An Introduction to Contextual Priming Effect
The term “priming” was first invented by Segal and Cofer (1960) to refer to the effect of using a particular concept in one task on the probability that the same context would be used in a subsequent, unrelated context. It essentially refers to the processing of information with awareness or intent.
The priming methodology was, however, acquainted to experimental social
psychology by Higgins, Rholes, and Jones (1977), in a set of studies that involved priming personality concepts and tracking effects on subsequent impression formation, and in which also remains representative of the basic priming paradigm used today.
Oblivious to the priming effects they have been exposed to, consumers hardly make impartial and unbiased judgments in their daily product evaluations and purchases. When primed, accessibility of some category or construct is increased in the consumer’s memory (Sherman, Mackie, and Driscoll, 1990), consequently causing him to utilise this relationship in his product judgment. The priming effect typically portrays unintended influence of recent or recurrent experiences on subsequent thoughts, feelings, and behaviour (Bargh and Chartrand, 2000). Therefore, when a related attribute is primed by the category context, a positive or negative relationship associated with this attribute is activated in the consumer’s mind, and this relationship may be utilised in his consideration of the product.
Assimilation Effect versus Contrast Effect
There are two types of contextual effects – assimilation and contrast effect.
Assimilation effects are demonstrated when the introduction of an anchor (the primed category) leads to the displacement of the judgment of objects toward the anchor (Helson, 1964; Sherif & Hovland, 1961). On the other hand, contrast effect is noted to be the judgment
of a given target stimulus that is inversely related to the values of the context stimuli that accompany it. For example, in a study conducted by Sherman et al. (1978), participants were reported to judge recycling to be much less important in the context of the more important issues than in the context of trivial issues. Simply, contrast effect refers to a target judgment shift away from the context, as opposed to a target judgment shift toward the context, in the case of an assimilation effect.
Empirical Theories of Assimilation and Contrast Effect
Past theories and models on priming effects have been developed extensively to show that judgments are sensitive to the context in which they are made (Herr, 1989; Meyers-Levy, 1989). Of which, demonstrations attempting to define underlying conditions for assimilation and contrast effects have also been aplenty. In one such experiment done by Herr (1989), prior knowledge of the participant was found to be a condition affecting both categorisation and judgments effects. His first experiment found that assimilation occurs only in the condition where the primed category is moderate (price of real cars) and the target to be judged is ambiguous (hypothetical cars). All other conditions were found to produce contrast effects. Herr then conducted a second experiment using both real and hypothetical cars as target stimuli, to examine the impact of pre-existing knowledge differences on priming effects. Results showed that priming effects are present only when participants have high knowledge; participants with lack of knowledge would not be affected by priming effects, thus strongly indicating that priming effects are in fact a function of the level of pre-existing knowledge held by the individual (Herr, 1989).
In a similar study, Higgins et al. (1977) found that ambiguous stimuli are more likely to be captured by the primed category. Participants in the study were asked to do a seemingly unrelated experiment after being discreetly primed to either positive or negative trait
adjectives that can equally describe an ambiguous target person’s behaviour. It was revealed that participants who had been exposed to the positive traits had a significantly more
favourable impression of the target person than other participants who had been exposed to the negative traits.
Herr, Sherman and Fazio (1983) also suggested that the ambiguity of the target as well as extremity of the primed category would dictate which direction priming would take place. The study first primed participants with well-known animals of ambiguous ferocity, before asking them to evaluate animals known to be moderately ferocious, including two fictitious animals invented to serve as ambiguous stimuli (“jabo” and “lemphor”). Contrast effect was evident in the extreme and ambiguous category (e.g. “extremely high” or
“extremely low”), while assimilation effect was shown in judgments where moderate categories are primed and ambiguous stimuli are judged.
In another example, Meyers-Levy and Sternthal (1993) attempted a two-factor theory to explain the effects of assimilation and contrast. The theory successfully proved that when substantial cognitive resources are devoted to message processing and that the categories of the context has a slight overlap with the target, contrast would occur, while the absence of either one of these factors would prompt assimilation.
The Dimensional Range Overlap Model
While previous theories and models for judgment targets, like the social judgment theory (Sherif and Hovland, 1961) which described latitudes of acceptance and rejection, have suggested a possible range and overlap in contextual priming effects, the Dimensional Range Overlap Model (DROM) is the first context effect model that directly addresses the concept of a possible dimensional range and overlap (Chien et al., 2010).
The Dimensional Range Overlap Model suggests that contextual effects, whether assimilation or contrast, is determined by the “perceptions of amount of overlap between the target range and context range”, thereby giving a measurable range to both the context and target, and going on to establish context effects by way of a presence, or lack of, of overlap between plausible values of both the context and target. Plausible value is defined as the range of possible interpretations for a context or target.
According to the model tested over three experiments, the three factors that determine the amount of substantial overlap are: the width of the context range, the width of the target, as well as the distance between the context and the target. Where there is a substantial
overlap between the context and the target, assimilation is expected to occur, while contrast is expected in the absence of the said substantial overlap.
In Experiment 1 where an advertisement for a new car was used, the target range was manipulated to be sufficiently wide enough to overlap with the context range, or sufficiently narrow so as not to overlap with the context range, ceteris paribus. The results showed that where a substantial overlap was present between the ranges, assimilation was produced, and an absence of an overlap would create contrast, supporting the hypothesis.
