Summary: This is a prototype test that I ran as part of my studies and make recommendations to the client organization. The research showed that identifiable characters positively influence trust and enjoyment derived while reading a narrative. Those findings have an implication for improving brand credibility and recall which are important elements for brand differentiation and increasing purchases.
Prototyping
The need for prototyping and doing this research occurred after the content analysis, card sorting, and the creation of content prototypes in the pitchbook. Thus, the information that was obtained during the content creation and with the help of the prototypes that were modified for this experiment, the research aims to identify what type of content and which modifications will have an effect on the content perception by the target audience.
Participants
The ideal number of participants per group in A/B tests should be around 40 to 50 for a normal distribution of the spread (Menger, personal communication, 3 February 2021). The current study employs 4 groups; so, the ideal number of participants was set to 20 per group. The total received responses was 90, and slightly above 20 per group.
The target respondents are aged between 19 to 35 years old, with bachelor’s degrees, and enthusiasts in adventure activities like hiking, kayaking, biking, and swimming. Gender is also measured in the current study, though due to the nature of the research it was from the criteria for the participants. The decision was made to include both male and female participants even though the target persona is male. See more in Design.
To measure the persona fit, questions related to the frequency of engagement in an outdoor activity, the preferred outdoor activity, age, education, and gender were collected in the final part of the questionnaire.
Materials
The materials used in the study are two design elements with an identifiable character on each mock-up. Each character was shortly introduced in the beginning. The participant also could see an image of the character and read the narrative. The variables that were modified are character similarity and persuasion principle.
Factors
The character similarity was developed using two degrees – high and low – based on the findings of the target persona.
In the introduction that was in the form of text and in third-person perspective, the characters that were high in similarity talked about their love for nature and adventure activities while characters low in similarity mentioned how they don’t often do physical exercise or are not risk-takers at all and prefer to feel safe and comfortable.
Additionally, the high in similarity characters were only introduced by their Western name in order not to create a greater dissimilarity between the target persona and the character. While the low similarity characters had names that were indicating Latin and Spanish origins, those aged above 60, and nationality that is different from the ones of the targeted participants for this research.
| Character Similarity: Introduction | |
|---|---|
| High in similarity | Low in similarity |
| Story 1: This is Mark. He is an open-minded person who enjoys being in nature. Mark likes experiencing a variety of adventure activities. The physical activities allow him to get rid of stress and daily responsibilities. Here he talks about his experience on a hike. | Story 1: This is Miguel. He is 62 years old and comes from Argentina. Miguel enjoys hiking now and then and spending time in nature, even though he rarely does physical activities in his daily life. Here he talks about his experience on a hike. |
To ensure that the character similarity will have an effect on the story, all stories were told in 1st person narrative style (‘I liked the hike…’) as it was found that the greatest similarity was in a condition of 1st person narration (Igartua, Rodríguez-Contreras, Ramos, Garay, & Frutos, 2020).
Also, the images on the prototype represented the characters visually in terms of age and context (dressed in sports clothing).
In the authority condition, the characters were talking about the help they received from the guides about the area, equipment, and skills they’ve learned.
In the liking condition, the characters were talking about the activities and being in nature as those are the favourite aspects of the trip for the persona.
As it is intended to create similarity, the likes of the persona based on the research insights have been used against the dislikes to create a low similarity condition.
In research (Zorn, Mata, & Alves, 2022), it was found that people who share their likes (positive attitudes) are more likely to like each other and find similarities in contrast to people who openly shared their dislikes (negative attitudes).
So, based on previous insights into the persona, it is expected that the liking condition will have a stronger effect on enjoyment and intent to visit while the authority condition will have an effect on perceived trustworthiness and intent to visit.
Theoretical Framework
The character similarity was chosen to be manipulated as it was found that the content that showed greater character similarity alongside character identification and narrative transportation had an influence on the intentions of the target market (Igartua, Rodríguez-Contreras, Ramos, Garay, & Frutos, 2020).
Thus, the research aims to explore whether participants identify with the character and what effect the identification has on the perception of the content and the brand in the adventure tourism context. Though, effects are expected to be seen more prominently in the high similarity conditions in contrast to the low similarity groups.
Additionally, the liking principle is used to create greater similarity as it was found that people like others who are like them and have more positive features in contrast to people who they dislike have more negative features (Alves, Koch, & Unkelbach, 2016). So, it is expected that liking will create similarity and when one likes someone, they find them to be like each other (similar).
The narrative message that is defined by the imaginable plot, verisimilitudes, and the character has been constructed for the purpose of exploring the character identification and narrative transportation effects on the dependent variables (van Laer, Feiereisen, & M.Visconti, 2019).
Research (Igartua, Rodríguez-Contreras, Ramos, Garay, & Frutos, 2020) has found that character similarity, identification, and narrative transportation had greater effects on changing behavioural intentions in comparison to other conditions where the effects of those variables lessened. And through the narrative plot of the ups and downs of the characters sharing about their hiking adventure in Northern Europe the researcher expects to be able to recreate the effects of narrative transportation and character identification.
The second independent variable is the persuasion principle. Similarly, the persuasion strategies are observed in the study to understand to what extent they have an influence on brand attitudes based on the perception of the content.
The two levels that were tested were liking and authority as defined by Cialdini (2012). Moreover, it was found that people who score high in agreeableness and extroversion (and those traits match the target audience) are more likely to be persuaded using the principle of liking (Wall, Campbell, Kaye, Levy, & Bhullar, 2019).
Overall, when it comes to authority, the principle was shown to have a greater effect on people high in agreeableness and consciousness and was less associated with other personality traits from the Big 5 model.
The researcher seeks to know what influence the expertise of the guides in the authority condition and the liking effect in the liking condition has on the participants regarding the content and intention to visit the destination.
Manipulation Check
The manipulation check included 3 items that measure character identification such as “I understood Katy’s feelings or emotions” (Igartua & Barrios, 2012) and “I tried to see things from Katy’s point of view” (Igartua & Barrios, 2012) and 2 items on measuring the narrative transportation effect – “I felt very involved mentally during the reading of the story” (Igartua, Rodríguez-Contreras, Ramos, Garay, & Frutos, 2020) and “I wanted to know how the story would end” (Igartua, Rodríguez-Contreras, Ramos, Garay, & Frutos, 2020).
For the persuasion principles, there were two questions in total. One in the authority condition asked if the guides in the story were knowledgeable (nominal question) and one in the liking condition asked to what degree (5-point Likert scale) the participant liked the content.
Measurement Instrument
The measurement instrument consisted of the items measuring the effect, the items referring to the manipulation check, and the psycho-demographic questions. In the following table, you can see the measurement method and resources.
| Item | Measurement | Source |
| I enjoyed reading the material (enjoyment) | Strongly disagree – strongly agree (5-Likert scale) | Lensvelt & Steenbakker (2014) |
| The material helped me get an idea of the experience (informative) | Strongly disagree – strongly agree (5-Likert scale) | Chen, Bell & Tylor (2016) |
| I would read more information like this to gather information (trustworthy) | Strongly disagree – strongly agree (5-Likert scale) | Chen, Bell & Tylor (2016) |
| I felt a connection with the story (enjoyment) | Strongly disagree – strongly agree (5-Likert scale) | Chen, Bell & Tylor (2016) |
| I found the material convincing (trustworthy) | Strongly disagree – strongly agree (5-Likert scale) | Chen, Bell & Tylor (2016) |
| I found the content to be informative (informative) | Strongly disagree – strongly agree (5-Likert scale) | Lensvelt & Steenbakker (2014) |
| If I get the chance to travel, I intend to visit the destination mentioned in the story | Strongly disagree – strongly agree (5-Likert scale) | Chen, Shang, & Li (2014) |
| I tried to see things from Katy’s [character’s name] point of view | Strongly disagree – strongly agree (5-Likert scale) | Igartua & Barrios (2012) |
| I understood Katy’s [character’s name] feelings or emotions | Strongly disagree – strongly agree (5-Likert scale) | Igartua & Barrios (2012) |
| I had the impression of living Katy’s story myself | Strongly disagree – strongly agree (5-Likert scale) | Igartua & Barrios (2012) |
| I felt very involved mentally during the reading of the story | Strongly disagree – strongly agree (5-Likert scale) | Igartua, Rodríguez-Contreras, Ramos, Garay, & Frutos (2020) |
| I wanted to know how the story would end | Strongly disagree – strongly agree (5-Likert scale) | Igartua, Rodríguez-Contreras, Ramos, Garay, & Frutos (2020) |
| According to you, were the guides knowledgeable in the story? (authority condition) | Yes No Unclear | – |
| To what degree did you like the story? (liking condition) | Very much – not at all (5-Likert scale) | – |
| How often do you take part in an outdoor activity? | A few times a week Once a week Once a month Few times a year Once a year I don’t do that | Barcelos, Dantas, & Sénécal (2018) |
| What is your preferred outdoor activity? | Fishing Hiking Kayaking Skiing Swimming SUP Biking Other… (an open answer) | – |
| What is your age? | Open question | |
| What is your gender? | Male Female Divers Other (open answer) Prefer not to say | Hughes, Camden & Yangchen (2016) |
| What is your highest level of education? | High School Bachelor’s Degree Master’s Degree Doctor’s Degree Other | Hughes, Camden & Yangchen (2016) |
Most of the items use a 5-Likert scale measurement (1 – totally disagree, 5- totally agree) and some items offer either a closed or an open answer where the participant can fill in with their own words and numbers for the age and preferred activities. Items 1 to 7 aim to measure the effect on the dependent variables – how informative, enjoyable, and trustworthy the content is and intent to visit based on the manipulations.
The items from 8 to 11 serve as a manipulation check to verify whether the manipulations had an effect on the participant in terms of identification and narrative transportation and whether they were able to recognize the content as enjoyable and the authority of the guides.
The results from the study will inform the strategy in terms of the use of character similarity and narrative message in all visual and textual materials published by the client.
More specifically since the current content objectives focus on brand awareness and purchase phases from the buyer’s funnel, the study will be helpful to determine if such type of content will be effective in convincing the reader of the brand trustworthiness to achieve higher conversion rates and inspire and attract new customers. The participant stories can also increase engagement from loyal customers who follow the brand online if the reader finds them to be enjoyable.
Validity and Reliability
The validity of the concepts and frameworks used in the prototype testing was ensured by using concepts tested in published articles in scientific journals. Similar studies ( (Juan José Igartua, 2019), (Chen, Bell, & Taylor, 2017), (Igartua, Wojcieszak, Cachón-Ramón, & Guerrero-Martín, 2017)) have been performed on narrative persuasion. The measurement items were taken from the research papers where they were tested for internal consistency using Cronbach’s alpha tests.
Additionally, the manipulation check served as a validity measure as it allows the researcher to know if the manipulations that were made had the desired effect on the participant while the repetitions ensured the reliability of the outcomes and minimized the chances to obtain random outcomes that are not based on the manipulations (Hu, Bao, & Wang, 2011).
The ability to reproduce the research is also an important element of a well-designed experiment. A clear explanation of the study steps should guarantee further reproduction by other researchers, so they are able to confirm or counterargue the findings in the current study (Bryman, 2012).
Design
Prototype Groups
The study consists of 4 groups of prototypes each group including a repetition. Based on the levels of the independent variables, character similarity (low/high) and persuasion principles (liking/authority), the following groups are tested – high similarity, authority conditions; high similarity, liking conditions; low similarity, authority conditions; low similarity, liking conditions; The study design is 2 x 2 factorial design involving between subject testing.
The repetition of the conditions was included as a benchmark to make sure that when the narrative changes, the conditions still have the same effect on the audience regardless of the elements of the plot and the characters apart from the intended manipulations.
Added to the male character, the female character across the repetitions was created to mitigate the effects of gender bias effects that were also discovered in research (Roden, Mustafaj, & Saleem, 2021) examining online endorsements and gender.
As a side note: Later on, I realized that I couldn’t achieve the confidence level that I wanted through the use of repetitions as it makes a difference when the participant identifies as male or female, or another gender. I realized that it would have been better to compare against a control group of only male or only female participants to control for the differences.
Test Factors
As it was noticed that identifiable characters are commonly used in the content that is published on social media in comparison to the competitor.
The researcher wanted to know to what extent the use of characters is beneficial and what effects those characters can have on the persona when it comes to gaining trust and delivering content that engages the community. The focus is on building trust for new customers, so they choose the company for their next hiking trip.
The second factor is the persuasion principles that were selected based on the insights of the content analysis and the persona. Also, it was suggested to the client to work on building the brand as an authority, so the current research aims to test liking versus authority.
And discover if there is any significant difference between the use of liking by the client and the persuasion principle of authority that the persona seeks when making a decision.
Based on the study results, content can be more carefully created including identifiable character and persuasion principles when creating content for the website and social media. Moreover, a study (Higgins & Rholes, 1987) has found that when the reader likes the character in a narrative, that has an impact on lasting memory recall and positive evaluation of the character.
That means that the reader will be able to associate the brand with the character and could improve brand recall and attitudes as part of the content strategy.
| Factors | