The survey was conducted for a period of two months and finally a sample of 421 usable cases was obtained. 34.9% of the cases were answered by male participants and 65.1% were answered by females. The average age of participants was 26.7 years.
The following bar charts illustrate some mentionable findings of the descriptive data analysis. The graphs show data grouped either by gender and video, or by language and video. Male participants are denoted by the number 1, female participants by the number 2; the languages by the numbers 1 to 4 (English =1;
German = 2; Chinese = 3; Other = 4); the videos are denoted by the numbers 1 to 6 (Gangnam Style = 1; Charlie bit my finger = 2; Old Spice = 3; Harlem Shake = 4;
Abercrombie Fitchthehomeless = 5; School fight = 6).
As expected, participants of different linguistic background scored the ease of understanding of the videos differently (see Figure 4-1). English speakers and Chinese speakers especially differ in scoring ease of understanding of Gangnam Style, the Old Spice commercial and the school fight. This may be partially due to language barriers or cultural differences (Korean culture or memes being more popular or more well-understood in Asia), or the commercial’s humor being targeted at an American audience.
Figure 4-1: Ease of Understanding grouped by language
Video 2 (featuring babies) gave the female audience a greater visual stimulus, while video 6 (the school fight) gave the male audience a greater visual stimulus (see Figure 4-2). Again, this finding is not surprising but goes to confirm old believes.
Targeting genders differently has long been an important part of marketing.
Figure 4-2: Visual stimulus grouped by gender
The participants perceived video 5 (A&F Fitchthehomeless) to provide the most ways for the viewer to get involved (see Figure 4-3). An interesting observation is that male participants overall gave higher scores for user involvement than females. It is not clear whether this is because male participants have more experience with video sharing or because they are more aware of opportunities to share videos.
Figure 4-3: User involvement
Figures 4-4 and 4-5 illustrate the results concerning positive and negative emotions felt by the participants while watching the videos. Video 1 (Gangnam Style) evoked the most positive emotions in the participants. Regarding Video 6, male participants felt negative emotions about as much as they felt positive emotions, while female participants had much more negative emotional responses toward the video.
This goes to show that, generally speaking, males do not perceive violence as negative as females do.
Figure 4-4: Positive Emotional Impact
Figure 4-5: Negative Emotional Impact
Figure 4-6 shows that the provocative power of the six videos differs for viewers of different language backgrounds. Gangnam Style, for example, scored worse among native English speakers but achieved quite similar scores among speakers of other languages. Gangnam Style was a global success. The Old Spice commercial scored the best among native English speakers but scored worse among speakers of other languages; the commercial was a huge success in the United States, but is not as well-known in other parts of the world. This may also be because the Old Spice brand is not as well-known outside the USA.
The Fitchthehomeless video scored the highest in terms of provocative power.
Although this does not directly correlate with its number of views on Youtube, it shows that the video makes people want to share its message (and not just view it).
This may have led to the aforementioned newspaper reports and A&F’s change in marketing strategy.
Figure 4-6: Provocative Power grouped by participants’ native languages
To test for internal consistency and reliability, we checked the values of Cronbach’s Alpha for the constructs of “Ease of Understanding”, “Simplicity” and
“Staying Power”. “Ease of Understanding” and “Simplicity” each have good Cronbach’s alpha values with 0.869 and 0.858, respectively. “Staying Power” has an excellent Cronbach’s alpha value of 0.928.
As a measure of the validity of the constructs proposed by the IVP Model, correlation analysis was performed (see Table 4-1).
相關
The findings of the correlation analysis support the hypothesized correlations between the variables proposed by the model. All variables are positively correlated as the model suggested. The “simplicity”-factor is represented by the variables “ease
of understanding” and “ease of sharing”. These two variables have very high correlation at a value of 0.749. This could be because, in many cases, viewers who think that a video is easy to understand for them may also deem it to be easy to understand for others. Furthermore we see that the variable of surprise (representing the concept of “Novelty”) has a high correlation with emotional impact with a value of 0.586. Although surprise was measured as part of emotions in the questionnaire, in the statistical analysis it is not part of emotional impact. User involvement and staying power also both have high correlation with provocative power, respectively, with values of 0.482 and 0.578.
