目標物、促發物與不重疊範圍對對比效果影響之研究: 解釋範圍重疊模型之延伸

全文

(1)國立臺灣師範大學管理學院管理研究所碩士論文 Graduate Institute of Management College of Management National Taiwan Normal University Master Thesis. 目標物、促發物與不重疊範圍對對比效果影響之研究: 解釋範圍重疊模型之延伸 How may Target Width, Context Width and Non-Overlap Range lead to differential Contrast Effect The Extension of Dimensional Range Overlap Model. 許益盛 I-Sheng HSU. 指導教授：蕭中強博士 Advisor：Chung-Chiang Hsiao, Ph.D.. 中華民國 106 年 7 月 July, 2017.

(2) 中文摘要此篇論文主要延伸(Chien, Wegener, Hsiao, and Petty 2010)所提出的「解釋範圍重疊模型」(The Dimensional Range Overlap Model, DROM)，他們認為決定對比效果是否發生是在於促發物的解釋範圍和目標物的解釋範圍不重疊。而我們此篇研究主要操弄了促發物與目標物解釋範圍不重疊的三大因素: 促發物的解釋範圍大小、目標物的解釋範圍大小以及促發物與目標物的解釋範圍之間的距離。因此，我們建立了三個假說模型，來驗證影響對比效果的關鍵因素。假說一: 在固定促發物的解釋範圍和目標物的解釋範圍情況下，促發物與目標物的解釋範圍之間的距離越遠，對比效果越強。假說二: 在固定目標物的解釋範圍情況和促發物與目標物的解釋範圍之間的距離，促發物的解釋範圍越大，對比效果越強。假說三: 在固定目標物的解釋範圍和目標物與促發物的最適代表值，在促發物與目標物解釋範圍不重疊情況且寬促發物範圍可涵蓋窄促發物範圍，我們預期較寬的促發物可以產生更強的對比效果，並且促發物在對比效果後的解釋範圍會變寬。而本研究經由實驗驗證了我們的假說一，具有顯著性效果，而假說二與假說三則是部分成立。此外，我們也發現促發物在對比效果後，促發物最適代表值產生了遠離目標物的現象，可以一併驗證相互假設的理論，促發物不僅會使目標物產生對比效果，且目標物也會給予促發物同等的對比效果，類似於牛頓第三運動定律，其效果相等，方向相反。未來可以再進一步探討比較促發物與目標物解釋範圍的不重疊範圍和促發物的解釋範圍對於對比效果之影響，期許藉由本研究的貢獻，更進一步延伸與推廣解釋範圍重疊模型之理論。. 關鍵字: 解釋範圍重疊模型、促發效果、同化效果、對比效果、相互假設 II.

(3) ABSTRACT People evaluate various things in their daily life, although they believe their judgments to be generally rational, the contest in which their judgments take place nevertheless impacts the choices they make which is so-called context effects. Many researchers have since concentrated on the issue of non-overlap of target and context will lead to contrast effects. However, we want to know more detail about people who form their judgments under different non-overlap range of target and context. Therefore, we are interested in how the target width, context width and non-overlap range lead to differential contrast effects. The present study based on the foundation of Dimensional Range Overlap Model (Chien, 2010) and Reciprocity Hypothesis (Hsiao, 2002) to extend the contrast effects under different target width, context width and non-overlap range. The effects of non-overlap range under same range of target and context are also examined to better know the mechanism of people evaluation. In this study, we tried to further explore the Dimensional Range Overlap Model and Reciprocity Hypothesis (Hsiao, 2002), under the condition that fixes the target width and non-overlap range, will wider or narrower context width lead to stronger contrast effect? Will the context width or target width change after contrast effect? What if we switch the target and context but set the other variables fixed? We’ll go through the experiments and try to create the scenario and verify the Reciprocity Hypothesis. Fortunately, the results support the hypotheses we proposed, we are able to III.

(4) demonstrate wider non-overlap range will lead to stronger contrast effect; under the condition that fixes the target width and non-overlap range, wider context width will lead to stronger contrast effects; despite switch the target and context, the contrast effect is same; the context width will change after contrast effect. In conclusion, the thesis offers a special view to better understand different contrast effect under the differential target width, context width and non-overlap range, also, extend the Reciprocity Hypothesis: despite switch the target and context, the contrast effect is same.. Key. words:. Priming. Effects,. Assimilation. Effects,. Dimensional Range Overlap Model, Reciprocity Hypothesis. IV. Contrast. Effects,.

(5) TABLE of CONTENT 中文摘要....................................................................................................................... II ABSTRACT ................................................................................................................ III TABLE of CONTENT ................................................................................................ V LIST of TABLES ..................................................................................................... VII LIST of FIGURES .................................................................................................. VIII Chapter 1 Study Purpose ............................................................................................ 1 Chapter 2 Literature Review ...................................................................................... 3 2.1 Priming Effects ............................................................................................... 3 2.2 Assimilation and Contrast Effects ................................................................ 3 2.3 The Dimensional Range Overlap Model ...................................................... 4 2.4 Reciprocity Hypothesis .................................................................................. 5 Chapter 3 Theory Foundation and Hypotheses ........................................................ 7 3.1 Theoretical Background ................................................................................ 7 3.2 Hypotheses ...................................................................................................... 8 3.3 Uniqueness of the current study ................................................................. 10 Chapter 4 Pretest ....................................................................................................... 12 4.1 The purpose of pretest ................................................................................. 12 4.2 Design of Pretest ........................................................................................... 12 4.3 Pretest Sample Size ...................................................................................... 13 4.4 Result of pretest............................................................................................ 13 Chapter 5 Main Experiment ..................................................................................... 15 5.1 Overview ....................................................................................................... 15 5.2 Participants and Design ............................................................................... 15 5.3 Manipulation checks .................................................................................... 16 5.4 Hypotheses test ............................................................................................. 24 Chapter 6 General Discussion................................................................................... 28 6.1 Conclusion Review ....................................................................................... 28 6.2 Experiment Limitation and Suggestion ..................................................... 30 6.3 Future Research and Creative Ideas .......................................................... 31 V.

(6) 6.4 Innovation of this study ............................................................................... 34 References ................................................................................................................... 36 Appendix (I): Pretest ................................................................................................. 39 Appendix (II): Main Experiment Stage 1 ................................................................ 56 Appendix (III): Main Experiment Stage 2............................................................... 73. VI.

(7) LIST of TABLES Table 4-4. The result of pretest (price dimension)……………………………..….14. Table 5-3-1 Manipulation check of Range width (Hypothesis 1)………….…… ..17 Table 5-3-2 Manipulation check of Non-overlap (Hypothesis 1)……………..........17 Table 5-3-3 Pair T test for Manipulation check (Hypothesis 1)…………...……….18 Table 5-3-4 Manipulation check of Range width (Hypothesis 2)……………..…...19 Table 5-3-5 Manipulation check of Non-overlap (Hypothesis 2)……………..........20 Table 5-3-6 Pair T test for Manipulation check (Hypothesis 2)………………...….20 Table 5-3-7 Manipulation check of Range width (Hypothesis 3)……………….....21 Table 5-3-8 Manipulation check of Non-overlap (Hypothesis 3)…………..........…22 Table 5-3-9 Manipulation check of Representative value (Hypothesis 3)……..…..22 Table 5-3-10 Pair T test for Manipulation check (Hypothesis 3)………………..…23 Table 5-4-1 Pair T test of Hypothesis test (Hypothesis 1)…………...…………...24 Table 5-4-2 Pair T test of Hypothesis test (Hypothesis 2)….…………..…….…..25 Table 5-4-3 Pair T test of Hypotheses test (Hypothesis 3)……………..………...26 Table 6-1. Summary of hypotheses and experiment outcomes………………….29. Table 6-2. The example for the future research scenario………………………..31. VII.

(8) LIST of FIGURES Figure 3-2-1. Manipulated Non-Overlap Condition (Hypothesis 1)……………..….8 Figure 3-2-2. Manipulated Non-Overlap Condition (Hypothesis 2)…………..….…9 Figure 3-2-3. Manipulated Non-Overlap Condition (Hypothesis 3)……………….10. VIII.

(9) Chapter 1 Study Purpose People access all kinds of things and need to form impressions in their daily life. They attempt to form fair judgments using only relative information of it; however, their judgments are actually influenced by irrelevant factors unconsciously. Although people believe their judgments to be generally rational, the contexts in which their judgments occur nevertheless impacts the choices they make, albeit unconsciously so. One of the well-known theories on explaining how the irrelevant factors impact people’s judgments is “priming effect”. “Priming effects” demonstrate that “latter” judgments are sensitive to the “former” context in which they are made (Herr, 1989). It means that there are several different interpretations of a “target object” resulting from the previously given information with so-called “the primed category”. When people receive a piece of information (contextual stimulus) prior to judgment, the contextual stimulus will activate their stored knowledge and thus influence their judgment of the specific target. The evaluation of a target stimulus may be either assimilated toward or contrasted away from the context. There are many studies to investigate what factors and how the priming will affect the evaluation of target. For example, the extremity and distinctness of the activated information and the primed category (Herr, 1983), priming attributes versus target category exemplar (Stapel, 1998), and the inclusion/ exclusion model (Schwarz & Bless, 1992), etc. However, despite the fact that many researchers have devoted 1.

(10) on the issue, there is still discrepancy to explain the direction of priming effects. Therefore, the “Dimensional Range Overlap Model” (Chein, 2010) provided a more general view to clarify the reasons behind the occurrence of assimilation and contrast. The model suggests that the direction of priming effects shall be determined by the overlap or non-overlap between the target’s range and the contextual stimulus’s range. Assimilation occurs when there is an overlap, whereas non-overlap will result in contrast effects. However, under the process of literature review, we still can’t tell that how may target width, context width and non-overlap range lead to differential contrast effect. Therefore, we set many kinds of conditions to further explain the Dimensional Range Overlap Model.. 2.

(11) Chapter 2 Literature Review 2.1 Priming Effects When people make judgments, they give evaluations not only based on the facts, but also unconsciously influenced by the context. Many studies have shown that contextual information activates people’s stored knowledge, and thus leads to a passive and automatic processing which affects their judgment (Higgin and King 1981). The term “priming effect” refers to the influence of priming on target evaluation. Even without awareness, the primed category still increase accessible information such that stored knowledge is more likely to be activated and to be used subsequently in processing of target information (Higgin and King 1981). The process can be further elaborated in three steps by Fazio, Powell and Herr (1983) as first, the presented prime can make a person retrieve attitude from his memories; second, through the process of selective perception, the accessed attitude will impact the person’s interpretation of the target; and last, the altered interpretation will influence his behavioral response. Furthermore, the primes will be used automatically without awareness. That is, people will affected by primes unconsciously and they will use the primes as a standard to judge the following information (Herr, et al. 1983). 2.2 Assimilation and Contrast Effects Priming increases the accessibility of available knowledge such that the stored knowledge is more likely to be activated and used in subsequent processing 3.

(12) of stimulus information (Higgins, 1996).As the result of priming effects, we can distinguish the subsequent judgment into two parts, which are assimilation effects and contrast effects. Assimilation effect occur when moderate exemplars are primed and ambiguous stimuli are judged. The concept is that when exemplars are moderately extreme, ambiguous stimuli are likely to be categorized as instances of accessible category. Contrast effect occur when extreme exemplars are primed, ambiguous stimuli are judged. Besides, judgments make away from the context occurred when the judged stimulus are unambiguous. The idea is that under such circumstances, the prime is likely to serve as a standard of comparison for judgments.. 2.3 The Dimensional Range Overlap Model Although many context effects theories have been developed, Chien, Wegener, Hsiao and Petty (2010) has proposed a more comprehensive model to explain the underlying mechanism of context effects. The Dimensional Range Overlap Model suggests that assimilation and contrast effects are determined by the overlap or non-overlap between the target’s range and the prime’s range on the relevant judgment dimension. When context and target ranges overlap, assimilation occurs (as more similarity between context and target), but the ranges do not overlap, contrast occurs (as less similarity between context and target). Besides, according to this model, if the target and the context cannot be related to 4.

(13) the same judgmental dimension, the accessible context would not be able to yield any context effect on target judgments. 2.4 Reciprocity Hypothesis Past researches in Context Effect only concentrate on how perception of a target stimulus affected by different kind of contextual stimulus. Nevertheless, Hsiao (2002) proposed the Reciprocity Hypothesis and suggested that not only the target shall shift when there’s a prime, the perception of context itself shall be affected as well by the target. While a contextual prime makes the target contrast away, meanwhile, the context will also move from the target. Reversely, judgment of contextual stimulus shifted toward the target would be the case of assimilation effects. Two studies are implemented to support the Reciprocity Hypothesis. Study 1 tried to test the hypothesis that under the condition of contrast effect, contextual stimulus would be also affected by target stimulus and contrast away from it. There are three groups of students taking part in this study. Group 1 evaluated the target stimulus only, group 2 was asked to evaluate the contextual stimulus and group 3 evaluated both the target stimulus and contextual stimulus. The evaluation of group 1-2 can be served as the baseline for target and context, and thus can be used to detect the perception shifts. The results showed as predicted, contextual stimulus and target stimulus contrasted away from each other. In study 2, the procedure of the experiment is similar to study 1. The selected 5.

(14) target and context are determined to have overlaps so assimilation is expected, hence, the results shows that contextual stimulus and target stimulus assimilated toward each other. Assimilation effects are observed. Through two experiments, the Reciprocity Hypothesis is proved, noticing the shift of context rather than only focusing on the target. When either assimilation or contrast effects occurs, it’s not only the target shift, but the context does, too.. 6.

(15) Chapter 3 Theory Foundation and Hypotheses 3.1 Theoretical Background Succeeding the literature review, we would illustrate the theory foundation that supported us to develop the hypotheses in this study. According to the proposition of the Dimensional Range Overlap Model, no matter what kind of targets and contexts are used, the model can be applied as long as both targets and contexts are on the same dimension. Therefore, thanks to the generalizability of the Dimensional Range Overlap Model, the main theory foundation of this thesis will follow the proposition of Dimensional Range Overlap Model: Irrespective of which type of target and context, it is presence or non-presence of overlap between interpretation range of target and context. If the Dimensional Range Overlap Model is supported again in our research under different width of targets and contexts, we can further extended current model. In addition, based on Reciprocity Hypothesis (Hsiao, 2002), not only is the target judgment affected by the earlier contextual stimulus, but the judgment of contextual stimulus is affected by the latter target stimulus. Hence, we presume that the magnitude of contrast effect will be same when given the switch of target and context. Reciprocity Hypothesis provides us a broader perspective to understand what the process underlying the switch of target and context. In this thesis, we will also verify the Reciprocity Hypothesis and further explain the 7.

(16) current hypothesis. 3.2 Hypotheses Hypotheses 1: Under the condition that fixes the target width and context width, wider non-overlap range will lead to stronger contrast effect.. Figure 3-2-1. Manipulated Non-Overlap Condition (Hypothesis 1) In hypothesis 1, we would go through the experiment and verify that the magnitude of contrast effect will increase by the non-overlap between target and context. The anchor effect can also explain the phenomenon. If hypothesis 1 can be verified, we can extend the Dimensional Range Overlap Model.. Hypotheses 2: Under the condition that fixes the target width and non-overlap range, wider context width will lead to stronger contrast effect.. 8.

(17) Figure 3-2-2 Manipulated Non-Overlap Condition (Hypothesis 2) In hypotheses 2, following the verification of wider non-overlap range will lead to stronger contrast effect. In hypothesis 2, the yellow area will lead to more magnitude of contrast effects, we would expect that given the same width of target and non-overlap range between target and context but different width of context, wider contest range will lead to more contrast magnitude.. Hypotheses 3: Under the non-overlap condition that fixes the target width and the distance of representative value of target and context, wider context width will lead to stronger contrast effect and context width will wider.. 9.

(18) Figure 3-2-3 Manipulated Non-Overlap Condition (Hypothesis 3). In hypothesis 3, according to the previous conclusion, wider non-overlap range will lead to stronger contrast effects. So, point B will make more magnitude of contrast effect than point A; point C will make more magnitude of contrast effect than point D. Therefore, we expect that the context width will become wider than before. 3.3 Uniqueness of the current study After presenting the structure of our hypotheses, the distinctive features of the current research from previous related studies are as following: 1. Possible Range of Interpretation When it comes to evaluation, past studies only focus on single point estimate on judgment dimension, which might solely represent the central tendency of 10.

(19) evaluation. So, following the Dimensional Range Overlap Model, we use the interval estimate, to represent the plausible values in consumer judgment settings. 2. Overlap and Non-Overlap Instead of using different kind of target and prime to explain direction of priming effect, this thesis mainly base on the Dimensional Range Overlap Model. In this model, it’s assumed that when there’s overlap between target and prime, assimilation is more likely to occur, while when there’s non-overlap, contrast is expected. 3. Reciprocity Hypothesis In explanation of context effects, most researchers merely notice that presented context will influence the target to shift, but actually the context itself will be influence by the target. As Reciprocity Hypothesis (Hsiao 2002) suggests, context and target will move toward each other when assimilation happen, while context and target will shift away from each other when contrast effect take place. This concept is adopted in this research to provide a more complete view to explain the same magnitude of contrast effect when switch the target and context.. 11.

(20) Chapter 4 Pretest 4.1 The purpose of pretest According to our three hypotheses above, we need to test the target participant’s interpretation range of a number of materials on certain dimension and select suitable dimensional range of target and context to prepare the subsequent main experiment. 4.2 Design of Pretest At the beginning, we had one booklet for the pretest. The first page of the questionnaire explained that the study was in order to investigate people’s comprehension ability. The participants would be going to read several product advertisement, which were randomly extracted from the market. The questionnaire is separated in three parts, the first part is an introduction of the measurement, the second part is a practice, and the last part is the main test. In the second part, there were two exemplifications, and participants could learn how to answer and how to measure them. In the last part, participants were asked to evaluate the representative value and their interpretation for a certain dimensions (price and prestige). The measurement scale used is from 0 (the least extreme) to 10 (the most extreme) in interval of 0.5, for example, in the dimension of “price”, 0 stands for the cheapest, and 10 represents for the most expensive (See Figure 4-1). After finishing all tasks, the participants were thanked and dismissed. 12.

(21) Figure 4-1 An example of the measurement scale for the dimension of “Price”. 4.3 Pretest Sample Size Participants are mainly composed by undergraduates and graduates of the college of Management in National Taiwan Normal University. They were requested to answer questions similar to the above intuitionally. As the result, the valid number of questionnaire is fifteen. 4.4 Result of pretest Manipulation in the made-up target’s advertisement (Belissimo) seems quite effective in this pretest. Table 4-4 in the below will give a clearer illustration.. 13.

(22) Table 4-4 The result of pretest (price dimension) Brand. Ave Rep.. Ave Dis.. Ave Non-Over.. Ave Range. Belissimo. 4.4. N/A. N/A. 3.2-5.2. Timberland. 7.5. 3.1. 1.7. 6.9-8.2. GU. 5.3. 0.9. -0.9. 4.4-6.5. Coach. 8.0. 3.6. 1.9. 7.1-9.1. Note. Ave Rep. =Average representative value Ave Dis. =Average Representative value of context – Average Representative of target Ave Non-Over. = Average Context lower bound – Average Target higher bound Ave Range = Higher bound – Lower bound. Based on above table, participants evaluated the target (Belissimo) as we expected that approximately fell into 4-5. Although GU was averagely rated 4.46.5, it was not as we expected that fell into 6-7. Nevertheless, the results of Timberland and Coach were close to our expectation that Timberland fell into 7-8 and Coach fell into 8-9. What’s more, we found that two of pretest participants were suitable for our hypotheses 1; one was suitable for hypotheses 2 and one was suitable for hypotheses 3. In addition, the result of analyzing whether there was an order effect exhibited disposition of the prime didn’t affect participants’ evaluation.. 14.

(23) Chapter 5 Main Experiment 5.1 Overview The study was designed to see how may target width, context width and nonoverlap range lead to differential contrast effect. The main experiment was within-participants design, and was divided into two stages. In stage 1, we investigated the participant’s attitude toward the target and contexts’ representative value and boundaries. After the stage 1 of main experiment, we selected those participants who were suitable for our three hypotheses conditions. In stage 2, we customized the questionnaire for each participant. The participants would see the context’s advertisement first then target’s, following, they would judge the target’s representative value and boundaries then context’s. 5.2 Participants and Design 150 undergraduate and graduate students from National Taiwan Normal University participated the first stage of the main experiment. They were rewarded with a gift and the chance to win free coupons. Thanks to the result of pretest. We used the same target and contexts of pretest questionnaire to be our stage 1 main experiment’s target and contexts. The first part shows the instructions; the second part is practice questions; and the last part contains one fictitious target, four contexts and two fillers. The order was Belissimo(target) first; Timberland(context 1), GU(context 2) and Coach(context 3) would be the second, fourth and sixth; OPPO(filler 1) and Mos Burger(filler 2) 15.

(24) would be the third and fifth. The participants were inquired to evaluate representative value, upper and lower bounds. Both interpretation range and representative value were measured on a 10 point scale with interval of 0.5. Our 3 hypotheses involved with 2 stages, asking every participant to answer two questionnaires in 4 week. After two weeks from Stage 1, the screened participants were continued to the stage 2 of main experiment. In stage 2, we customized the questionnaire for each participant. The participants would see the context’s advertisement first then target’s, following, they would judge the target’s representative value and boundaries then context’s. In the end, the participants were thanked and dismissed after all 2 questionnaires were finished in 4 weeks. 5.3 Manipulation checks 5.3.1 Hypothesis 1 In order to check our selected participants from hypothesis 1 target and contexts manipulation, which was under the condition that fixed the target width and context width. We expected that wider non-overlap range would lead to stronger contrast effect. We used the pair T test to compare Stage 1 target’s range and Stage 1 context’s range; Stage 1 target and context’s non-overlap range (narrow and wide). Table 5-3-1, Table 5-3-2 and Table 5-3-3 in the below will give a clearer illustration.. 16.

(25) Table 5-3-1 Manipulation check of Range width(Hypothesis 1) Range Width Non-Overlap. Target. Context. Wide. 1.6389. 1.6389. (.61371). (.61371). 1.6389. 1.6389. (.61371). (.61371). Narrow. Table 5-3-2 Manipulation check of Non-overlap(Hypothesis 1) Non-Overlap Wide. 2.3056 (.73041). Narrow. 1.2778 (.64676). Note: M=(2.3056 v.s 1.2778), pair t(17)=20.955, p***<.001.. 17.

(26) Table 5-3-3 Pair T test for Manipulation check (Hypothesis 1) Pair 1. 2. 3. Var.. Mean. N.. S.D.. S1_wide_nonoverlap. 2.3056. 18. .73041. S1_narrow_nonoverlap. 1.2778. 18. .64676. S1_wide_target_range. 1.6389. 18. .61371. S1_narrow_target_range. 1.6389. 18. .61371. S1_wide_context_range. 1.6389. 18. .61371. S1_narrow_context_range. 1.6389. 18. .61371. No matter wide or narrow scenario, target and context’s range were identical (M=1.6389). The selected target and context’s wide non-overlap one were manipulated to be larger than narrow one. M=(2.3056 v.s 1.2778), pair t(17)=20.955, p<.001. The result revealed these selected participants were successfully suitable for our stage 2 of hypothesis 1.. 5.3.2 Hypothesis 2 In order to check our selected participants from hypothesis 2 target and contexts manipulation, which was under the condition that fixed the target width and 18.

(27) non-overlap range. We expected that wider context width would lead to stronger contrast effect. We used the pair T test to compare Stage 1 target’s range and Stage 1 context’s range; Stage 1 target and context’s non-overlap range (narrow and wide). Table 5-3-4, Table 5-3-5 and Table 5-3-6 in the below will give a clearer illustration.. Table 5-3-4 Manipulation check of Range width(Hypothesis 2) Range Width Context width. Target. Context. Wide. 1.8182. 1.9318. (1.02986). (.35508). 1.8182. 1.1818. (1.02986). (.39477). Narrow. Note: M=(1.9318 v.s 1.1818), pair t(21)=4.054, p**<.05.. 19.

(28) Table 5-3-5 Manipulation check of Non-overlap(Hypothesis 2) Non-Overlap Wide. 1.5682 (.67780). Narrow. 1.5682 (.67780). Table 5-3-6 Pair T test for Manipulation check of hypothesis 2 Pair 1. 2. 3. Var.. Mean. N. 22. S.D.. S1_wide_nonoverlap. 1.5682. .67780. S1_narrow_nonoverlap. 1.5682. 22. .67780. S1_wide_target_range. 1.8182. 22. 1.02986. S1_narrow_target_range. 1.8182. 22. 1.02986. S1_wide_context_range. 1.9318. 22. .35508. S1_narrow_context_range. 1.1818. 22. .39477. No matter wide or narrow scenario, non-overlap range were identical (M=1.5682). The selected wide context’s range was manipulated to be larger than. 20.

(29) narrow one. M=(1.9318 v.s 1.1818), pair t(21)=13.748, p<.001. The result revealed these selected participants were successfully suitable for our stage 2 of hypothesis 2. 5.3.3 Hypothesis 3 In order to check our selected participants from hypothesis 3 target and contexts manipulation, which was under the non-overlap condition that fixed the target width and the distance of representative value of target and context. We expected that wider context width would lead to stronger contrast effect and context width would wider. We used the pair T test to compare Stage 1 target’s range and Stage 1 context’s upper bound and lower bound; Stage 1 target and context’s nonoverlap range (narrow and wide); Stage 1 target and context’s representative value. Table 5-3-7, Table 5-3-8 and Table 5-3-9 in the below will give a clearer illustration.. Table 5-3-7 Manipulation check of Range width (Hypothesis 3) Context Width. Wide. Narrow. Target Width. Upper bound. Lower bound. 9.7624. 7.5333. 1.8000. (.41871). (.51640). (.41404). 9.2000. 7.9667. 1.8000. (.31623). (.63994). (.41404). 21.

(30) Note: M=(9.7624 v.s 9.2000), pair t(14)=3.609, p*<.01. M=(7.5333 v.s 7.9667), pair t(14)=-3.166, p*<.01. Table 5-3-8 Manipulation check of Non-overlap(Hypothesis 3) Non-Overlap Wide. 2.4667 (.83381). Narrow. 2.9000 (.92967). Note: M=(2.4667 v.s 2.9000), pair t(14)=-4.373, p*<.01. Table 5-3-9 Manipulation check of Representative value(Hypothesis 3) Representative value Wide. 8.5667 (.49522). Narrow. 8.5667 (.49522). 22.

(31) No matter wide or narrow scenario, target’s range(M=1.8000) and context’s representative value(M=8.5667) were identical. The selected target and context’s were manipulated to have non-overlap (M=2.4667 v.s. 2.9000) and wide context’s upper bound was higher than narrow context’s upper bound; narrow context’s lower bound was higher than wide context’s lower bound. M=(2.4667 v.s. 2.9000) pair t(14)=-3.166, p<.001. The result revealed these selected participants were successfully suitable for our stage 2 of hypothesis 3.. Table 5-3-10 Pair T test for Manipulation check (Hypothesis 3) Pair. Var.. Mean. N.. S.D.. 1. S1_wide_context_upper. 9.7624. 15. .41871. S1_narrow_context_upper. 9.2000. 15. .31623. S1_wide_target_lower. 7.5333. 15. .51640. S1_narrow_target_lower. 7.9667. 15. .63994. S1_wide_nonoverlap. 2.4667. 15. .83381. S1_narrow_nonoverlap. 2.9000. 15. .92967. S1_wide_context_rep. 8.5667. 15. .31997. S1_narrow_context_rep. 8.5667. 15. .31997. 2. 3. 4. 23.

(32) 5.4 Hypotheses test 5.4.1 Hypothesis 1 We expected that wider non-overlap range would lead to stronger contrast effect. M=(3.9722 v.s 2.8056), pair t(17)=8.332, p<.001 (Table 5-4-1). Target was as expected had contrast effect and the distance of representative value was wider than stage 1. No matter wide or narrow non-overlap, stage 2 were all wider than stage 1. What’s more, target indeed shifted in stage 2. We also found that wider nonoverlap would lead to stronger contrast effect, wide target shift was larger than narrow target shift. As the result, hypothesis 1 was confirmed. In addition, the contexts also shifted, although the p-value was insignificant, partially verified the Reciprocity Theory.. Table 5-4-1 Pair T test of Hypothesis test (Hypothesis 1) Pair 1***. Var.. Mean. N.. S.D.. S2_wide_nonoverlap. 3.9722. 18. .71686. S2_narrow_nonoverlap. 2.8056. 18. .75027. -1.1944. 18. .54608. narrow_targetshift. -.6944. 18. .68897. wide_contextshift. .8056. 18. .57238. 2*** wide_targetshift. 3. 24.

(33) narrow_contextshift. .7500. 18. .52159. 4*** wide_distance_shift. 2.0000. 18. .93934. 1.4444. 18. 1.08314. narrow_distance_shift Note.. Targetshift = Stage 2 target representative value - Stage 1 target representative value Contextshift.= Stage 2 context representative value - Stage 1 context representative value Distance_shift. = (Stage 2 context rep.- Stage 2 target rep.) – (Stage 1 context rep. – Stage 1 target rep). 5.4.2 Hypothesis 2 We expected that wider context width would lead to stronger contrast effect. M=(2.1675 v.s 1.1537), pair t(21)=.489, p=.631. Only wide scenario had stronger contrast effect, but narrow one didn’t (M= 2.1675 v.s. 1.5682)(M=1.1537 v.s. 1.5682). Unfortunately in our hypothesis 2, the hypothesis was partially confirmed. (wide targetshift P*<.01). In addition, we found that context range no matter in wide or narrow scenario shifted, could verify the Reciprocity Theory.. Table 5-4-2 Pair T test of Hypothesis test (Hypothesis 2) Pair 1. Var. S2_wide_nonoverlap. 25. Mean. N.. S.D.. 2.1675. 22. .97596.

(34) S2_narrow_nonoverlap. 1.1537. 22. .95835. -.2437. 22. .84950. narrow_targetshift. .4944. 22. .88934. wide_contextshift. .3856. 22. .86744. narrow_contextshift. .0891. 22. .93628. wide_distance_shift. 0.6293. 22. .96934. narrow_distance_shift. -0.4053. 22. .58736. 2* wide_targetshift. 3. 4. Note. Targetshift = Stage 2 target representative value - Stage 1 target representative value Contextshift.= Stage 2 context representative value - Stage 1 context representative value Distance_shift. = (Stage 2 context rep.- Stage 2 target rep.) – (Stage 1 context rep. – Stage 1 target rep). 5.4.3 Hypothesis 3 We expected that wider context width would lead to stronger contrast effect and context width would wider. M=(4.1333 v.s 4.0000), pair t(14)=.718, p=.484.(Table 5-4-3) Target was as expected had contrast effect but p-value wasn’t significant. Both wide and narrow non-overlap range were larger than stage 1. What’s more, both wide and narrow targets indeed shifted in stage 2 and wide target representative value shift was larger than narrow one (M=-1.6667 v.s. -.9667). We also found that wider context range would lead to stronger contrast effect. 26.

(35) (M=4.1333 v.s. 4.0000), but the contexts’ range weren’t wider than stage 1 (M=2.4667 v.s 1.6667; M=1.2333 v.s. 1.0667). As the result, hypothesis 3 was only partially confirmed.. Table 5-4-3 Pair T test of Hypothesis test (Hypothesis 3) Pair 1. Var.. Mean. N.. S.D.. S2_wide_nonoverlap. 4.1333. 15. .69351. S2_narrow_nonoverlap. 4.0000. 15. .59761. -1.6667. 15. .67259. narrow_targetshift. -.9667. 15. .48058. wide_contextshift. .4000. 15. .20702. .4000. 15. .43095. S1_wide_context_range. 2.4667. 15. .51640. S2_wide_context_range. 1.6667. 15. .58757. S1_narrow_context_range. 1.2333. 15. .49522. S2_narrow_context_range. 1.0667. 15. .17593. 2* wide_targetshift. 3. narrow_contextshift 4. 5. Note. Targetshift = Stage 2 target representative value - Stage 1 target representative value Contextshift.= Stage 2 context representative value - Stage 1 context representative value Distance_shift. = (Stage 2 context rep.- Stage 2 target rep.) – (Stage 1 context rep. – Stage 1 target rep). 27.

(36) Chapter 6 General Discussion 6.1 Conclusion Review The critical issue of this research was to test how may target width, context width and non-overlap range lead to differential contrast effect. Based on the purpose, we created different kinds of scenario to verify our hypotheses. What’s more, to apply Dimensional Range Overlap Model in a more generalized way, within participant design was adopted. The result in our experiment partially support the hypotheses derived from Dimensional Range Overlap Model and Reciprocity theory. First, in hypothesis 1, we went through the experiment and verify that the magnitude of contrast effect increase by the non-overlap between target and context. We confirmed that under the condition that fixed the target width and context width, wider non-overlap range would lead to stronger contrast effect. Second, in the hypothesis 2, we found that target indeed shifted in the end (P*<.01). As a result, we would like to say in our experiment partially supported under the condition that fixed the target width and non-overlap range, wider context width would lead to stronger contrast effect. We found that wider context would lead to stronger contrast effect but in the narrow scenario it had assimilation effect. Maybe we had sample selection bias so we should choose the other sample to conduct the survey again. Third, in hypothesis 3, the experiment results partially confirmed our expectation that under the non-overlap condition that fixed the target width and the distance of representative value of target and context, wider context width would lead to stronger contrast effect and context width would wider. The reason why was we 28.

(37) found target indeed shifted in the end. (P*<.01) Based on the above principle, the outcomes in the current study generally supported the Dimensional Range Overlap Model, and even strengthened the model more tenable and extensive. Disputes about the Reciprocity theory were also clarified successfully. In sum, the summary of hypotheses results will show in Table 6-1. Also, we assumed that if we could modify the current experiment design, it’s more likely that the current second hypotheses could be verified.. Table 6-1 Summary of hypotheses and experiment outcomes Predictions. What was found. Hypotheses 1: Under the condition that fixes the. Confirmed. target width and context width, wider non-overlap range will lead to stronger contrast effect.. Hypotheses 2: Under the condition that fixes the. Partially Confirmed. target width and non-overlap range, wider context width will lead to stronger contrast effect.. Hypotheses 3: Under the non-overlap condition that fixes the target width and the distance of representative value of target and context, wider context width will lead to stronger contrast effect and context width will wider. 29. Partially Confirmed.

(38) 6.2 Experiment Limitation and Suggestion 6.2.1 Sample Size In the stage 2 of main experiment, for our special condition need, we decreased a lot of samples, especially in the manipulated hypothesis 3. The reason was that it’s hard to find the distance of representative value of target and context in narrow condition and wide condition. Hence, there were only fifteen valid samples in manipulated hypothesis 3. If there was more ample time and efforts in future related studies, effective samples could be collected as many as possible to make analyses more tenable and plausible. 6.2.2 Within Participant Design In this research, we noticed that the phenomenon of systematic shift due to the fatigue of participants. The reason why we adopted within participant design was because we have considered the individual variance across participants, so we assumed that if people could use what they have evaluated to be their target and context, the shift of the perception could be reflected through the process. Nevertheless, we found that the participants didn’t do the questionnaires in exchange of credit but rather for “kindness”, so when they’re required to do questionnaires for two times four weeks, they significantly lost patience and only wanted to answer the questions with less efforts, which might result in thoughtless and careless ratings of representative value and interpretation of target and context. 30.

(39) Besides, the repeated measure made the participants who were familiar with the design of questionnaire, so in order to save time, they might directly jump to evaluate the representative value and interpretation but ignore writing sentences for the reason why they chose those numbers. Therefore, even though within design has the advantage of allowing us to collect individual perception, we should not neglect the disadvantages that might significantly influence the outcome, which would be the bias of data and unable to reflect their true perception. 6.2.3 Questionnaire Design Generally speaking, the questionnaire was quite long itself, containing 20 pages. If participants were asked to focus on all details on the questionnaire, it would take 15 to 25 minutes. Second, in order to let the participants to be familiar with the answering skills of our question, such as point estimation and range estimation, the first part of the main experiments was an instruction and a practice of four questions. After reading and practicing, the participants seemed to lose their patience when they’re just about to give evaluation to the target and contexts. Hence, this kind of long questionnaire made them fatigue and no patience to read the whole questionnaire sentence by sentence, so their given perception might not reflect their true thoughts. 6.3 Future Research and Creative Ideas As previous discussed, there were several limitation in the current research method. In order to eliminate the unexpected influence on data, following are some 31.

(40) suggestions. First, even though within participant design requires for 2 lengthy questionnaires in a series will easily make participants fatigue, but it’s still a valuable way. We can use the individual perception to generalize Dimensional Range Overlap Model scenario. In this way, we suggest within participant design can be modified. One way is to eliminate the explanation sentence part of each evaluation, and the other way is to ensure that they can be motivated to devote to the questionnaires, for example, by changing for a credit in course. In our experiment result of hypothesis 1, under the condition that fixes the target width and context width, wider non-overlap range will lead to stronger contrast effect is confirmed. We verify that the magnitude of contrast effect will increase by the non-overlap between target and context. But we still can’t to clarify whether or not the target and context’s range influence the contrast effect. For example, fix the narrow non-overlap is 1; wide non-overlap is 1.5, how may different target and context range lead to differential contrast effect (Table 6-2)?. Table 6-2 The example for the future research scenario Narrow Non-Overlap(1) Target range Context range Scenario 1. Wide Non-Overlap(1.5) Target range Context range. 5-6. 7-8. 5-6. 7.5-8.5. 4.5-6. 7-8.5. 4.5-6. 7.5-9. (range=1) Scenario 2. 32.

(41) (range=1.5) Scenario 3. 4-6. 7-9. 4-6. 7.5-9.5. (range=2). Second, in hypothesis 2, following the verification of wider non-overlap range will lead to stronger contrast effect (Hypothesis 1). The extra area of wider context range in theory should lead to more magnitude of contrast effects (Hypothesis 2), we would say that given the same width of target and non-overlap range between target and context but different width of context, wider contest range will lead to more contrast magnitude. Maybe due to the limitation of experimental design or selection bias, we only partially verified our second hypotheses as expected, however, future research can manipulate the condition again and try to confirm it. According to the Reciprocity Hypothesis (Hsiao, 2002), we’ll expect that given the switch of target and context, the magnitude of contrast effect will also be the same. To further explain and verify the Reciprocity theory, future research can extend our current model. For example, given the switch of target and context of all our hypotheses, will the magnitude of contrast effect also be the same and will the consequence approximately be symmetric to our current finding?. 33.

