Network Analysis Lim 97

Editor's Notes

• #19: Although there are lots of attitude definitions, this study follows Fazio’s definition. That is, attitudes can be defined as associations, represented in memory, between attitude objects and their evaluations. In this definition, there are two key properties, attitude object and evaluation. Attitude objects are anything that can be designated including concrete objects and abstract objects. Concrete objects are such as desk, chair, car, dog, cat, and human beings, and abstract objects are such as freedom, power, good, bad, and poor. So, attitude objects can be considered symbols or concepts that can be expressed by a language. Second key property of attitude is evaluation. Evaluation is a summary assessment including all aspects of evaluative responding, such as cognitive beliefs, emotions, and past behaviors. That is, when we evaluate an attitude object, we consider our beliefs based on our knowledge. Also, we consider our emotions, such as happy, sad, guilty, and angry. Finally, we consider our past experience. So, an attitude can be represented as a simple two-node network like this diagram.
• #20: An attitude is often associated with other attitudes. The associations among attitudes are called interattitude structure. Traditionally, interattitude structure have been studied. Balance theory is the most representative. The diagram is a simple example based on the balance theory. Now, I will explain the diagram using my wife case. My wife is allergic to chicken, so she has a negative evaluation of all chicken cuisine. When I asked her “how about duck couisine?”, she said “I don’t like duck either because it is a kind of chicken.” In my wife’s case, the attitude toward duck is associated with the attitude toward chicken. From this simple example, we can consider that attitudes are interrelated to each other rather than isolated in our minds.
• #21: Although I simply explained the interattitude structure, it’s not simple. We know that a neuron in our brain stores information and it is interconnected to other neurons. Also, the neural networks in our brain are very complex like the left figure. This may indicate that the interattitude structure in our brains is very complicated. The Galileo model considers the complex neural networks in our brains and attempt to mimic and describe the interattitude structure in a multi-dimensional space as the right figure.
• #22: First, the Galileo model considers the categorization process that is a basic cognitive process in human brains. Through the categorization process, we can easily recognize our world. For your better understanding, I’d like to give a simple example based on my son’s case. I have a son. He is three years old. He likes apple juice but he doesn’t like milk. One day, I gave him a lemonade instead of apple juice. Actually, he had never seen the lemonade. He sipped a lemonade then he looked like thinking for a while. Then, he said to me “lemonade is apple juice. I like it.” So, I asked him, “how about milk?” He said, “Milk is different. It is not apple juice.” In my son’s case, he categorized his favorite by using apple juice. That is, lemonade is included in apple juice category but milk is not. The categorization process is based on similarity and dissimilarity between objects. Likewise, although my son is only three years old. He is able to perceive the world by using the categorization process. So, the categorization process is basic and inevitable to understand our world.
• #23: Second, the Galileo space is a cognitive map in human brains. As I said before, the Galileo model mimics the neural networks in our brains. That is, the Galileo model is mapping the mental structure. The Galileo model measures the concept distances based on the dissimilarity between every possible pair of concepts. In my son’s case, lemonade and apple juice are very similar concepts. So, their positions in the Galileo space are very close to each other. On the contrary, milk is different from lemonade and apple juice. So. The position of milk is relatively far from the other concepts. That is, the Galileo model makes a simple cognitive map considering concept distances.
• #24: Lastly, the self concept is a very special concept in the Galileo model. It is a conscious representation of a person. In the neuroscience, many researchers empirically found differentiated neural activations in self-related cognitive process. These findings supports that the representation of self exists as distinct information processing in human brains. In the Galileo space, like the other concepts, the self concept is close to relevant concepts that describe the person. Conversely, it is far from irrelevant concepts that do not define the person. In my son’s case, apple juice and lemonade may be close to his self concept, but milk may be far from the self in the Galileo space. So, an attitude can be identified as the distance between the self concept and the designated concept. Lastly, the Galileo measure is the direct-magnitude scaling for precise measurement. The scaling is based on the real number system including absolute zero point and unbounded real number. So, the scaling can be fully satisfied with the property of ratio scale and mathematical operations.
• #25: To examine the interattitude structure of the target audiences in IDCs, the Galileo survey was conducted. This study used a convenient sample of college students as a segment of the campaign targets. The Galileo survey instrument is constructed by the 12 concepts. So, a complete list of 66 pair comparisons was included in the survey instrument.
• #26: This is an example of the Galileo survey. Cooperation and conflict are 100 units apart. That is the criterion pair which helps the respondents to judge the differences between concepts as a standard distance. Respondents were instructed to report a real number without any upper limit. But note that , a number less than 100, it means that the differences between any paired concepts are less different than the standard distance. A number above 100, it means that the differences are more different. Also, if any paired concept is perceived to be the same, zero point should be entered. Lastly, if respondents do not know the differences, blank answers were allowed.
• #28: To better understand the distances among concepts, this figure represents the interrelationships among 12 concepts in the three dimensional space accounting for about 67% of the real variance. As you can see, poverty and international aid are very far from the self concept. Also, international aid does not have neighboring concepts. On the contrary, health, education and human rights are relatively close to the self concept.
• #29: This diagram shows the Galileo message strategy based on the target audiences’ interattitude structure. As you can see, if a campaign message emphasizes the strong connections between international aid and the three relevant concepts, such as education, health, and human rights, the combined pulling effects of the three concepts will move the international aid close to the self concept. The principle of the Galileo message strategy is simple. Actually, it’s basic physics. I will show you a practical example. Could you help me? The Galileo message strategy is based on the mathematical representation of physical vector in the spatial coordinate system. So, the Galileo message optimizing procedure is calculating all possible concept combinations from one to four concept combinations. Then, through the message optimizing procedure, the Galileo results shows the best concept pair that makes the best expected position of international aid close to the target. As you can see, if the full effects of the best message strategy were obtained, the distance between the self concept and the expected position of international aid becomes 14.41 units. It’s very close.
• #30: Based on the Galileo message strategy, I made a simple campaign message like this. As you can see, the message does not include emotional and negative concepts.
• #31: The different concept locations is graphically displayed as the figure. As you can see, non-manipulated concept locations are relatively stable across the three group spaces, whereas the manipulated concept locations are relatively different. Particularly, the locations of international aid are more different than others.
• #32: This table presents the differences of manipulated concept-pair distances. As you can see, all distances in two treatment groups decreased. Particularly, the distances between international aid and the self concept relatively more decreased. The results of Kruska-Wallis and Mann Whitney U test indicated that the distances are significantly different and the distances in the two treatment groups are smaller than the distances in the control group. These results statistically supports the Galileo message effectiveness for attitude change.