How the correlation of Istagram tweets readability and brand hedoism affect audiecne engagment
Dublin Core
Title
How the correlation of Istagram tweets readability and brand hedoism affect audiecne engagment
Creator
Jiehong Wu
Date
Sep 8th 2021
Description
Social media marketing is increasing in importance and more and more brands are embracing social media to increase their brand reach and communicate with their audience. However, there is still little empirical research on how brand message features affect consumer engagement. This study focuses on the impact of readability as an influence on consumer engagement while also noting that the effect of hedonic value of a brand may potentially moderate the level of audience engagement. An experiment based on a sample of 20 of the 100 brands covered by Forbes Media was conducted for this study. In total, a sample of 400 Instagram tweets were collected and analysed for their text readability and audience engagement. Still, the results did not indicate a significant interaction between readability and engagement. A careful analysis of the difficulties and shortcomings encountered in this experiment provides some insights for any subsequent research on the readability of short-form communication by brands.
Subject
Readability, Brand hedonism, readbility formula, audience engagment
Source
Research Question & Hypotheses
The research question for this study is: can the readability of tweets influence the level of audience engagement?
As readability increases the perception associated with processing fluency (Rennekamp, 2012), the ability to process information fluently makes the target message more appealing to the audience, and visual fluency in processing information can also increase people's perception of the processing target (Novemsky et al., 2007). Language that can be processed fluently also enhances consumer perceptions (Lee & Aaker, 2004; Lee & Labroo, 2004). Thus, for most brands with low levels of hedonism, higher tweet readability means higher processing fluency which can reduce audience metacognitive difficulties and thus increase tweet engagement levels.
At the same time, for products with high hedonic demand, lower familiarity and uniqueness may provide consumers with greater signals of value, with metacognitive difficulties increasing the appeal of the product by making it appear unique or unusual. More easily processed messages reduce the appeal of the product, possibly because they appear too familiar and therefore less consistent with the perception of uniqueness (Pocheptsova, Labroo, 2004；Pocheptsova, Labroo, and Dhar 2010).
It is therefore hypothesised that text features associated with greater readability will be positively associated with consumer engagement with the message. However, given the presence of brand hedonistic features, it can be argued that low readability of messages may increase consumer engagement in brand tweets with higher levels of hedonism instead.
Data collection for the experiment
From the above, whether the readability of the tweet text and the level of brand hedonism of the brand to which the tweet belongs combine to influence consumer engagement with the brand's social tweets must be determined.
Instagram was chosen because it is one of the world's most popular social networks, with around one billion active users per month, and over two-thirds of the Instagram audience is under the age of 34, making the platform particularly attractive to marketers. At the same time, Instagram is an open public platform and information on experiments can be easily accessed by searching for the brand name to use for experiments. This included the number of followers of the brand, the history and content of the tweets, the number of comments and the number of likes. To make the experiment practical, 20 tweets from each of the 20 brands (see Step 1 below) were selected for the experiment. The process of collecting information was as follows.
Step 1 involved the selection of the experimental subject brands. The results of a hedonistic study of the TOP 100 most valuable brands in the world on the Forbes list (Davis et al., 2019) were used to rank the brands from the highest to lowest level of hedonism using the hedonism index (from Davis et al 2019 survey, for detail see Degree of brand hedonism) as the key indicator. A computer generated a random series of 20 numbers from 1 to 100, and the numbers in this series were used to correspond to the serial numbers of the brands in the hedonism table. The following 20 brands for this experiment were selected: Goldman, Sachs, HSBC, Walmart, Thomson Reuters, IBM, Subway, Verizon, HP, Hyundai USA, Boeing, Chanel, Coach, ESPN, Starbuck coffee, Nike, Gucci, amazon, Mercedes-Benz, Google, Porsche（For the logic behind the selection of these brands, please see Degree of brand hedonism）.
