360 likes | 472 Views
When does Choice Reveal Preference? Moderators of Heuristic versus Goal-based Choice AIMEE DROLET MARY FRANCES LUCE ITMAR SIMONSON. 學生 姓名 :吳承恩 學生座 號: 99353019 指導 老師:胡凱傑 老師. Reveal Preference. Heuristic. Simon(1957)-- 有限理性 (Bounded Bationality )
E N D
When does Choice Reveal Preference?Moderators of Heuristic versus Goal-based ChoiceAIMEE DROLETMARY FRANCES LUCEITMAR SIMONSON 學生姓名:吳承恩 學生座號:99353019 指導老師:胡凱傑 老師
Heuristic Simon(1957)--有限理性(Bounded Bationality) Bettman, Luce, and Payne(1998)—使用Heuristic決策的動機 Heuristic決策程序:態度—Eagly and Chaiken (1993) Heuristic決策程序:知覺,NFC(Need for Cognition)—Cacioppo et al. (1996)
Goal-Based Choice Self-Goal Choice:消費者會根據其價值觀、偏好、限制來判定任何有關產品實用程度和選擇後的滿意程度。(A、C點) CompromiseChoice:基於偏好不確定和有限理性的情況下,折衷選擇被認為是比較保險而且不會失誤的選擇。(B點)
Compromise Choice 舉例: 有A、B、C的烤肉架,三個烤爐比較之下,A的烤肉架SIZE比較大,C的烤肉架比較輕,而B烤肉架的屬性則若再A和C之間。 消費者選擇了B烤肉架(Compromise),因為他想減少買錯東西的風險以及後悔的情況。
背景 • 從前的研究著眼於:在依賴Self-Goal這種決策程序的情況下,認知資源(cognitive resources)和動機對做出正確的決策有何影響。 • 本篇研究: • 消費者的認知負荷(Cognitive Load)對高/低認知需求者(NFC)在決策過程(Heuristic /Goal-based Choice)上有何影響
Cognitive Load 消費者決策過程受到多重因素影響,若心有旁騖則購物時之選擇將趨於複雜,甚至有違個人實際目標或偏好。 Load存在的情況下會影響消費者的決策模式(存在干擾或是弱化能力的功能),使其決策過程變得更簡化。
NFC(Need for Cognition) NFC與動機存在著間接的關係,NFC是一種知覺的能力(cognitive ablity),是消費者做決策前深思熟慮的程度-Cacioppo and petty 1982 需求認知-高(HNFC) 需求認知-低(LNFC) • 願意花較多的時間去蒐集決策資訊 • 較依賴Self-Goal Choice • 喜歡抽象思考,解決複雜問題,對於接受的訊息會進行反覆的思考 • 比較沒有意願花時間收集決策資訊 • 較依賴Compromise Heuristic • 低喜歡簡單、需要較少認知資源的工作,且習慣依賴捷思和態度或是別人的意見。 文獻整理: Haugtvedt, petty, and Cacioppo 1992;Cacioppo et al. 1996; Pteey, Cacioppo, and Schumann 1983;
Study 1 Compromise Choice Load ? NFC
Study 1 作者推論: H1a:Under no load, HNFC predicts lower choice ofthe compromise option.non-Load的情況下HNFC會有較少的Compromise Choice。 H1b:Load will dampen the effect of (higher) NFCon (lower) choice of the compromise option.HNFC(with load)會比HNFC(without load)偏向選用較為簡化的決策過程。
Study 1-Method 受測者:128位UCLA的學生-Consumer decision-making學分為獎勵。
Study 1-ProtocolCoding 目的:We developed a protocol-coding scheme to test our assumption that load disrupts use of self-goalinformation. Interjudge reliability was 84%. We summed across the fivepossible types of goal-relevant statements (i.e., direct reference,local evaluation for size, local evaluation for weight,evaluative inferences for size, and evaluative inferences forweight) to create a six-level “goal mention” index; indexvalues varied between 0 and 5.
Study 1-Protocol Results 結果 Main effect of load (F(2,123)=5.48, p<.2) 在no-load的情況下受測者的Goal Mention Index(M=1.58)在比Load情況下高(M=0.75) Load對goal mention的影響不會受NFC操弄(F(1,123)=1.23, p=0.27) =>顯示出Load會減少self-goal的情況。
Study 1-Protocol Results 結果 根據羅吉斯回歸(Compromise(1),not compromise(0)) Load和NFC之間有顯著的交互效果(=4.08,p<.04) 在no-load的情況下,HNFC和Compromise之間有關係()=4.34, p<.04,estimate=-0.59) 在no-load的情況下,HNFC和Compromise之間沒有關係 ()=0.54, NS,estimate=-0.19)
Study 1-Protocol Results 結果 At 1.5 standard deviations above the mean forNFC, load marginally significantly increases compromise()= 2.76, p<.10, estimate=1.41) At 1.5 standard deviations below the mean, the direction of the load effect reverses()= 2.97, p<.10, estimate=-1.33)
Study 1-Choice results 根據羅吉斯回歸 Dependent variables: NFC和Load顯示出有交互作用()= 6.60, p<.01 )
Study 1-Choice results 在no-load的情況下,Slope analysis顯示NFC對Compromise有顯著的負相關()= 5.22, p<.05, estimate= -.60 )H1a:Under no load, HNFC predicts lower choice ofthe compromise option. 假設成立 在load的情況下,顯示NFC對Compromise沒有顯著的影響()= 1.71, p<.20, estimate=.30 )H1b:Load will dampen the effect of (higher) NFCon (lower) choice of the compromise option.假設成立
Study 1-Choice results At 1.5 standard deviations above the mean forNFC, load increases compromise ()= 5.49, p<.20, estimate=1.71) At 1.5 standard deviations below the mean, the direction of the load effect reverses()= 4.02, p<.05, estimate=-1.42)
Study 1-Results 總結Study1: Load會減少self-goal在做決策時資訊的使用。 Study1也顯示Load使HNFC(決策時偏好使用self-goal做決策)增加使用Compromise。
Study 2 從前研究顯示HNFC的人,其選擇會偏向個人的偏好,他們搜尋有用的訊息和self-goal,並且避免使用簡單的決策程序> H2:The interaction effect of load and NFC on compromise choice will be dampened by a manipulation that obscures the compromise relationship among choice set options. 操弄問項使表示的型態較模糊時會使Load和CompromiseChoice of NFC之間的交互作用減少。
Study 2-Method 受測者:145位UCLA修Consumer decistion-making的學生(每人給予$5報酬)
Study 2-Method 實驗設計: 本實驗為加入表達問項的透明程度(組間變異)和產品的等級—極端/妥協(組間設計)
Study 2-Results • 根據羅吉斯回歸 Dependent variables: • NFC和Load以及資料透明度間顯示出有交互作用()= 6.27, p<.01 ) • H2:The interaction effect of load and NFC on compromise choice will be dampened by a manipulation that obscures the compromise relationship among choice set options.支持H2
Study 2-Results )= 12.78, p<.01 )= .04, NS
Study 3 選擇Compromise選項並非一個最令人滿意的答案,選擇Self-goal則可以令人滿意。 Study3將討論產品屬性的比重:當產品屬性權重equal時,消費者會基於不確定性會傾向Middle option的選擇;但,當產品屬性權重不一致的時候,消費者會依照self-goal來選擇產品。 H3:There will be a significant interaction effect between load and NFC on compromise when at tribute importance weights are less equal but not when they are more equal.
Study 3 H4:The interaction effect of load and NFC on compromise will be dampened by a manipulation that reduces participants’ motivation to use self-goals.
Study3-Method 受測者:584位UCLA修 Consumer decistion-making的學生(每個學生給$5當獎勵)。
Study3-Method NFC(continuous),屬性比重(continuous) 組間變異 Design
Study 3-Method 前測: 受測者:n=80(1/4的機率隨機從本實驗中抽出)=>選中Load且Low-motivation instructions 受測者回答四個動機問項(Cronbach alpha=.083) F(1,79)=15.45, p<.0002
Study 3-Result Motivation instructions, load, or NFC 對受測者在產品屬性權重沒有顯著的影響。 Slopes for NFC at 1.5 standard deviations above and below the meanforimportant difference
Study 3-Result As expected, the slope of NFC significantly predicts lower compromise only when importance differences are high, low-motivation instructions are absent, and load is also absent.
Not All Resource Constrains Are Equal 未來的研究可以建立一個宏觀的方向: How Cognitive Capacitymaps onto decision outcome。
感謝各位的聆聽 各位,辛苦了