4-1. Experiment 1
4-1-1. Overview. Experiment 1 was conducted to observe how non-overlap between two primes, so-called positive and negative primes, leads to assimilation or contrast effects after priming. Chien et al’s Dimensional Range Overlap (2010) proposed that an overlap and non-overlap range between the context and target ranges will produce an assimilation or contrast effect. According to Hisao’s (2002) Reciprocity Hypothesis, both contexts will have a simultaneous shift toward or away from each other. From these studies, we developed two stages for this experiment. We compared the width of non-overlap range which we manipulated by 0.5 between Prime 1 and Prime 2 in Stage 1 and its range after the contextual effect in Stage 2.
4-1-2. Hypothesis Recap.
H1: When prime 1 and 2 have different valences and there is no overlap between the two primes initially, then prime 1 and prime 2 will contrast away from each other.
Figure 4-1 Hypothesis 1
4-1-3. Participants and Design. A total number of 20 participants aged between 20 to 30 years, are mostly undergraduate and graduated-students from the Business department and MBA at National Taiwan Normal University participated in Experiment 1. These students were recruited on campus during their break time between classes.
Two stages of Experiment 1 was held over two to three weeks. Responding to each questionnaire took less than 15 minutes for Stage 1 and less than 5 minutes for Stage 2, in exchange for a chance to win NTD 500 worth of 7-Eleven vouchers.
Two stages of Experiment 1 ware conducted by way of within-participants design.
Stage 1 is designed for providing items which comes to a participants’ mind within a given range. The ranges of 2.5-4.5 and 5-7 were fixed for two ranges, which can be Prime 1 and Prime 2 in order to manipulate a 0.5 non-overlap range between them. Before filling in the question of items, bunch of brands were randomly listed with the purpose of easier accessibility of items. In Stage 2, participants gave ratings to two contexts which they
chose in Stage 1. There are one to two weeks between Stage 1 and Stage 2 to observe whether both contexts are influenced mutually after context effect.
4-1-4. Procedure. A whole procedure was conducted with paper based questionnaires. It was a color printed booklet including twelve-pages for Stage 1 and four pages for Stage 2. Participants were told that they joined the study for the purpose of researching the cognitive responses of young adults for print ads in magazines. They were told that those ads in the questionnaires, were randomly selected from mainstream magazines including very popular brands or brands that are fairly new to the market. In the beginning part of the Stage 1 questionnaire, participants were asked to exercise a range on judgmental dimension with their observation sensitivity. The first two sample questions in this exercise compose two dimensions, “hardness of article” and “sweetness of food”, followed by two exercise questions on dimensions of “hardness of article” and
“the speed of animal”, as duplicated from Lu’s (2014) Context Effect under Dual Contexts: The Influence of Positive and Negative Primes to Product Judgement.
For evaluation of each dimension, participants were asked to give ratings of the representative value, and dimensional range. The representative value is one- point score rating that best represents items from a scale of 0 to 10 in intervals of 0.5. Figure 4.2, the example of the representative value for “hardness of article”, is shown before the main experiment in our questionnaire in order to help them understanding. The dimensional range is a suitable range rating with lowest and highest values in the same scale of 0 to 10. Figure 4.3 indicates an example of dimensional range for “Hardness of article”.
Besides these two patterns of rating questions, participants were given a fill-in-the- break type of question to list specific items. Participants were assigned to offer an item within the specific range as seen in Figure 4.3.
Figure 4.2 An example of representative value most suitable for “Hardness of article”
Figure 4.3 An example of dimensional range for “Hardness of article” explained in the questionnaire
Figure 4.4 An example of listing of an object, given specific range on the scale for the
“Hardness of article”
The next part of questionnaire Stage 1 asked about participants’ pre contextual effect item of apparel brands, which would be used as our Prime 1 and 2 in Stage 2. As mentioned earlier in this chapter, we inquired participants to write an apparel brand for Prime 1 as falling between 2.5-4.5 prestige level and Prime 2 as 5-7. Participants may also answer these contexts from a list which contained 34 apparel brands that we prepared.
After filling in the questions of apparel brand, filler questions of automobile and mobile phone brands are also arranged to reduce the possibility of participants guessing the aim of our questionnaires. In detail, we asked about the automobile safety levels between 5-7 and 6-8 on the dimensional ranges and level of phone camera quality between ranges 4-6 and 8-10 ranges.
Figure 4.5 A list of apparel brands’ logos. Participants check mark the ones that they recognize before answering the context free question of prestige level for apparel brands.
Stage 2 was conducted after one to two weeks to reduce participants’ accessibility of previously given ratings and item they chose. Participants were told that we randomly selected product advertisement to read before starting the Stage 2. We asked them to mark
a representative value and dimensional range rates for both Prime 1 and 2.
4-1-5. Results.
Manipulation Check. The manipulation of non-overlap between Prime 1 and 2
was unsuccessful. That is, even though the non-overlap range between Prime 1 and 2 was manipulated to be 0.5, participants reported overlap rather than non-overlap between Prime 1 and 2 in the manipulation checks Moverlap 0.275, t(19) = 0.18, p=
0.094.
Prime 1 and 2 evaluations. One sample t-test is applied to analyze the lower and
upper bounds of two primes. The lower bound of Prime 1 (M= 3.38) showed a large difference from the manipulated lower bound of M= 2.50, t(19)= 2.98, p=0.009. The upper bound of Prime 1 also generated a higher rating (M= 5.63) from the manipulated upper bound (M=4.50, t(19) =3.957, p= 0.001). On the other hand, the lower bound (M=
5.35) and upper bound of Prime 2 (M= 7.60) did not differ significantly from its initial value of 5 and 7 individually; (t(19) = 0.86, p= 0.401) and (t(19)= 2.22, p= 0.039).
The representative values were compared between pre and post-primed conditions.
The negative context, Prime 1, created higher post-context prestige rates (M= 4.38; t(19)=
3.523, p=.002) than its average of manipulated upper bound and lower bound (M= 3.50).
Therefore, the post context result of P1 shifted upward. Similarly, the positive context, Prime 2, also produced higher prestige rates (M= 6.45; t(19) = 1.616, p=0.122) than its average of manipulated upper bound and lower bound (M= 6.00). Prime 2 was moved upward as well.
This experiment found that representative values in both primes shifted toward higher ratings in stage 2 from the values in Stage 1. Prime 1 and Prime 2 did not shift away from each other simultaneously, and as a result, our Hypothesis 1 was not confirmed.
Table 4.1 Means for the two primes and the target of Experiment 1
Prime 1 and 2 Non-overlap Range Distance between Representative Values
Stage 1 Mean 0.50 2.50
Figure 4-6. Diagram of the shift of Prime 1 and 2 in Experiment 1, showing representative values, and lower and upper bound values.
6.45 Prime 1’s representative value 4.1
Prime 2’s representative value
4.50
4-2. Experiment 2
4-2-1. Overview. Experiment 2 for Hypothesis 2A and 2B were examining the context effect of overlap of Prime 1 and Prime 2 with the target respectively. We expected an occurrence of the assimilation effect for these experiments. This study ‘s concept of contexts, separates tests of the target with Prime 1 and Prime 2, condition where an overlap between them from Hsiao(2002)’s Reciprocity Hypothesis.
4-2-2. Hypothesis Recap.
H2A: For the condition in which the prime on the negative valence of the target is wide enough to create an overlap with the target initially, then both the prime and target will assimilate towards each other.
Figure 4.7 Hypothesis 2A
H2B: For the condition in which the prime on the positive valence of the target is wide enough to create an overlap with the target initially, then the prime and target will assimilate toward each other.
Figure 4.8 Hypothesis 2B
4-2-3. Participants and Design. 42 young adults, with 21 participants each for experiment 2A and 2B take part in our questionnaires. Most of our participants were undergrads and graduate students from the Business department and MBA at National Taiwan Normal University. These students were recruited on the school campus during their break time between classes. Three stages of Experiment 1 were held over two, three, and four weeks. Responding to the questionnaire took less than 15 minutes for Stage 1 and less than 10 minutes for Stage 2 and 3, and all participants were entered in a raffle for a chance to win NTD 500 worth of 7-11 vouchers.
Three stages of Experiment 2 were conducted by way of participants design. Stage 1 was designed to give a score of representative values and lower and upper bounds for the target in order to test the target’s dimensional range pre context effects.
Stage 2 was designed to identify participant’s prime 1 and 2 for Experiments 2A and 2B separately. Participants were asked to provide items which come to their mind within a given range. The range was adjusted for prime 1, overlapping 0.5 with negative valence of the target. Prime 2 in Stage 2 of the questionnaire was overlapped with the positive
valence of the target, the 1 overlap range. Before filling in the question about items, a variety of brands were randomly listed with the purpose of easier accessibility of items.
In Stage 3, participants give ratings to the target with Prime 1 and 2 separately, which they chose in Stage 1 and 2. There is approximately one week each between Stage 1, Stage 2 and Stage 3 to observe whether both contexts and the target are mutually affected after context effects.
4-2-4. Procedure. A whole procedure went through with paper based questionnaires as Experiment 1. It was a color printed booklet including fourteen pages for Stage 1, eight pages for Stage 2 and four pages for Stage 3. The introductory statement and exercises in this experiment were exactly the same as that of Experiment 1. In this Stage, participants were asked their rates to the representative value and lower and upper bounds for a prestige level of the target. The target’ advertisement was retrieved from the fictitious advertisement, Belissimo, which Lu (2014) produced for Context Effect under Dual Contexts: The Influence of Positive and Negative Primes to Product Judgement.
The target advertisement of apparel brand prestige level was designed within neutral ratings, with a lower mean of 3.72 and higher mean of 5.88. After the target advertisement, participants were asked to evaluate several advertisements (i.e. Motorola, McDonalds, and Mercedes-Benz), however, these are filler questions to ensure participants did not guess our experiment’s purpose.
In Stage 2 for our questionnaires, participants were requested to list one apparel brand each in Experiment 2A and 2B. The score we gave for Prime 1 and Prime 2 were manipulated to overlap with target 0.5 and 1 each in Experiments 2A and 2B. Participants may also answer these contexts from a list which contained 34 apparel brands we prepared.
Similar to Stage 1, irrelevant items such as automobile and cellphone brands were
included in this part of the questionnaires to reduce the possibility of guessing the purpose of our study.
Stage 3 was conducted after one to two weeks of the previous stages to forget given rates and apparel brands they filled in earlier. Participants were told that we randomly selected product advertisement to read before starting Stage 3. We asked them to mark a representative value and dimensional range rates for the target and Prime 1 and 2 respectively for Experiment 2A and 2B.
4-2-5. Results
4-2-5-1. Experiment 2A.
Manipulation Check. The manipulation of overlap between Prime and the target was successful. Overlap range between the Prime 1 and target resulted in greater overlap (M= 2.05) than the initial manipulated overlap, 0.5. (t(20) = 5.365, p<0.001) as we expected.
Prime 1 and target evaluation. Paired-sample t-tests were conducted to compare each a shift of pre & post-context effect lower bound and its upper bound of Prime 1. The manipulated lower bound (M=1.57) significantly shifted to the post context effect lower bound (M=3.07; t(20) = -7.46, p<0.001). Correspondingly, the manipulated upper bound (M= 3.57) did significantly move in the direction of post context effect upper bound (M=
5.19; t(20) = -7.082, p<0.001). Both lower and upper bounds shifted upwards. Therefore, Prime 1 evidently showed the shift to the target as we anticipated.
For the shift of target’s lower bound and upper bound, we also applied the paired-sample t-test. The post lower bound (M=3.143) didn’t shift from the manipulated lower bound of the target (M= 3.071; t(20)= -0.278, p = 0.78). The initial upper bound (M=
5.95) and post context effect upper bound (M= 5.93) indicated a similar rating as well
(t(20)= 0.085, p=0.93). The target stayed in the same place after post context effect although Prime 1 shifted toward the target. Hence, this experiment did not prove the simultaneous shift.
The representative value for Prime 1 and the target were also analyzed to compare the manipulated and post context ratings. The representative value of Prime 1 indicated a higher value (M= 4.18) than the manipulated value (M=2.58; t(20) =-7.01, p<0.001).
However, the target only produced a slightly lower prestige rating (M=4.35) than its manipulated value (M=4.45; t(20) = 0.298, p= 0.77). The representative value of the target moved downward, but it couldn’t statistically prove the shift. The representative target value of Prime 1 and the target didn’t show the simultaneous shift of the assimilation effect. As a result, hypothesis 2A was not supported.
Table 4.2 Means for Prime 1 and the target in condition 1 of Experiment 2
Prime 1 Representative
Target Overlap Range Distance between Representative Values
Stage 1 Mean 0.50 1.87
Stage 2 Mean 2.05 0.17
p <0.001 p <0.001
Figure 4-9. Diagram of shift of Prime 1 and the target in Experiment 2A, showing representative values, lower and upper bound values.
4-2-5-2. Experiment 2B . One participant was eliminated from 21 participants for data analysis due to missing data.
Manipulation check. Overlap range between the Prime 2 and target resulted in the width of overlap being slightly wider (M=1.45) than the manipulated overlap (t(19)= 1.23 , p= 0.234). However, this overlap was not statistically significant.
Prime 2 and target evaluation. The manipulated lower bound of Prime 2 (M= 3.75) slightly shifted to the post context effect lower bound (M= 3.88; t (19)= -3.98, p = 0.7) while the initial upper bound of Prime 2 (M=5.75) significantly moved to the post upper bound (M= 6.52, t (19) = -3.44, p= 0.03). However, Prime 2 moved upward, which we did not expect. The initial lower bound of the target (M=2.05) significantly moved to the
0
Prime 1’s representative value Target’s representative value
post context effect lower bound (M= 2.68, t(19) = -2.27, p= 0.04) while its shift of upper bound didn’t indicate the significant rating. (Pre-context effect upper bound of Prime 2:
M=4.75, its post upper bound: M=5.36; t= -2.03, p= 0.06). Thus, the assimilation between prime 2 and target were not confirmed.
Paired sample t-tests were also conducted to compare the representative values of Prime 2 and the target. Similarly, the representative value of Prime 2 contrasted away from its manipulated context. Its manipulated representative value was M= 4.75 shifted in the direction of post context representative value (M=5.35; t (19) = -2.77, p= 0.01). On the other hand, the manipulated representative value of the target (M = 3.15) significantly shifted toward to its post rating (M= 3.92; t (19) = - 2.68, p= 0.02). The representative value of the target also moved upward. The representative values also didn’t show the assimilation effect between Prime 2 and the target, therefore Hypothesis 2B was not supported.
Table 4.3 Means for Prime 2 and the target in condition of Experiment 2
Prime 2 Representative
Value Lower bound Upper Bound
Stage 1 Mean 4.75 3.75 5.75
Stage 2 Mean 5.35 3.86 6.52
p = 0.012 p = 0.695 p = 0.003
Target Representative
Value Lower bound Upper Bound
Stage 1 Mean 3.15 2.05 4.75
Stage 2 Mean 3.93 2.68 5.38
p = 0.015 p = 0.035 p = 0.056
Prime 2 and
Target Overlap Range Distance between Representative Values
Stage 1 Mean 1.0 1.60
Stage 2 Mean 1.45 1.425
p = 0.234 p = 0.0630
Figure 4-10. Diagram of shift of Prime 2 and the target in Experiment 2B, showing representative values, lower and upper bound values.
4-3. Experiment 3
4-3-1. Overview. Experiment 3 was conducted to examine the influence of mixed prime conditions, including positive and negative primes and the target. We expected to see the target shift toward the positive prime than the negative prime as well as the
Prime 2’s representative value Target’s representative value
occurrence of the contrast effect between Prime 1 and Prime 2
4-3-2. Hypothesis Recap.
