Overview
In Experiment 2A and 2B, we set to test the context effect of range overlap of Prime 1 and 2 with target separately according to our Hypotheses 2A and 2B, on occurrence of
assimilation. This study is conceptually identical to Hsiao’s (2002) Reciprocity Hypothesis where both contexts, tested separately with the target, will move simultaneously with the target, towards each other, in the situation where an overlap exists between them. The current case however, shall lay ground for our final hypotheses hereinafter, to test the post-context effect movements of the target stimuli and contexts when they were presented at once.
Hypotheses Recap
Each of the two reported experiments tested two key hypotheses on the occurrences of assimilation effect where a dimensional range overlap exist between the context and target, one on each valence of the target.
Hypothesis 2A: 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 6.1 Hypothesis 2A
Hypothesis 2B: 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 both the prime and target will assimilate towards each other.
Figure 6.2 Hypothesis 2B
Participants and Design
The four-section Experiment 2A & 2B was carried out by way of a within-participants design, conducted on 50 young adults, with 25 participants each for each condition. These participants, aged between 20 and 30 years old, are mostly undergraduates and postgraduate students from National Taiwan Normal University, either recruited in the school campus or via online postings. Participants signed up voluntarily for one of the 32 sessions held to conduct all three main experiments, each lasting about ten to fifteen minutes, held over four days of the study, in exchange for a lucky draw chance to win NTD500 worth of Uni-President vouchers. They were subsequently randomly assigned to either one of the experimental conditions in this study.
Each of the reported experiment 2A and 2B also consist of two stages; the first stage was designed to establish participants’ context-free representative score, as well as both lower and higher bound values of items’ ranges, while the second stage would be their post-priming representative values and item boundaries. In either of the experimental conditions,
the VBA programme in which the questionnaire was written on was manipulated to generate Prime 1’s range to be exactly 0.5 on the negative valence of the target range (Experiment 2A), while Prime 2 would appear to have an overlap range of exactly 1 on the positive valence of participants’ target range ratings (Experiment 2B).
Procedure
The introductory statements and practice exercise given in this study was identical to that of Experiment 1, after which participants working on both questionnaires were shown several advertisements (i.e. Motorola, McDonalds, Mercedes-Benz) and asked to give representative values, as well as item boundaries for each brand advertisement shown. Our hypothetical target apparel brand advertisement, Belìssimo, which was determined to be of a range between 3.72 to 5.88 in the pretest, was used here in this main experiment was included and introduced before all other advertisements, and all values given would serve as context-free target ratings. This was followed by a psychology test.
In a similar section two of the questionnaire, participants were told they would be tested on their imagination abilities, and asked to input one apparel brand each in Experiment 2A and Experiment 2B, in which they thought would be associated with a specific prestige range. This second section of the questionnaire was that of our context-free situation; Prime 1 and 2 were manipulated to overlap with target 0.5 and 1 respectively in Experiment 2A and 2B initially. The brands suggested would later be used as our Prime 1 and 2 respectively in Experiment 2A and 2B. To reduce the likelihood that participants may guess the purpose of the study, brands of two other irrelevant items (i.e. cars, fast-food chains) and dimensions (i.e.
safety level, health level) were also asked in both questionnaires. A Sudoku game meant to compel the use of participants’ cognitive abilities, and possibly disrupt their concentration
and accessibility of previously primed item, were inserted for the participants to complete, before they proceeded to the next section.
In the final section of both questionnaires, participants were similarly told that the computer has chosen the clothing industry for them to answer subsequent questions on. They were also reminded of the respective Prime 1 and 2 brands they specified in the section before, and asked to re-rate the representative value and interpretative range (lower and upper bounds) on the “prestige level” dimension of Belìssimo (the target) and the corresponding prime of the condition.
Results
Experiment 2A
Manipulation overlap. Results from our within-participants t-tests (Table 6.1) showed that when there was an overlap in the perceived prestige ranges of Prime 1 on the negative
valence of the target, assimilation effect arose and both the context and target shifted towards each other concurrently. Shifts in the respective context and target range ratings resulted in a substantially greater overlap (M = 1.14) than the 0.5 manipulated overlap initially; t(24) = 4.11, p < 0.001, thereby supporting Hypothesis 2A. (Please refer to Figure 6.3)
Prime 1 and target evaluation. The initial lower bound of M = 2.68 in relation to the post-influenced lower bound M = 3.50; t(24) = -3.64, p < 0.01) and initial upper bounds of Prime 1 M = 4.68 as compared to the upper bound of M = 5.50 after; t(24) = -3.31, p < 0.01, showed that Prime 1 has moved considerably towards the target as anticipated. The same cannot be said for the target, whose results are mixed with a relatively shortened range in Stage 2; its post-context effect lower bound contrasted away from Prime 1 (M = 4.40) compared with its initial lower bound (M = 4.18); t(24) = -0.68, p = 0.51, although its upper
bound did significantly shift in the direction of Prime 1 (M = 6.16) from its initial upper bound of 6.82; t(24) = 2.50, p < 0.05.
Paired sample t-tests were also conducted to compare the representative values (Table 6.1); Prime 1 generated higher prestige ratings (M = 4.42) than its initial rating (M = 3.68);
t(24) = -2.70, p = 0.01, while the target also produced a slightly lower post-context prestige ratings as expected (M = 5.00) than its initial context-free ratings (M = 5.18); t(24) = 0.78, p
= 0.45. The distance between the post-context effect representative values of Prime 1 and target (M = 0.58) was found to be shorter than in context-free Stage 1 (M = 1.50); t(24) = 3.07, p < 0.01, indicating an assimilation effect and further confirming our hypothesis.
