Academia.eduAcademia.edu

Figure 1 – uploaded by Phoey Lee Teh

See full PDFdownloadDownload figure
example of capitalization used for emotional expression is addressed by Anil et al. [1]. According to them, one of the factors to detect emotions is referring to the user style of typing including the usage of letter capitalization. They stated that capitalization is applied more often when the expression is in the category of mood, which is 80% of the cases. Similar to the study of Byron [2] which agreed that capitalization could be used to express emotions in text. In fact, usage of capitalization indicates how intense the emotions are [2]. Moreover, Ledbetter et al. [6] stated that fully capitalized text has interpreted as shouting and unwelcome emotion. However, whether the usage of capitalization is used to express a happy mood or unwelcome emotions, it has impact on the reader opinions when Figure 1 The Impact of Capitalized Social Media Text towards Purchase Likelihood We have collected social media texts from online reviews manually. There is a total of 1014 reviews based ©n [Opdifferent

Figure 1 example of capitalization used for emotional expression is addressed by Anil et al. [1]. According to them, one of the factors to detect emotions is referring to the user style of typing including the usage of letter capitalization. They stated that capitalization is applied more often when the expression is in the category of mood, which is 80% of the cases. Similar to the study of Byron [2] which agreed that capitalization could be used to express emotions in text. In fact, usage of capitalization indicates how intense the emotions are [2]. Moreover, Ledbetter et al. [6] stated that fully capitalized text has interpreted as shouting and unwelcome emotion. However, whether the usage of capitalization is used to express a happy mood or unwelcome emotions, it has impact on the reader opinions when Figure 1 The Impact of Capitalized Social Media Text towards Purchase Likelihood We have collected social media texts from online reviews manually. There is a total of 1014 reviews based ©n [Opdifferent

Related Figures (4)

Table 2 Sentiment tools used in this study
Table 2 Sentiment tools used in this study
The list of selected statements is shown in Table’1. Then, we simulated the statement with three ‘capitalization formats. Five hundred respondents were requestéd to rate after they have read the text. Likelihood includes 7-pointsiscale from 1-strongly dislike till 7-strongly like. Table 1. Positive and negative reviews for the survey
The list of selected statements is shown in Table’1. Then, we simulated the statement with three ‘capitalization formats. Five hundred respondents were requestéd to rate after they have read the text. Likelihood includes 7-pointsiscale from 1-strongly dislike till 7-strongly like. Table 1. Positive and negative reviews for the survey
Table 3 Correlation statists The negative 0.133 coefficient demonstrates a connection between these two variables in the same opposite way that positive correlation coefficient does, with the relative strengths ‘ef likelihood are the same of 0.133. In others words, when_a written comment is in capital letters, it has certainly impact,and influence on the reader, and it is in 2-tailed with a significant‘value of 0.003. Interesting that letter capitalization used for negative statement intrigues a higher likelihood. Even greater comparing to text appeared in it low case letters. 99 66 very happy”, “my favourite” and “I am glad”. And looking from the less likely to purchase a product, which is below 3, most comments are negative polarity, such as the word “Never, ever and never”, “what a lie” and “J hate it”. From Figure 1, the impact for the positive texts has shown a clear Likelihood in opinion, mostly above 3 of the Likert scale. We observed that both capital letters used for the whole statement content words in capital letters had its impacts toward the reader. Respondents had answered with a higher likelihood to purchase after reading those texts in capital letters. In addition, when it comes to the negative comments, surprisingly, it showed a slightly different opinion on texts that appear in negative polarity. The result indicates that there is contradictory even the statement is negative in polarity, and with a capital letter in use.
Table 3 Correlation statists The negative 0.133 coefficient demonstrates a connection between these two variables in the same opposite way that positive correlation coefficient does, with the relative strengths ‘ef likelihood are the same of 0.133. In others words, when_a written comment is in capital letters, it has certainly impact,and influence on the reader, and it is in 2-tailed with a significant‘value of 0.003. Interesting that letter capitalization used for negative statement intrigues a higher likelihood. Even greater comparing to text appeared in it low case letters. 99 66 very happy”, “my favourite” and “I am glad”. And looking from the less likely to purchase a product, which is below 3, most comments are negative polarity, such as the word “Never, ever and never”, “what a lie” and “J hate it”. From Figure 1, the impact for the positive texts has shown a clear Likelihood in opinion, mostly above 3 of the Likert scale. We observed that both capital letters used for the whole statement content words in capital letters had its impacts toward the reader. Respondents had answered with a higher likelihood to purchase after reading those texts in capital letters. In addition, when it comes to the negative comments, surprisingly, it showed a slightly different opinion on texts that appear in negative polarity. The result indicates that there is contradictory even the statement is negative in polarity, and with a capital letter in use.
From the results, they are all\significant: The highest correlations that highly correlate with*eachjother/is when we correlate the not capitalized negative text with the text that contains all capital etters, which is 0°65 with a 2-tailed significant value of 0.000. Other results overall have Shown a significant impact of all of the not capitalized social media text (both in positive and negative polarity) andyhave‘aneGative correlation with all the social media text that,contain a particular word in capital letters (for both in positive‘and negative polarity). 4N is referred to Negative in Table 4 Furthermore, we continued to explore the significance‘of our study among all these scenarios. Table 3 shows all the résults processed with the computational values organized in the-way, that 1s shown in the correlations (first column of Table 3),“NoCapitalization” is referred to a text without any capitalization. We computed by summarizing up all the 500 respondents likelihood results for both positive and negative comments and’separate them accordingly to the range of positive to negative, drawingand,etganizing similar to the statement from Figure 1.
From the results, they are all\significant: The highest correlations that highly correlate with*eachjother/is when we correlate the not capitalized negative text with the text that contains all capital etters, which is 0°65 with a 2-tailed significant value of 0.000. Other results overall have Shown a significant impact of all of the not capitalized social media text (both in positive and negative polarity) andyhave‘aneGative correlation with all the social media text that,contain a particular word in capital letters (for both in positive‘and negative polarity). 4N is referred to Negative in Table 4 Furthermore, we continued to explore the significance‘of our study among all these scenarios. Table 3 shows all the résults processed with the computational values organized in the-way, that 1s shown in the correlations (first column of Table 3),“NoCapitalization” is referred to a text without any capitalization. We computed by summarizing up all the 500 respondents likelihood results for both positive and negative comments and’separate them accordingly to the range of positive to negative, drawingand,etganizing similar to the statement from Figure 1.
Related topics:
PsychologyComputer Science
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved