程序之下皆出現負向遷移指標。實驗二延續實驗一單例程序的實驗，檢視可能造成Williams（1995）之單例程序實驗失效的原因，共包括四個子實驗。實驗二a操弄單例組的內容，在學習階段以P1，P2＋（P2為陌生事件）取代PT＋，同時在測試階段以C、N，、TC及TN，（幾乎皆為陌生之嘗試組合）取代P、T、TN，及PT，檢視陌生嘗試類別的影響，結果仍獲致負向遷移的現象。實驗二b以實驗二a的事件安排為基礎，操弄嘗試類別之外的因素，以虛構股市交易指導語、表格呈現資料方式以及去除中間指導語的操弄，取代吃食與腹瀉指導語、個別呈現資料方式以及給予中間指導語的操弄，結果負向遷移現象消失，受試出現整體抑制的表現。實驗二c延伸自實驗二b的設計，個別檢視三項因素對於受試表現的影響，結果僅發現去除中間指導語足以使得負向遷移現象消失，但虛構股市交易指導語具有較特別之影響，該因素使得受試的表現既不屬於成分抑制亦不屬於整體抑制。實驗二d延續實驗二。的設計，將測試階段的嘗試修改成受試較熟悉的P、T、TN1及PT，去除了陌生的測試階段嘗試之可能影響，再次檢視去除中間指導語的作用，結果受試的表現不再屬於整體抑制表現，而變成介於成分抑制與整體抑制之間的表現。實驗三延續實驗一的雙例程序實驗，使標的N1事件不具與正向事件配對的學習階段經驗，初步檢視雙例程序與預期違逆的關係（預期違逆指的是在P＋而後PN－的學習中，首先由P＋經驗產生P之後將發生後續事件的預期，但是此一預期卻將在PN－經驗中受到反駁的概念，在此一概念中P＋與PN－皆不可缺乏），結果發現N1事件仍然能夠通過學習測試，暗示雙例程序具有花不符合預期違逆的條件下，導致類似學習表現的可能性。 本研究主要的發現可歸納為兩個部分：第一，在單例程序之下，受試的表現仍受到某些事件安排之外的實驗操弄之影響，無法完全歸因於P＋而後PN－的學習，但是若能盡量減低受試對於實驗操弄的疑惑，單例程序仍可能是相當良好的研究工具；第二，雙例程序之下的學習指標似乎有可能出現在學習條件不符合預期違逆的情況之下，因此雙例程序是否僅是較單例程序更穩定的實驗設計，抑或牽涉到不同的學習運作，仍有待進一步的澄清。 zh_TW dc.description.abstract (摘要) In this research we reexamined the usefulness of conditioned inhibition paradigm (Cl, Cl is a P+ then PN- learning which means outcome appear after P but absent after PN) under the theme of human negative contingency learning. The animal studies of negative contingency learning have used Cl as basic experimental design since Pavlov (1927) and constantly observed indexes of successful learning like negative transfer (The ability that target event transfer its negative contingency to other conditions). Recently human negative contingency learning studies have also started to use Cl, but some of these researches failed to find Cl`s usefulness as animal studies. For example, in Williams (1995) and Shanks, Charles, Darby, & Azmi (1998), subjects didn`t learn that target event was negative contingent with the outcome. Williams (1995) therefore created a new paradigm called double discrimination procedure (P+ then PN- and N-, which means outcome appear after P but absent after PN and N) in order to replace therparadigm (Cl, also called single discrimination procedure). The observations under single and double discrimination procedure were respectively called configural inhibition (negative transfer appear) and elemental inhibition (negative transfer absent) by Williams. Although the usefulness of single discrimination procedure was doubted by researchers like Williams, the quantity of relevant researches and relevant theories are still not strong enough to support this point of view. As a result, in our research we reexamined the usefulness of single discrimination procedure and tried to find out why discrepancy appeared in relevant human researches. In experiment 1, we respectively examined the usefulness of single and double discrimination procedure. The subjects in single discrimination group received P + then PN- as learning material, and the subjects in double discrimination group received P+ then PN- and N-----. Results indicated that both conditions were sufficient for negative transfer. Experiment 2 included 4 single discrimination sub-experiments based on the design of experiment 1 and intended to find out what made Williams (1995) fail to observe negative transfer under single discrimination procedure. In experiment 2a, we replaced PT+ in the learning stage by P1P2 + (P2 was a novel event) and replaced P, T, TN1 and PT in the test stage by C, N1, TC and TN1 (all four trails were novel trails), but negative transfer remained. In experiment 2b, the learning material kept the same as experiment 2a but we changed three parts of the former design as manipulations: first, we used fictional stock market version instruction instead of the former eating and diarrhea version; second, we showed all possible events on a table and indicated which event is relevant instead of only plotting the relevant events on the screen; third, we gave no more instruction for test stage like before. In this experiment, the negative transfer index disappeared as reported in Williams (1995). In other words, this experiment showed configural inhibition. Later, a relevant examination continued in the experiment 2c as we respectively checked the three possible causes indicated by experiment 2b. And the results showed that only the withhold-instruction-for-test-stage manipulation was sufficient to abolish negative transfer but the fictional stock market version instruction or table presenting were not. However, the fictional stock market version instruction acted in a special fashion that subjects under this manipulation performed in between configural and elemental inhibition. The independent function of withholding instruction for test stage was then checked in experiment 2d by changing those test stage trials back to familiar ones(P, T, TN) and PT). This time subjects didn`t act that extremely but performed in between configural and elemental inhibition. Experiment 3 was a double discrimination experiment based on the design of experiment 1, but we let the target event N1 lack of the compound experience with positive contingent event in order to understand the relationship between double discrimination procedure and expectancy violation (In the P+ then PN-learning, subjects first learned that P will be followed by outcome, but this expectancy should be violated in the later PN- experience. The point of expectancy violation is that both P+ and PN- should be necessary for learning). Even so the target event N1 still passed the test. Such result rose the possibility that the learning under double discrimination might not necessarily correspond to negative contingency learning. In conclusion, our research had two main implications: First, subjects under single discrimination procedure are easily affected by some experimental manipulations. But if subjects` doubts about experiment manipulations could be reasonably reduced, single discrimination procedure would still be good enough for human researches; Second, the learning under double discrimination procedure might not necessarily correspond to negative contingency learning. As a result, its usefulness compared to single discrimination procedure should be reconsidered. en_US dc.description.abstract (摘要) 負向關聯學習-----01 條件抑制-----03 阻斷效果與預期違逆-----04 區辨學習、負向遷移與CI下的學習機制-----06 人類負向關聯學習-----10 人類的預期違逆表現-----11 人類面對CI情境有不同的反應傾向?-----13 關於單例程序失效的理論探討-----17 對於雙例程序之學習內容的檢討-----23 測試方法的考量-----25 研究目的-----29 實驗一-----30 實驗二-----37 實驗二 a-----39 實驗二 b-----43 實驗二 C-----48 實驗二 d-----53 實驗二 綜合討論-----56 實驗三-----61 綜合討論-----65 人類受試的處理策略-----66 雙例程序的暗示-----71 本研究結果的可能應用-----72 參考文獻-----74 表一、Chapman與Robbins（1990）實驗二的實驗設計-----11 表二、Wi11jams（1995）實驗二b的實驗設計-----24 表三、實驗一的材料安排-----32 表四、Wi11iams（1995）實驗二b單例組與本研究實驗一單例組主要的不同-----37 表五、實驗二a的材料安排-----41 表六、實驗二d的材料安排-----54 表七、可能影響受試表現為成分處理或整體處理的操弄-----58 表八、實驗三的材料安排-----62 圖一、N之負向調節直接作用於Us表徵上的CI運作機制-----06 圖二、N之負向調節作用於P與US聯結關係上的CI運作機制-----08 圖三、在PearCe（1987）的理論架構下，負向關聯學習可能的運作方式-----21 圖四、實驗一單例組與雙例組中兩類關鍵判斷的次數-----33 圖五、實驗二a中兩類關鍵判斷的次數-----42 圖六、實驗二b中兩類關鍵判斷的次數-----46 圖七、實驗二c中各組中兩類關鍵判斷的次數-----50 圖八、實驗二d中兩類關鍵判斷的次數-----54 圖九、本研究所有單例程序實驗申兩類關鍵判斷的次數-----56 圖十、實驗三與實驗一雙例組兩類關鍵判斷的次數對照-----63 - dc.description.tableofcontents 負向關聯學習-----01 條件抑制-----03 阻斷效果與預期違逆-----04 區辨學習、負向遷移與CI下的學習機制-----06 人類負向關聯學習-----10 人類的預期違逆表現-----11 人類面對CI情境有不同的反應傾向?-----13 關於單例程序失效的理論探討-----17 對於雙例程序之學習內容的檢討-----23 測試方法的考量-----25 研究目的-----29 實驗一-----30 實驗二-----37 實驗二 a-----39 實驗二 b-----43 實驗二 C-----48 實驗二 d-----53 實驗二 綜合討論-----56 實驗三-----61 綜合討論-----65 人類受試的處理策略-----66 雙例程序的暗示-----71 本研究結果的可能應用-----72 參考文獻-----74 表一、Chapman與Robbins（1990）實驗二的實驗設計-----11 表二、Wi11jams（1995）實驗二b的實驗設計-----24 表三、實驗一的材料安排-----32 表四、Wi11iams（1995）實驗二b單例組與本研究實驗一單例組主要的不同-----37 表五、實驗二a的材料安排-----41 表六、實驗二d的材料安排-----54 表七、可能影響受試表現為成分處理或整體處理的操弄-----58 表八、實驗三的材料安排-----62 圖一、N之負向調節直接作用於Us表徵上的CI運作機制-----06 圖二、N之負向調節作用於P與US聯結關係上的CI運作機制-----08 圖三、在PearCe（1987）的理論架構下，負向關聯學習可能的運作方式-----21 圖四、實驗一單例組與雙例組中兩類關鍵判斷的次數-----33 圖五、實驗二a中兩類關鍵判斷的次數-----42 圖六、實驗二b中兩類關鍵判斷的次數-----46 圖七、實驗二c中各組中兩類關鍵判斷的次數-----50 圖八、實驗二d中兩類關鍵判斷的次數-----54 圖九、本研究所有單例程序實驗申兩類關鍵判斷的次數-----56 圖十、實驗三與實驗一雙例組兩類關鍵判斷的次數對照-----63 zh_TW dc.language.iso en_US - dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G91NCCU1102012 en_US dc.title (題名) 單例程序和雙例程序在導致人類負向關聯學習上的效果之再探 zh_TW dc.title (題名) Further studies on the usefulness of single and double discrimination procedures in inducing human vegative contingency learning en_US dc.type (資料類型) thesis en
