國小學童在條件句推理上所表現的偏誤分析

Analysis of conditional reasoning biases in primary students

郎慧瀛(Hui-Ying Lang)

指導教授 ： 林文瑛 ； 王琳

中原大學/理學院/心理學研究所/碩士(2015年)

https://doi.org/10.6840/cycu201500812

摘要

Manktelow（1999）認為人們在推理上容易受到信念偏誤、驗證偏誤、配對偏誤及否定結論偏誤的影響產生推理錯誤，在條件句推理中，信念偏誤會使推理者忽略邏輯法則，傾向認為擁有符合自身信念結論的條件句為有效論證；驗證偏誤使受試者傾向認為擁有符合自身信念小前提的條件句為有效論證；配對偏誤使受試者傾向認為，小前提或結論中有出現與大前提相同字詞的條件句為有效論證；否定結論偏誤使受試者在小前提為否定描述時，傾向認為否定描述的結論為有效論證。而根據過去研究顯示，推理偏誤的產生主要為受到過去經驗、邏輯知識及受試者本身認知資源的影響。
本研究採問卷實驗法，欲探討四種推理偏誤的發展趨勢及造成偏誤的可能原因，並假設信念偏誤及驗證偏誤主要受到過去經驗的影響，因此年齡越大，越容易認為符合信念組別為有效論證；配對偏誤及否定結論偏誤主要受到認知負荷量的影響，因此年齡越小，越容易認為重複出現過的組別為有效論證。受試者為台北市、桃園市、台中市各一所小學之二、四、六年級學生，採受試者間設計。
本研究實驗一：「結論內容效果：過去經驗」，假設隨著年齡與經驗增加，越容易認為結論符合信念的條件句為有效論證，故得分下降。結果顯示：過去經驗可能不是主要影響信念偏誤的原因。
本研究實驗二：「小前提內容效果：過去經驗」，假設隨著年齡與經驗增加，越容易認為小前提符合信念的條件句為有效論證，故得分上升。結果顯示：過去經驗可能不是主要影響驗證偏誤的原因。
本研究實驗三：「小前提與大前提間文字重複出現效果：認知資源」，假設隨著年齡與認知能力增加，推理者較能依邏輯法則進行推理，大、小前提文字重複出現組別得分將會上升。結果顯示：無法確認認知資源多寡是否為造成配對偏誤的主要原因。
本研究實驗四：「否定小前提與結論間文字配對效果：認知資源」，假設隨著年齡與認知能力增加，推理者較能依邏輯法則進行推理，擁有否定描述的結論組別得分會上升。結果顯示：認知資源可能不是影響推理表現的主要因素。
綜合本研究結果顯示，邏輯思考受到許多因素影響，包含過去經驗、邏輯知識以及認知資源等，雖本研究無法複製過去研究的結果，但從發展趨勢可發現推理者邏輯知識的影響效果。未來研究應先瞭解推理者對邏輯知識的瞭解程度，再進行過去經驗以及認知資源量的研究，如此才能瞭解真正影響推理表現的原因。

關鍵字

條件句推理 ； 信念偏誤 ； 驗證偏誤 ； 配對偏誤 ； 否定結論偏誤

並列摘要

Manktelow (1999) believes there are belief bias, confirmation bias, matching bias, and negative conclusion bias can be considered as influences in the process of reasoning. In conditional reasoning, belief bias makes people ignore the rules of logic, and tend to believe the conclusion more conform to their own belief is valid. Confirmation bias makes reasoners tend to believe that the minor term of a syllogism conform to their own belief is valid. Matching bias makes reasoners tend to believe the conditional has the same words between the minor term of a syllogism or conclusion and major term of a syllogism is valid. Negative conclusion bias makes reasoners making mistakes when the minor term of a syllogism describe negative, and tend to believe the conclusion describe negatively is valid. According to the past survey, the main reasons of having reasoning bias are from past experiences, logical knowledge and cognitive resource.
The aim of this study attempts to explore the trend of the four reasoning bias and the possible causes of making reasoning bias. We assume that belief bias and confirmation bias are influenced by the past experiences. As people age, they are disposed to believe the conditional conform to their own experiences is more valid; matching bias and negative conclusion bias are mainly affected by cognitive loading. That makes the younger generation are more likely to believe the condition with repeat words is valid. This research adopts survey-experimental method with between-subject design, participants are 2nd. 4th and 6th students.
Experiment 1 “Content effect of conclusion: Past experiences”. We assume the older participants are, the more experiences the have, and the easier they think conclusion of the conditional conform to their belief is valid, it decreases the correct response rate. The results revealed that past experiences may not be the main reason of belief bias.
Experiment 2 “Content effect of minor term of a syllogism: Past experiences”. We assume the older participants are, the more experiences they have, and the easier they think the conditional of the minor term of a syllogism conform to their own belief is valid conclusion, so the correct response rate increased. The results revealed that past experiences may not be the main reason of confirmation bias.
Experiment 3 “Matching of minor and major term of a syllogism: Cognitive resource”. We assume the older participants are, the more cognitive resource they have, and the easier follow the logical rules, so the correct response rate of matching group increased. The results revealed that the more or less cognitive resource is the main reason of matching bias.
Experiment 4 “Matching of negative minor term of a syllogism and conclusion: Cognitive resource”. We assume the older participants are, the more cognitive resource they have, and the easier follow the logical rules, so the correct response rate of matching group increased. The results showed that cognitive resource may not be the main reason of negative conclusion bias.
To conclude, past experiences and cognitive resource may not be the main reason of reasoning bias. However, the findings go against the past study, but we found the influence of logical knowledge does affect their performance. Future research should first test the logical knowledge of participants before testing other factors, so as to better understand the other reason of the reasoning bias.

並列關鍵字

conditional reasoning ； belief bias ； confirmation bias ； matching bias ； negative conclusion bias

參考文獻

洪文東（2003），創造性問題解決化學單元教學活動設計與評估。科學教育學刊，11，407-430。

蔡明珠（2010）。影響國小學童演繹推理能力相關因素之研究-以台北縣市六年級學童為例。台北市立教育大學數學資訊教育學系碩士論文，台北市。

謝佳鈺（2009）。國小學童邏輯推理能力養成之研究。國立台北教育大學數學教育研究所碩士論文，台北市。

Artman, L., Cahan, S., & Avni-Babad, D. (2006). Age, schooling and conditional reasoning. Cognitive Development, 21, 131–145.

Barrouillet P. & Lecas J. (2002). Content and context effects in children’s and adults’ conditional reasoning. The quarterly journal of experimental psychology, 55, 839–854.