(42) 6.4 Innovation of this study After presenting the structure of all our hypotheses and results, the distinctive features of the current research from previous related studies are as following: Firstly, our research verified the contrast effect in the condition that target and context were non-overlap, wider non-overlap would lead to stronger contrast effect. We manipulated the target and context’s width and non-overlap range. This is the first thesis to conduct the survey and verify the contrast effect and dimensional range overlap model. Secondly, we tried to clarify whether or not context range would influence the contrast effect, so in our second hypothesis, we fixed the target width and non-overlap range to see the how may context range influence the contrast effect, although only partially confirmed our hypothesis that wider context would lead to stronger contrast effect. Future research can further expand this hypothesis and do the experiment. Thirdly, we tried to clarify the reciprocity in our hypothesis 3. We wanted to know would context range wider after the contrast effect, although only partially confirmed our hypothesis that context range would become wider than before. It is the first study tried to clarify this issue. Further, from the manipulation procedure of all our hypotheses, this research more steadied the Dimensional Range Overlap Model and Reciprocity Theory. In spite of not all the hypotheses were confirmed, this study initiated the new hypotheses to expand the dimensional range overlap model, which contributed to the field of 34.

(43) contrast effect by fixing target and context’s width, non-overlap range and representative value of target and context.. 35.

(44) References Brown, D. R. (1953). Stimulus-similarity and the anchoring of subjective scales. The American Journal of Psychology, 66, 199-214. Bassili, John N. (1996), Meta-judgmental versus Operative Indexes of Psychological Attributes: The Case of Measures of Attitude Strength, Journal of Personality and Social Psychology,71 (4), 637–53. Chien, Y. and Hsiao, C. (2001), Dimensional Range Overlap Model for Explanation of Contextual Priming Effects, Proceedings of the Association for Consumer Research 2000 Annual Conference. Chien, Y. (2002). Dimensional range overlap model for explanation of contextual priming effects on product judgments. A thesis submitted to the faculty of Purdue University. Chien, Y., Wegener, D., Hsiao, C., and Petty, R., (2010), Dimensional Range Overlap and Context Effects in Consumer Judgments, Journal of Consumer Research, 37(3), 530-542 Fazio, R. H., Powell, M.C., and Herr, P. M. (1983). Toward a process model of the attitude-behavior relation: Accessing one’s attitude upon more observation of the attitude object. Journal of Personality and Social Psychology, 44, 723-735 Herr, P. M., Sherman, S. J., and Fazio, R. H. (1983). On the consequences of priming: Assimilation and contrast effects. Journal of Experimental Social 36.

(45) Psychology, 19, 232-340 Herr, P. M. (1986). Consequences of priming: Judgment and behavior. Journal of Personality and Social Psychology, 51, 1106-1115. Herr, P.. M. (1989). Priming price: Prior knowledge and context effects. Journal of Consumer Research, 16, 67-75 Hsiao, C. (2002). The Reciprocity Hypothesis As an Explanation of Perception Shifts in Product Judgment. Dissertation, Purdue University. Joan Meyers-Levy, B. S. (1993). A two-factor explanation of assimilation and contrast effects. Journal of Marketing Research, 30(3), 359-368. Kruglanski, Arie W. (1975), The Human Subject in the Psychology Experiment: Fact and Artifact. Advances in Experimental Social Psychology, Vol. 8, ed. L. Berkowitz, Orlando, FL: Academic Press, 101-147 Meyers-Levy, J., (1989). Priming Effects on Product Judgements: A Hemispheric Interp. Journal of Consumer Research, 16, 76-86. Meyers-Levy, J., & Sternthal, B. (1993). A two-factor explanation of assimilation and contrast effects. Journal of Marketing Research, 30, 359-368. Petty, Richard E., John T. Cacioppo and David Schumann (1983), “Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement,” Journal of Consumer Research, 10(September), 2, 135-146. Petty, Richard E. and Duane T. Wegener (1993), Flexible Correction Processes 37.

(46) in Social Judgment: Correcting for Context-Induced Contrast, Journal of Experimental Social Psychology, 29(March), 137-165. Russell H. Fazio, M. C. P., and Paul M. Herr. (1983). Toward a process model of the attitude-behavior relation: accessing one's attitude upon mere observation of the attitude object. Journal of Personality and Social Psychology, 44, 723-725. Shimp, Terence A., Eva M. Hyatt, and David J. Snyder (1991a), A Critical Appraisal of Demand Artifacts in Consumer Research, Journal of Consumer Research, 18 (December), 273-283 Stapel, D. A., Koomen, W., & Velthuijsen, A. S. (1998). Assimilation or contrast?: Comparison relevance, distinctness, and the impact of accessible information on consumer judgments. Journal of Consumer Psychology, 7(1), 1-24. Yi, Y. (1990). The effects of contextual priming in print advertisements. Journal of Consumer Research, 17, 215-222. Yi, Y. (1993). Contextual priming effects in print advertisements: The moderating role of prior knowledge. Journal of Advertising, 22, 1-10.. 38.

(47) Appendix (I): Pretest 親愛的朋友，您好：我們正在調查 20~30 歲的年輕人對於雜誌或平面廣告的閱讀與認知能力。在每一份問卷中我們會隨機抽取出一些來自市面上雜誌或廣告的內容，因此每一位受測者所看到的廣告或內容會有所不同。在問卷內的廣告是隨機擷取市面上主流的雜誌，其中可能會包含一些你我所熟知的品牌，也可能包含一些剛上市的新品牌。因此當您接觸到您不熟悉的廣告時，也希望您能仔細閱讀廣告並且耐心回答問題。您所填寫的資料純粹只提供本研究之用，請安心作答，並依據個人感受及想法作答。在作答過程中無法回到不同部分修改答案，不可跳題，因此請認真填答。如有任何疑問，請立即向施測人員詢問。您的熱心參與將有助於本研究的順利完成，在此先向您致上衷心的感謝！認真且填答完整的受測者將有機會抽中$1000 的 7-11 禮券。國立台灣師範大學管理研究所指導教授:蕭中強博士研究生:許益盛. 第一部分以下將進行一些測驗，目的是測驗大家的觀察力及敏感度，煩請仔細閱讀下面說明內容，再翻頁回答測驗。說明：根據認知學習理論，人們對於訊息的處理，無論是針對特定的東西，或者是模糊的概念，都會在心中形成一個盡可能準確的印象，並且嘗試用量化的方式來描述該印象所表達的程度，或其所能解釋的範圍，以下我們將舉例說明此種印象形成的過程。 39.

(48) 【範例】物品的軟硬程度 (1) 最能代表的值若請您由下列量表中圈選您認為最能代表「棉被」的「軟硬程度」的分數，其中 0 表示「最軟」，10 表示「最硬」。您可能會想說，棉被普遍來說較鬆軟，但又不是很極端的軟，若放在 0~10 的尺度中，「2.5」可能是最能代表棉被的柔軟，則請在下方的量表中，將數字「2.5」圈選出來。. (2) 適當的範圍但若將問題改成：請您由下列量表中選出您認為可以表現「棉被」的「軟硬程度」的適當範圍，並且找出此範圍的「最高值」與「最低值」。您可能尌會回想印象中最軟的棉被，像是蠶絲被或羽絨被，大約是在「1.5」的柔軟度；另一方面也回想著最硬的棉被，可能是放在櫥櫃中好幾年的舊棉被，大約是「5」的硬度，所以覺得棉被能表現的「軟硬程度」的範圍，可能落於最低值「1.5」和最高值「5」之間，則請在下方的量表中，將數字「1.5」和「5」圈選出來。. (3) 寫出一至三個具體的物品問題也可能由另一個方式問，例如：請您試著列出「軟硬程度」介於「9.5~10」之間的一至三個物品(最少一個，最多三個)。則您可能會試著去想一些很硬的東西（因為此範圍靠近最硬的極端值），且此物品類別內的每一項目硬度都很接近，最硬與最不硬時差異不大（因為此範圍很窄），再從您的記憶中去搜尋，發現「鋼筋水泥、鑽石」恰好符合以上的條件，因此決定以鋼筋水泥與鑽石作為「軟硬程度」落在「9.5~10」之間的代表，則請在量表下方的答案格中填上「鋼筋水泥、鑽石」。. (ANS) _______鋼筋水泥________ (ANS) _______鑽石____________ (ANS) _______________________. 在明白上述作答說明後，請繼續往下一頁作答，謝謝！ 40.

(49) §Exercise 為了讓您更了解測驗進行的方式，以下將會呈現練習題。上頁的範例已經說明了問題的三種形式和填答步驟，請試著以相同的方法完成練習。以下將詢問您對於某些事物的知覺，請務必用心回答，謝謝! Exercise 01【動物的行進速度】（以下量表，0 表示「行進速度最慢」，10 表示「行進速度最快」） (1) 最能代表的值請由下列量表中圈選您認為最能展現「烏龜」的「行進速度」的分數（請圈選出「一個」適當的數字）. (2) 適當的範圍請由下列量表中選出您認為「烏龜」可能展現「行進速度」的適當範圍，並圈選出代表此範圍的「最高值」與「最低值」（請圈選出「兩個」適當的數字）. (3) 寫出一至三個具體的形容詞或物品請試著列出「行進速度」介於「9~10」之間的一些動物。請試著想像一些動作極快的動物（因為此範圍靠近最快的極端值），而且各個品種的行進速度差異不大，最快的也不會比最慢的快太多（因為此範圍很窄）。（請將答案填在量表下方的答案格中）. (ANS)___________________ (ANS)___________________ (ANS)___________________. 41.

(50) Exercise 02【回答問卷的認真程度】(以下量表，0 表示「最不認真」，10 表示「最認真」。) (1) 最能代表的值請由下列量表中圈選您認為「目前」最能展現自己填寫問卷的「認真程度」的分數（請圈選出「一個」適當的數字）。. (2) 適當的範圍請由下列量表中選出您認為「目前」可能展現自己填寫問卷的「認真程度」的適當範圍，並圈選出代表此範圍的「最高值」與「最低值」。（請圈選出「兩個」適當的數字）. ~第一部分到此結束，請繼續前往第二部分~. 第二部分此部份將觀察您對品牌的觀察力與認知能力，請閱讀以下四則廣告。. 說明: 以下您將會看到市面上六則隨機抽取的平面廣告，請您仔細閱讀完每則廣告，在閱讀完畢後，將會詢問您一些相關問題，請您圈選心中適合的答案。在作答過程中無法回到不同部分修改答案，因此請認真填答。 42.

(51) 43.

(52) 接下來，請您針對廣告中品牌「belíssimo」的看法，依照您的直覺填答。在回答問題時，千萬不要往前翻閱或是修改答案。一、【價格】（以下量表，0 表示「最便宜」，10 表示「最貴」） (1) 最能代表的值請由下列量表中圈選您認為「belíssimo」最能表現「價格」的分數（請圈選出「一個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 (2) 適當的範圍請由下列量表中選出您認為「belíssimo」可能表現「價格」的適當範圍，並圈選出代表此範圍的「最高值」與「最低值」（請圈選出「兩個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。二、【尊貴程度】（以下量表，0 表示「最不尊貴」，10 表示「最尊貴」 (1) 最能代表的值請由下列量表中圈選您認為「belíssimo」最能呈現「尊貴程度」的分數（請圈選出「一個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 (2) 適當的範圍請由下列量表中選出您認為「belíssimo」可能呈現「尊貴程度」的適當範圍，並圈選出代表此範圍的「最高值」與「最低值」（請圈選出「兩個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 44.

(53) 45.

(54) 接下來，請您針對廣告中品牌「Timberland」的看法，依照您的直覺填答。在回答問題時，千萬不要往前翻閱或是修改答案。一、【價格】（以下量表，0 表示「最便宜」，10 表示「最貴」） (1) 最能代表的值請由下列量表中圈選您認為「Timberland」最能表現「價格」的分數（請圈選出「一個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 (2) 適當的範圍請由下列量表中選出您認為「Timberland」可能表現「價格」的適當範圍，並圈選出代表此範圍的「最高值」與「最低值」（請圈選出「兩個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。二、【尊貴程度】（以下量表，0 表示「最不尊貴」，10 表示「最尊貴」 (1) 最能代表的值請由下列量表中圈選您認為「Timberland」最能呈現「尊貴程度」的分數（請圈選出「一個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 (2) 適當的範圍請由下列量表中選出您認為「Timberland」可能呈現「尊貴程度」的適當範圍，並圈選出代表此範圍的「最高值」與「最低值」（請圈選出「兩個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 46.

(55) 47.

(56) 接下來，請您針對廣告中品牌「OPPO」的看法，依照您的直覺填答。在回答問題時，千萬不要往前翻閱或是修改答案。一、【時尚程度】（以下量表，0 表示「最低」，10 表示「最高」） (1) 最能代表的值請由下列量表中圈選您認為「OPPO」最能表現「時尚程度」的分數（請圈選出「一個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 (2) 適當的範圍請由下列量表中選出您認為「OPPO」可能表現「時尚程度」的適當範圍，並圈選出代表此範圍的「最高值」與「最低值」（請圈選出「兩個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。二、【價格程度】（以下量表，0 表示「最便宜」，10 表示「最貴」 (1) 最能代表的值請由下列量表中圈選您認為「OPPO」最能呈現「價格程度」的分數（請圈選出「一個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 (2) 適當的範圍請由下列量表中選出您認為「OPPO」可能呈現「價格程度」的適當範圍，並圈選出代表此範圍的「最高值」與「最低值」（請圈選出「兩個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 48.

(57) 49.

(58) 接下來，請您針對廣告中品牌「GU」的看法，依照您的直覺填答。在回答問題時，千萬不要往前翻閱或是修改答案。一、【價格】（以下量表，0 表示「最便宜」，10 表示「最貴」） (1) 最能代表的值請由下列量表中圈選您認為「GU」最能表現「價格」的分數（請圈選出「一個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 (2) 適當的範圍請由下列量表中選出您認為「GU」可能表現「價格」的適當範圍，並圈選出代表此範圍的「最高值」與「最低值」（請圈選出「兩個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。二、【尊貴程度】（以下量表，0 表示「最不尊貴」，10 表示「最尊貴」 (1) 最能代表的值請由下列量表中圈選您認為「GU」最能呈現「尊貴程度」的分數（請圈選出「一個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 (2) 適當的範圍請由下列量表中選出您認為「GU」可能呈現「尊貴程度」的適當範圍，並圈選出代表此範圍的「最高值」與「最低值」（請圈選出「兩個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 50.

(59) 51.

(60) 接下來，請您針對廣告中品牌「摩斯漢堡」的看法，依照您的直覺填答。在回答問題時，千萬不要往前翻閱或是修改答案。一、【美味程度】（以下量表，0 表示「最不美味」，10 表示「最美味」） (1) 最能代表的值請由下列量表中圈選您認為「摩斯漢堡」最能表現「美味程度」的分數（請圈選出「一個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 (2) 適當的範圍請由下列量表中選出您認為「摩斯漢堡」可能表現「美味程度」的適當範圍，並圈選出代表此範圍的「最高值」與「最低值」(請圈選出「兩個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。二、【健康程度】（以下量表，0 表示「最不健康」，10 表示「最健康」 (1) 最能代表的值請由下列量表中圈選您認為「摩斯漢堡」最能呈現「健康程度」的分數（請圈選出「一個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 (2) 適當的範圍請由下列量表中選出您認為「摩斯漢堡」可能呈現「健康程度」的適當範圍，並圈選出代表此範圍的「最高值」與「最低值」（請圈選出「兩個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 52.

(61) 53.

(62) 接下來，請您針對廣告中品牌「COACH」的看法，依照您的直覺填答。在回答問題時，千萬不要往前翻閱或是修改答案。一、【價格】（以下量表，0 表示「最便宜」，10 表示「最貴」） (1) 最能代表的值請由下列量表中圈選您認為「COACH」最能表現「價格」的分數（請圈選出「一個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 (2) 適當的範圍請由下列量表中選出您認為「COACH」可能表現「價格」的適當範圍，並圈選出代表此範圍的「最高值」與「最低值」（請圈選出「兩個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。二、【尊貴程度】（以下量表，0 表示「最不尊貴」，10 表示「最尊貴」 (1) 最能代表的值請由下列量表中圈選您認為「COACH」最能呈現「尊貴程度」的分數（請圈選出「一個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 (2) 適當的範圍請由下列量表中選出您認為「COACH」可能呈現「尊貴程度」的適當範圍，並圈選出代表此範圍的「最高值」與「最低值」（請圈選出「兩個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。. 54.

(63) 所有實驗完成後，我們會進行抽獎活動，在所有認真且填答完整受測者中抽出 20 人可以獲得價值$1000 元的 7-11 禮卷，因此請您留下您的基本資料，以方便我們進行資料確認，感謝您。. 基本資料 1. 姓名： 2. 性別： □ 男 □ 女 3. 年齡： □ 18-25 □ 26-35 □ 36-45 □ 46-55 □ 56-65 4. 職業/學校系級： 5. E-mail: 您認為這份問卷的目的是什麼? ___________________________________________________________ ___________________________________________________________ ___________________________________________________________ ___________________________________________________________ ___________________________________________________________. 55.

(64) Appendix (II): Main Experiment Stage 1 親愛的朋友，您好：我們正在調查 20~30 歲的年輕人對於雜誌或平面廣告的閱讀與認知能力。在每一份問卷中我們會隨機抽取出一些來自市面上雜誌或廣告的內容，因此每一位受測者所看到的廣告或內容會有所不同。在問卷內的廣告是隨機擷取市面上主流的雜誌，其中可能會包含一些你我所熟知的品牌，也可能包含一些剛上市的新品牌。因此當您接觸到您不熟悉的廣告時，也希望您能仔細閱讀廣告並且耐心回答問題。您所填寫的資料純粹只提供本研究之用，請安心作答，並依據個人感受及想法作答。在作答過程中無法回到不同部分修改答案，不可跳題，因此請認真填答。如有任何疑問，請立即向施測人員詢問。您的熱心參與將有助於本研究的順利完成，在此先向您致上衷心的感謝！認真且填答完整的受測者將有機會抽中$1000 的 7-11 禮券。國立台灣師範大學管理研究所指導教授:蕭中強博士研究生:許益盛. 第一部分以下將進行一些測驗，目的是測驗大家的觀察力及敏感度，煩請仔細閱讀下面說明內容，再翻頁回答測驗。說明：根據認知學習理論，人們對於訊息的處理，無論是針對特定的東西，或者是模糊的概念，都會在心中形成一個盡可能準確的印象，並且嘗試用量化的方式來描述該印象所表達的程度，或其所能解釋的範圍，以下我們將舉例說明此種印象形成的過程。 56.

(65) 【範例】物品的軟硬程度 (1) 最能代表的值若請您由下列量表中圈選您認為最能代表「棉被」的「軟硬程度」的分數，其中 0 表示「最軟」，10 表示「最硬」。您可能會想說，棉被普遍來說較鬆軟，但又不是很極端的軟，若放在 0~10 的尺度中，「2.5」可能是最能代表棉被的柔軟，則請在下方的量表中，將數字「2.5」圈選出來。. (2) 適當的範圍但若將問題改成：請您由下列量表中選出您認為可以表現「棉被」的「軟硬程度」的適當範圍，並且找出此範圍的「最高值」與「最低值」。您可能尌會回想印象中最軟的棉被，像是蠶絲被或羽絨被，大約是在「1.5」的柔軟度；另一方面也回想著最硬的棉被，可能是放在櫥櫃中好幾年的舊棉被，大約是「5」的硬度，所以覺得棉被能表現的「軟硬程度」的範圍，可能落於最低值「1.5」和最高值「5」之間，則請在下方的量表中，將數字「1.5」和「5」圈選出來。. (3) 寫出一至三個具體的物品問題也可能由另一個方式問，例如：請您試著列出「軟硬程度」介於「9.5~10」之間的一至三個物品(最少一個，最多三個)。則您可能會試著去想一些很硬的東西（因為此範圍靠近最硬的極端值），且此物品類別內的每一項目硬度都很接近，最硬與最不硬時差異不大（因為此範圍很窄），再從您的記憶中去搜尋，發現「鋼筋水泥、鑽石」恰好符合以上的條件，因此決定以鋼筋水泥與鑽石作為「軟硬程度」落在「9.5~10」之間的代表，則請在量表下方的答案格中填上「鋼筋水泥、鑽石」。. (ANS) _______鋼筋水泥________ (ANS) _______鑽石____________ (ANS) _______________________. 在明白上述作答說明後，請繼續往下一頁作答，謝謝！ 57.

(66) §Exercise 為了讓您更了解測驗進行的方式，以下將會呈現練習題。上頁的範例已經說明了問題的三種形式和填答步驟，請試著以相同的方法完成練習。以下將詢問您對於某些事物的知覺，請務必用心回答，謝謝! Exercise 01【動物的行進速度】（以下量表，0 表示「行進速度最慢」，10 表示「行進速度最快」） (1) 最能代表的值請由下列量表中圈選您認為最能展現「烏龜」的「行進速度」的分數（請圈選出「一個」適當的數字）. (2) 適當的範圍請由下列量表中選出您認為「烏龜」可能展現「行進速度」的適當範圍，並圈選出代表此範圍的「最高值」與「最低值」（請圈選出「兩個」適當的數字）. (3) 寫出一至三個具體的形容詞或物品請試著列出「行進速度」介於「9~10」之間的一些動物。請試著想像一些動作極快的動物（因為此範圍靠近最快的極端值），而且各個品種的行進速度差異不大，最快的也不會比最慢的快太多（因為此範圍很窄）。（請將答案填在量表下方的答案格中）. (ANS)___________________ (ANS)___________________ (ANS)___________________. 58.

(67) Exercise 02【回答問卷的認真程度】(以下量表，0 表示「最不認真」，10 表示「最認真」。) (1) 最能代表的值請由下列量表中圈選您認為「目前」最能展現自己填寫問卷的「認真程度」的分數（請圈選出「一個」適當的數字）。. (2) 適當的範圍請由下列量表中選出您認為「目前」可能展現自己填寫問卷的「認真程度」的適當範圍，並圈選出代表此範圍的「最高值」與「最低值」。（請圈選出「兩個」適當的數字）. ~第一部分到此結束，請繼續前往第二部分~. 第二部分此部份將觀察您對品牌的觀察力與認知能力，請閱讀以下四則廣告。. 說明: 以下您將會看到市面上六則隨機抽取的平面廣告，請您仔細閱讀完每則廣告，在閱讀完畢後，將會詢問您一些相關問題，請您圈選心中適合的答案。在作答過程中無法回到不同部分修改答案，因此請認真填答。 59.

(68) 60.

(69) 接下來，請您針對廣告中品牌「belíssimo」的看法，依照您的直覺填答。在回答問題時，千萬不要往前翻閱或是修改答案。一、【價格】（以下量表，0 表示「最便宜」，10 表示「最貴」） (1) 最能代表的值請由下列量表中圈選您認為「belíssimo」最能表現「價格」的分數（請圈選出「一個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 (2) 適當的範圍請由下列量表中選出您認為「belíssimo」可能表現「價格」的適當範圍，並圈選出代表此範圍的「最高值」與「最低值」（請圈選出「兩個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。二、【尊貴程度】（以下量表，0 表示「最不尊貴」，10 表示「最尊貴」 (1) 最能代表的值請由下列量表中圈選您認為「belíssimo」最能呈現「尊貴程度」的分數（請圈選出「一個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 (2) 適當的範圍請由下列量表中選出您認為「belíssimo」可能呈現「尊貴程度」的適當範圍，並圈選出代表此範圍的「最高值」與「最低值」（請圈選出「兩個」適當的數字）. 您為什麼會這樣選擇:__________________________________________________。 61.

(70) 62.