In the next step, regression analysis was performed to test the hypotheses illustrated in Figures 2-2 to 2-5 (see Tables 4-2 to 4-7).
Anovab
模式 平方和 df 平均平方和 F 顯著性
1 迴歸 266.334 3 88.778 91.884 .000a
殘差 402.906 417 .966
總數 669.240 420
a. 預測變數:(常數), Surprise, Ease_of_Understanding, Sensory_Impact b. 依變數: Total_Emotion_minus_Boredom_and_Surprise
Table 4-2: Regression Analysis of Ease of Understanding, Sensory Impact, and Novelty in relation with emotional impact
模式 R 平方 調過後的 R 平方 估計的標準誤
1 .398 .394 .98295
a. 預測變數:(常數), Surprise, Ease_of_Understanding, Sensory_Impact b. 依變數: Total_Emotion_minus_Boredom_and_Surprise
Table 4-3: Regression Analysis of Ease of Understanding, Sensory Impact, and Novelty in relation with emotional impact
係數a
Ease_of_Understanding .057 .019 .118 2.945 .003 Sensory_Impact .096 .022 .180 4.348 .000 Surprise .206 .016 .514 12.878 .000 a. 依變數: Total_Emotion_minus_Boredom_and_Surprise
Table 4-4: Regression Analysis of Ease of Understanding, Sensory Impact, and Novelty in relation with emotional impact
The P-value is smaller than 0.05 (F = 91.884), which means that the model is statistically valid. The R-square value is 0.398, which means that around 40 % of the value of emotional impact can be explained by Ease of Understanding, Sensory impact and Surprise (Novelty).
The three variables “ease of understanding”, “sensory impact” and “Surprise” all have significant influence on the intervening variable “emotional impact”, with surprise (Novelty) being the biggest factor with a Beta value of 0.514. Sensory impact and ease of understanding have Beta values of 0.18 and 0.118, respectively.
Anovab
模式 平方和 df 平均平方和 F 顯著性
1 迴歸 884.046 6 147.341 100.480 .000a
殘差 607.079 414 1.466
總數 1491.124 420
a. 預測變數:(常數), Simplicity_UI_moderation, Staying_Power, Sensory_Impact, Total_Emotion_minus_Boredom_and_Surprise, Simplicity, User_Involvement b. 依變數: Provocative_Power
Table 4-5: Regression analysis of Provocative power
模式摘要
Table 4-6: Regression analysis of Provocative power
係數a
Total_Emotion_minus_Boredom_and_Surprise .504 .055 .338 9.231 .000
Staying_Power .295 .033 .327 8.996 .000
User_Involvement -.169 .169 -.109 -.997 .319
Simplicity -.083 .045 -.104 -1.854 .064
Simplicity_UI_moderation .081 .021 .477 3.752 .000 a. 依變數: Provocative_Power
Table 4-7: Regression analysis of Provocative power
Regression analysis showed that the variable “Surprise” was not statistically significant in measuring “Provocative power”, unlike originally assumed. Therefore the variable was left out the second time the regression analysis was performed. The results of the second regression analysis show that there is satisfactory statistical support for the remaining parts of the model, with F = 100.48 and the value of R-square being 0.593. An R-square value of 0.59 is considered quite high in social science because of the subjectivity of participants. “Sensory impact” has weak significance with a Beta value of 0.071. The variable of “User involvement” has to have “Simplicity” as the moderating variable to be statistically significant for measuring “Provocative Power”, just like the model suggested. With a beta value of 0.338, “Emotional Impact” proves to be a highly significant factor as already stated in prior research. “Staying Power” is also significant with a Beta value of 0.327.
Based on these findings, Figure 4-7 illustrates the adjusted model of innate virality potential. Out of the nine hypotheses, only hypothesis 5 was proven to be false.
Figure 4-7: The adjusted model of Innate Virality Potential