H3: For the condition in which the two primes have different valences and each overlap with the target initially, but the positive prime overlaps with the target at a larger amount than the negative prime, and both primes do not overlap with each other, the target will be more likely to shift toward the positive prime than toward the negative prime.
Figure 4-11. Hypothesis 3
4-3-3. Participants and Design. A total of 20 participants aged between 20 to 30 years who are mostly undergrads and graduated students at National Taiwan Normal University joined Experiment 3. These students were recruited on campus during their break time between classes. Three stages of Experiment 3 were held over three to four weeks. Responding to the questionnaire took less than 15 minutes for Stage 1 and less than 10 minutes for Stage 2 and 3 to finish, and all participants were entered in a raffle
for a chance to win NTD 500 worth of 7-11 vouchers.
Similar to the previous experiment, Experiment 3 followed a participant-sensitive design, consisting of three stages, and paper-based questionnaires. The questionnaire design of Stage 1 is identical to Experiment 2. For Stage 2 of Experiment 3, the manipulated Prime 1 and 2 ranges were overlapped from 0.5 and 1 respectively based on the target that participants rated. Participants were asked to give scores to all targets and two contexts in Stage 3. It takes approximately one week each between Stage 1, Stage 2 and Stage 3 to observe whether each context and the target are mutually affected after any context effect.
4-3-4. Procedure. The whole procedure of Experiment 3 is similar to Experiment 2 except that the condition of testing is two primes with the target at once, unlike in Experiment 2, which examines two primes with the target separately.
4-3-5. Results. Two participants didn’t reach Stage 3 of our questionnaires, although there were 20 participants for experiment 3 originally.
Manipulation check.
Prime 1 and Prime 2. The manipulation of non-overlap between Prime 1 and 2 was unsuccessful. Although the overlap between 1 and 2 is more overlapped (M= 0.14) than the manipulated overlap range of M= 0.5, it was not statistically significant.
(t (17) = -0.80 , p= 0.435)
Prime 1 and target. The manipulation of overlap between Prime 1 and the target was successful. The overlap range between Prime 1 and the range was manipulated to be 0.5, as a result, participants reported more overlap (M = 1.42, t(17) = 2.95 p< 0.01).
Prime 2 and target. The manipulation of overlap between Prime 2 and the target failed because less overlap was reported (M=0.64; t(17) = -1.36 p = 0.19) rather than the manipulated overlap range M= 1.0
Evaluation.
Prime 1. One sample t-test was conducted to analyze the lower and upper bounds of two primes and the representative value. The post-context effect lower bound of Prime 1 showed a significantly higher rating (M= 2.97) from its manipulated lower bound (M=
2.08, t(17) = -2.899, p = 0.01). Correspondingly, the post-context effect upper bound of Prime 1 was also greater (M= 5.29) than the manipulated upper bound of Prime 1 (M=
4.08, t (17) = -2.91, p = 0.01). Both lower and upper bound of Prime 1 were moved upward. The representative value of Prime 1 was similarly shifted from the manipulated value M= 3.08 to the post context effect value M= 4.28, therefore, it moved upward significantly (t (17) = - 2.842 p = 0.011). As a result, Prime 1 assimilated to the target in opposition of our expectation.
Prime 2. One sample t-test was once again applied to examine Prime 2. The manipulated lower bound of Prime 2 (M= 4.94) shifted to the direction of its post lower bound (M=5.42, t(17) = -2.152, p = 0.046). The manipulated upper bound of Prime 2 (M=
6.94) also significantly moved to a higher rate for its post context effect upper bound (M=7.42, t (17) = -2.189, p = 0.043). The representative value showed the same result of an upward shift. The pre- context effect value (M= 5.94) moved to its post context effect value (M=6.61, t(17) = -2.675, p = 0.016). That is, Prime 2 contrasted away from the target.
Target. The shift of the target was also evaluated by a paired one sample t-test. The lower bound of the target was shifted slightly from the pre-contextual effect rate (M=