In Step 2, text samples and audience engagement data were collected. In order to control the variables of the experiment as much as possible, text samples of tweets were collected from August 12 to August 13, 2021, and only tweets with 30-150 words were selected to control the discrete nature of the sample. To avoid the influence of rich media such as video/audio on audience engagement, tweets in the form of rich media were also excluded from the sample, ensuring that all samples contained only images and textual content. The number of likes and comments on each tweet was also recorded. To ensure that the selected sample of tweets accumulated enough likes and comments, all samples were posted before 7 August, ensuring that they had five days to accumulate interaction data with the audience. According to the official Twitter report (Twitter，2016), due to the instantaneous nature of the social media platform, in general tweets were largely ignored by audiences a week after they were posted and they therefore found it difficult to accumulate further feedback data.
Step 3 was the readability analysis of the text samples. Considering that some of the tweet samples were less than 100 words, and that The Flesch Reading Ease formula recommends a text count of 100 words or more, and considering the validity of the formula, this experiment combined two or more samples for tweets with a text count of fewer than 100 words to obtain at least 100 words before using the formula for analysis , so as to the average readability score for this group of samples was calculated (See Message readability for details of the Flesch Reading Ease formula)
Variables and measures
Message readability
Readability formulas have evolved to the point where there are now over 40 readability formulas (Heydari, 2012). The most widely known of these is Rudolph Flesch's formula, created in 1948 and published in the Journal of Applied Psychology in his article ' A New Readability Yardstick'. This formula is considered to be one of the oldest and most accurate formulas for readability, and has made Flesch an authority on readability scholarship. It was originally created to assess the readability of readers at grade level and is widely regarded as an accurate measure without much scrutiny. The formula is best suited to school texts, but it is also widely used by US government agencies (including the US Department of Defense) to assess the readability of their published documents and forms, and some states even require insurance policies to achieve a Flesch reading-ease score of 45 or higher. The Readability Formula is even installed in Microsoft Office Word, where the program checks the spelling and grammar of a text as well as its readability level (Heydari, 2012).
The specific mathematical formula is as follows:
RE = 206.835 – (1.015 x ASL) – (84.6 x ASW)
RE = Readability Ease the output is a number ranging from 0 to 100. The higher the number, the easier the text is to read
ASL = Average Sentence Length (i.e., the number of words divided by the number of sentences)
ASW = Average number of syllables per word (i.e., the number of syllables divided by the number of words)
Table1: Description and predicted reading grade for Flesch Reading Ease Score (Stein, 1984)
Score School level (US) Notes
100.00–90.00 5th grade Very easy to read. Easily understood by an average 11-year-old student.
90.0–80.0 6th grade Easy to read. Conversational English for consumers.
80.0–70.0 7th grade Fairly easy to read.
70.0–60.0 8th & 9th grade Plain English. Easily understood by 13- to 15-year-old students.
60.0–50.0 10th to 12th grade Fairly difficult to read.
50.0–30.0 College Difficult to read.
30.0–10.0 College graduate Very difficult to read. Best understood by university graduates.
10.0–0.0 Professional Extremely difficult to read. Best understood by university graduates.
As can be deduced, the text samples should ideally contain short sentences and words. As most texts on social media are short sentences or words, the Flesch Reading Ease Score was considered to be the most suitable tool for measuring the readability of tweets in this experiment. The Flesch Reading Ease readability formula in the online automatic readability checker was used in this study (https://readabilityformulas.com/free-readability-formula-tests.php).
Consumer engagement with brands
As Instagram retweets can only be sent to friends or groups of friends and not to the user's public page, this experiment only measured the number of "likes" (users click on the red love button below the tweet or double click on the tweet to like it) and comments on the tweet, as retweet data is difficult to collect. As described in the data collection process, the collected tweets were given at least 5 days to accumulate comments and likes. These two numbers (comments+likes) were then added together and divided by the number of brand trackers and multiplied by 10,000 to obtain the final audience engagement level score.
Degree of brand hedonism
As this experiment was limited by resources and practicability, the results of the Davis et al 2019 survey on the level of brand hedonism were used directly here. The following is an introduction to the process of Davis et al.'s 2019 survey on levels of brand hedonism which measured the level of hedonism of 100 brands primarily by human judges on a rating scale (four non-social media active brands were finally excluded, giving a final total of 96 brands).
In the Davis et al. experiment, a total of 200 human judges participated in scoring the level of brand hedonism. Each judge was randomly assigned to 10 brands and they scored each brand on four hedonism-related indicators: fun, excitement, thrill and pleasure, on a scale of 1 'not at all' to 7 'very much'. The final brand hedonism index was derived from these four indicators and then averaged across the 10 judges. The judges who participated in the experiment were recruited from the Amazon Mechanical Turk online panel. A total of 200 judges participated in the experiment, 61% of whom were male and the remainder female, all aged 35 years and of unknown ethnic background, but all participants were US residents. Detailed results of the original experiment can be found in Appendix A.
In this particular experiment, the brands were ranked from the highest to lowest hedonism level using the hedonism index of Davis et al. A computer generated a random series of 20 numbers from 1-96, and the numbers in this series were used to correspond to the serial numbers of the brands in the hedonism table as the experimental subjects. Table 2 shows the average hedonism scores of the 20 brands selected. Figure 1 shows the conceptual model for this experiment, the relevant experimental variables and the control variables.
Table 2 the brand hedonism scores
NO Brands Mean SD
1 Porsche 6.05 1.26
2 Google 5.92 0.95
3 Mercedes-Benz 5.68 1.13
4 Amazon
5.41 1.59
5 Gucci 5.29 1.13
6 Nike 5.05 1.40
7 Starbucks Coffee 4.89 1.19
8 ESPN 4.75 1.93
9 Coach. Inc 4.53 1.60
10 Chanel 4.40 1.26
11 Boeing 4.27 1.77
12 Hyundai USA 4.12 1.45
13 HP 3.86 1.75
14 Subway 3.75 1.67
15 Verizon 3.75 1.36
16 IBM 3.45 1.47
17 Walmart 3.15 1.39
18 Walmart 3.15 1.39
19 HSBC 2.89 1.35
20 Goldman Sachs 2.14 1.23
The research question for this study is: can the readability of tweets influence the level of audience engagement?
As readability increases the perception associated with processing fluency (Rennekamp, 2012), the ability to process information fluently makes the target message more appealing to the audience, and visual fluency in processing information can also increase people's perception of the processing target (Novemsky et al., 2007). Language that can be processed fluently also enhances consumer perceptions (Lee & Aaker, 2004; Lee & Labroo, 2004). Thus, for most brands with low levels of hedonism, higher tweet readability means higher processing fluency which can reduce audience metacognitive difficulties and thus increase tweet engagement levels.
At the same time, for products with high hedonic demand, lower familiarity and uniqueness may provide consumers with greater signals of value, with metacognitive difficulties increasing the appeal of the product by making it appear unique or unusual. More easily processed messages reduce the appeal of the product, possibly because they appear too familiar and therefore less consistent with the perception of uniqueness (Pocheptsova, Labroo, 2004；Pocheptsova, Labroo, and Dhar 2010).
It is therefore hypothesised that text features associated with greater readability will be positively associated with consumer engagement with the message. However, given the presence of brand hedonistic features, it can be argued that low readability of messages may increase consumer engagement in brand tweets with higher levels of hedonism instead.
Data collection for the experiment
From the above, whether the readability of the tweet text and the level of brand hedonism of the brand to which the tweet belongs combine to influence consumer engagement with the brand's social tweets must be determined.
Instagram was chosen because it is one of the world's most popular social networks, with around one billion active users per month, and over two-thirds of the Instagram audience is under the age of 34, making the platform particularly attractive to marketers. At the same time, Instagram is an open public platform and information on experiments can be easily accessed by searching for the brand name to use for experiments. This included the number of followers of the brand, the history and content of the tweets, the number of comments and the number of likes. To make the experiment practical, 20 tweets from each of the 20 brands (see Step 1 below) were selected for the experiment. The process of collecting information was as follows.