In the subsequent Experiment 2 testing the same two hypotheses as Experiment 1, the relative distance between the representative values were in turn varied, holding the ranges of
both the context and target constant. Participants were first asked to note adjectives that they associated with products of a certain prestige range, followed by rating a moderately
prestigious brand of car, the Honda Civic. Finally, they were asked to report the prestige range of a product described by each of the adjectives that they had listed in the first part.
Again, the results showed consistency as before, further validating the hypotheses that assimilation would arise as expected when an overlap is developed, and where a non-overlap was created due to a distance between ranges, contrast would follow.
In the final Experiment 3 similar to that of Experiment 2, effects of range overlap created by a manipulation in the width of the context range, everything other variable
constant, were examined. Positive results were produced, showing that an overlap created by a wide enough context range would warrant assimilation, while a non-overlap caused by a narrow enough context would generate contrast. The Dimensional Range Overlap Model was once again supported.
In essence, the core of the Dimensional Range Overlap Model lies in whether the context and target range overlaps substantially by way of the three variables abovementioned, to produce either assimilation or contrast effect. The model was substantiated steadily over three experiments, indicating that contextual effects are in fact dependent on the
“dimensional range overlap”.
Illustrated in Figure 2.1 is a presentation of assimilation effect produced, as proposed in the Dimensional Range Overlap Model.
Figure 2.1 Assimilation Effect: (1) Narrow context range; (2) wide target range; (3) narrow relative distance between both context and target’s representative values.
Similarly, the Dimensional Range Overlap Model justified that contrast effect will be produced in the following situation depicted in Figure 2.2.
Figure 2.2 Contrast Effect: (1) Narrow context range; (2) wide relative distance between both context and target’s representative values; (3) narrow target range.
Reciprocity Hypothesis
As opposed to the many priming studies focusing primarily on shifts in perception of target before, Hsiao (2002) proposed a reciprocal shift in the contextual stimuli, either in a direction towards or away from the target, in addition to a shift in the target towards or away from the contextual stimuli in assimilation and contrast situations respectively. Simply, it means to say when a context prompts assimilation, the perception of the target may shift towards the context, but the perception of the contextual stimulus itself will also move towards the target (Figure 2.3). The same applies for contrast effects, as depicted in Figure
2.4.
Figure 2.3 Assimilation Effect: The perception of the context shift toward the target will occur simultaneously with the perception of the target shift toward the contextual stimulus.
Figure 2.4 Contrast Effect: The perception of the context shift away from the target will occur simultaneously with the perception of the target shift away from the contextual stimulus.
Three groups of participants were involved in Study 1 of the Hsiao’s (2002)
experiment to test for contrast effect (illustrated in Figure 2.4 above). The first group was told to evaluate only the target (low-humour stimulus, Schindler’s List), while another was
instructed to evaluate only the contextual stimulus (high-humour stimulus, The Simpsons).
The final group was tasked to evaluate both – first the contextual stimulus, then the target.
The results showed that the perceptions of the target and contextual stimulus shifted away from each other simultaneously, thus supporting the Reciprocity Hypothesis.
In an identical Study 2 tested for assimilation effect held subsequently, both the context and target stimuli were selected to be moderately humorous, Jurassic Park and Lion King respectively. Similarly, one group of participants was tasked to evaluate only the target stimulus, another group to evaluate only the contextual stimulus, and a third group to evaluate both – first the context, then the target. The results showed that the perceptions of the target and contextual stimuli shifted toward each other simultaneously this time, supporting the Reciprocity Hypothesis again.
The Reciprocity Hypothesis clearly demonstrated the concept of a simultaneous shift in the contextual stimuli as the target stimuli over two experiments, proving that judgments of context would indeed be affected by target judgments the same way target judgments affect contextual judgments. It is also a development of past research which had primarily been focused on shifts in target perceptions.
Extensions of the DROM and Reciprocity Hypothesis
Following the establishment of Chien et al.'s DROM, several theories were formed in attempts to extend the model. Of which, Lu (2012) incorporated Hsiao’s Reciprocity
Hypothesis and attempted examine contextual effects under multiple contexts. The new factor of attitude certainty, expected to dictate the magnitude of evaluation shift, was introduced though failed to be proven.
Tse (2012) also examined the influence of dual primes, one positive and negative prime each on product judgment, making a similar proposition that a higher attitude certainty will result in lower magnitude of shift, while lower attitude certainty would result in higher magnitudes of shift. She too, however, did not manage confirm her hypotheses. In addition, Tse’s (2012) suggestion of a possible sequence effect in the presentation of primes was not proven to have an impact on product judgment as well.
In another earlier DROM extension experiment, Lin (2006) utilised both exemplar and attribute information primes, to test for contextual effects under multiple primes. Over two experiments, Lin successfully showed that when exemplar primes are accompanied by the attributes they represent, both assimilation and contrast would occur depending on the existence of overlap or non-overlap of the combined primes. It was also proven that the occurrence of assimilation or contrast was not limited by the type of primes (exemplar or attribute), and it was the final range of overlap or lack thereof, after sequence effect took place, that mattered. Further, Lin also showed that priming effects on the main judgment dimension were more influenced by dimensions of more relevance to the main judgement dimension.