Figure 6.3 Diagram of movements of Prime 1 and target in Experiment 2, indicating respective representative values, lower and upper bound ratings.
Table 6.1
Means for Prime 1 and the target in condition 1 of Experiment 2
Prime 1 Representative Value Lower Bound Upper Bound
Stage 1 Mean 3.68 2.68 4.68
Stage 2 Mean 4.42 3.50 5.50
Target Representative Value Lower Bound Upper Bound
Stage 1 Mean 5.18 4.18 6.85
Stage 2 Mean 5.00 4.40 6.21
Prime 1 and
Target Overlap Range Distance between Representative Values
Stage 1 Mean 0.50 1.50
Stage 2 Mean 1.14 0.58
Experiment 2B
Manipulation check. Similarly, results from our within-participants t-tests (Table 6.2) showed that when there was an overlap in the perceived prestige ranges of Prime 2 on the positive valence of the target, assimilation effect caused both the context and target to simultaneously shift towards each other.
Prime 2 and target evaluation. However, this assimilation between Prime 2 and target is not statistically proven; the context range of Prime 2 was stretched, with its lower (M = 4.26) assimilating with the target range but upper bound (M = 6.56) contrasting from the target.
These lower and upper bounds of Prime 2 in Stage 2 however did not significantly differ from that of context-free Stage 1, where the lower and upper bounds have means of 4.40 and 6.40 respectively; t(24) = 0.46, p = 0.65 and t(24) = -0.65, p = 0.53 for lower and upper
bound correspondingly. The target range on the other hand, provided more comforting results, with both the lower and upper bound of target shifting towards Prime 2. There was a
significant difference in the lower bounds of the target before (M = 2.98) and after context effect (M = 3.40); t(24) = -2.40, p < 0.05). Though the upper bound of the target also
assimilated with Prime 2, the boundary ratings were not significantly different in Stage 2 (M
= 5.70) as in Stage 1 (M = 5.40); t(24) = -1.29, p = 0.21. As a result, though a greater overlap (M = 1.44) range was produced after the mutual effect of Prime 2 and target, it was not confirmed statistically (t(24) = 1.07, p = 0.3), thereby not supportive of Hypothesis 2B.
(Please refer to Figure 6.4)
In subsequent paired sample t-tests conducted (Table 6.2), Prime 2 generated higher prestige representative value ratings (M = 5.56) in Stage 2 than it was in Stage 1 (M = 5.40);
t(24) = -0.16, p = 0.50). On the contrary, the perceived prestige representative value
generated from a target shift after context effect assimilated with Prime 2 as we expected (M
= 4.34), compared to the pre-context effect representative value (M = 3.98); t(24) = -1.76, p <
0.1. The distance between the pre- (M = 1.42) and post-context effect representative values of Prime 2 and target (M = 1.22) indicated an assimilation between the two, though the shift was not statistically proven; t(24) = 0.72, p = 0.48). Hypothesis 2B is therefore not confirmed.
Table 6.2
Means for Prime 2 and the target in condition 2 of Experiment 2
Prime 2 Representative Value Lower Bound Upper Bound
Stage 1 Mean 5.40 4.40 6.40
Stage 2 Mean 5.56 4.26 6.56
Target Representative Value Lower Bound Upper Bound
Stage 1 Mean 3.98 2.98 5.40
Stage 2 Mean 4.34 3.40 5.70
Prime 2 and
Target Overlap Range Distance between Representative Values
Stage 1 Mean 1.00 1.42
Stage 2 Mean 1.44 1.22
Figure 6.4 Diagram of movements of Prime 2 and target in Experiment 2, indicating respective representative values, lower and upper bound ratings.
Further Analyses (Experiment 2B)
Manipulation check. Similar to Experiment 1, we sieved out a total of 11 data sets, with which had a post-context effect overlap range satisfying our expectations (> 0.5) for further analyses. A general comparison of Prime 1 and the target, before and after context effect, reflected that both ranges have indeed assimilated towards each other, with all perceived prestige representative values as well as upper and lower bound ranges moving in directions as we had expected. Expectedly, results also revealed a significant difference in the degree of overlap (M = 2.91) from the initial manipulated range of 1.00; t(10) = 3.04, p = 0.01, thereby supporting our hypothesis in this case. (Please refer to Table 6.3 and Figure 6.5)
Prime 2 and target evaluation. Upon scrutiny, only shifts in the lower bound of Prime 2 (M
= 3.00) and the upper bound of the target range (M = 6.05) were significant; t(10) = 3.34, p <
0.01 and t(10) = -1.79, p = 0.1 respectively. In addition, there was a significant difference in the distance between the representative values of Prime 2 and the target, with the paired sample t-test presenting a Stage 1 mean of 1.09 compared to a Stage 2 mean of 0.27; t(10) = 1.85, p < 0.1, hence supporting Hypothesis 2B.
Table 6.3
Means for Prime 2 and the target in condition 2 of Experiment 1 on selected data sets Prime 2 Representative Value Lower Bound Upper Bound
Stage 1 Mean 5.32 4.32 6.32
Stage 2 Mean 4.73 3.00 6.05
Target Representative Value Lower Bound Upper Bound
Stage 1 Mean 4.23 2.91 5.32
Stage 2 Mean 4.46 3.14 6.05
Prime 2 and
Target Overlap Range Distance between Representative Values
Stage 1 Mean 1.00 1.09
Stage 2 Mean 2.91 0.27
Figure 6.5 Diagram of movements of Prime 2 and target on selected data sets in Experiment 2, indicating respective representative values, lower and upper bound ratings.