Step 1 involved the selection of the experimental subject brands. The results of a hedonistic study of the TOP 100 most valuable brands in the world on the Forbes list (Davis et al., 2019) were used to rank the brands from the highest to lowest level of hedonism using the hedonism index (from Davis et al 2019 survey, for detail see Degree of brand hedonism) as the key indicator. A computer generated a random series of 20 numbers from 1 to 100, and the numbers in this series were used to correspond to the serial numbers of the brands in the hedonism table. The following 20 brands for this experiment were selected: Goldman, Sachs, HSBC, Walmart, Thomson Reuters, IBM, Subway, Verizon, HP, Hyundai USA, Boeing, Chanel, Coach, ESPN, Starbuck coffee, Nike, Gucci, amazon, Mercedes-Benz, Google, Porsche（For the logic behind the selection of these brands, please see Degree of brand hedonism）.
In Step 2, text samples and audience engagement data were collected. In order to control the variables of the experiment as much as possible, text samples of tweets were collected from August 12 to August 13, 2021, and only tweets with 30-150 words were selected to control the discrete nature of the sample. To avoid the influence of rich media such as video/audio on audience engagement, tweets in the form of rich media were also excluded from the sample, ensuring that all samples contained only images and textual content. The number of likes and comments on each tweet was also recorded. To ensure that the selected sample of tweets accumulated enough likes and comments, all samples were posted before 7 August, ensuring that they had five days to accumulate interaction data with the audience. According to the official Twitter report (Twitter，2016), due to the instantaneous nature of the social media platform, in general tweets were largely ignored by audiences a week after they were posted and they therefore found it difficult to accumulate further feedback data.
Step 3 was the readability analysis of the text samples. Considering that some of the tweet samples were less than 100 words, and that The Flesch Reading Ease formula recommends a text count of 100 words or more, and considering the validity of the formula, this experiment combined two or more samples for tweets with a text count of fewer than 100 words to obtain at least 100 words before using the formula for analysis , so as to the average readability score for this group of samples was calculated (See Message readability for details of the Flesch Reading Ease formula)
Variables and measures
Message readability
Readability formulas have evolved to the point where there are now over 40 readability formulas (Heydari, 2012). The most widely known of these is Rudolph Flesch's formula, created in 1948 and published in the Journal of Applied Psychology in his article ' A New Readability Yardstick'. This formula is considered to be one of the oldest and most accurate formulas for readability, and has made Flesch an authority on readability scholarship. It was originally created to assess the readability of readers at grade level and is widely regarded as an accurate measure without much scrutiny. The formula is best suited to school texts, but it is also widely used by US government agencies (including the US Department of Defense) to assess the readability of their published documents and forms, and some states even require insurance policies to achieve a Flesch reading-ease score of 45 or higher. The Readability Formula is even installed in Microsoft Office Word, where the program checks the spelling and grammar of a text as well as its readability level (Heydari, 2012).
The specific mathematical formula is as follows:
RE = 206.835 – (1.015 x ASL) – (84.6 x ASW)
RE = Readability Ease the output is a number ranging from 0 to 100. The higher the number, the easier the text is to read
ASL = Average Sentence Length (i.e., the number of words divided by the number of sentences)
ASW = Average number of syllables per word (i.e., the number of syllables divided by the number of words)
Table1: Description and predicted reading grade for Flesch Reading Ease Score (Stein, 1984)
Score School level (US) Notes
100.00–90.00 5th grade Very easy to read. Easily understood by an average 11-year-old student.
90.0–80.0 6th grade Easy to read. Conversational English for consumers.
80.0–70.0 7th grade Fairly easy to read.
70.0–60.0 8th & 9th grade Plain English. Easily understood by 13- to 15-year-old students.
60.0–50.0 10th to 12th grade Fairly difficult to read.
50.0–30.0 College Difficult to read.
30.0–10.0 College graduate Very difficult to read. Best understood by university graduates.
10.0–0.0 Professional Extremely difficult to read. Best understood by university graduates.
As can be deduced, the text samples should ideally contain short sentences and words. As most texts on social media are short sentences or words, the Flesch Reading Ease Score was considered to be the most suitable tool for measuring the readability of tweets in this experiment. The Flesch Reading Ease readability formula in the online automatic readability checker was used in this study (https://readabilityformulas.com/free-readability-formula-tests.php).
Consumer engagement with brands
As Instagram retweets can only be sent to friends or groups of friends and not to the user's public page, this experiment only measured the number of "likes" (users click on the red love button below the tweet or double click on the tweet to like it) and comments on the tweet, as retweet data is difficult to collect. As described in the data collection process, the collected tweets were given at least 5 days to accumulate comments and likes. These two numbers (comments+likes) were then added together and divided by the number of brand trackers and multiplied by 10,000 to obtain the final audience engagement level score.
Degree of brand hedonism
As this experiment was limited by resources and practicability, the results of the Davis et al 2019 survey on the level of brand hedonism were used directly here. The following is an introduction to the process of Davis et al.'s 2019 survey on levels of brand hedonism which measured the level of hedonism of 100 brands primarily by human judges on a rating scale (four non-social media active brands were finally excluded, giving a final total of 96 brands).
In the Davis et al. experiment, a total of 200 human judges participated in scoring the level of brand hedonism. Each judge was randomly assigned to 10 brands and they scored each brand on four hedonism-related indicators: fun, excitement, thrill and pleasure, on a scale of 1 'not at all' to 7 'very much'. The final brand hedonism index was derived from these four indicators and then averaged across the 10 judges. The judges who participated in the experiment were recruited from the Amazon Mechanical Turk online panel. A total of 200 judges participated in the experiment, 61% of whom were male and the remainder female, all aged 35 years and of unknown ethnic background, but all participants were US residents. Detailed results of the original experiment can be found in Appendix A.
In this particular experiment, the brands were ranked from the highest to lowest hedonism level using the hedonism index of Davis et al. A computer generated a random series of 20 numbers from 1-96, and the numbers in this series were used to correspond to the serial numbers of the brands in the hedonism table as the experimental subjects. Table 2 shows the average hedonism scores of the 20 brands selected. Figure 1 shows the conceptual model for this experiment, the relevant experimental variables and the control variables.
Table 2 the brand hedonism scores
NO Brands Mean SD
1 Porsche 6.05 1.26
2 Google 5.92 0.95
3 Mercedes-Benz 5.68 1.13
4 Amazon
5.41 1.59
5 Gucci 5.29 1.13
6 Nike 5.05 1.40
7 Starbucks Coffee 4.89 1.19
8 ESPN 4.75 1.93
9 Coach. Inc 4.53 1.60
10 Chanel 4.40 1.26
11 Boeing 4.27 1.77
12 Hyundai USA 4.12 1.45
13 HP 3.86 1.75
14 Subway 3.75 1.67
15 Verizon 3.75 1.36
16 IBM 3.45 1.47
17 Walmart 3.15 1.39
18 Walmart 3.15 1.39
19 HSBC 2.89 1.35
20 Goldman Sachs 2.14 1.23
Publisher
Lancaster University
Format
Data/Excel.xlsx
Identifier
Wu 2021
Contributor
Chloe Keung, Elena Ball
Rights
Open
Relation
None
Language
English
Type
Word
Coverage
LA1 4YZ
LUSTRE
Supervisor
Robert Davies
Project Level
MSC
Topic
Marketing
Sample Size
20 tweets from each of the 20 brands
Statistical Analysis Type
Regression
Files
Collection
Citation
Jiehong Wu, “How the correlation of Istagram tweets readability and brand hedoism affect audiecne engagment ,” LUSTRE, accessed May 2, 2024, https://www.johnntowse.com/LUSTRE/items/show/113.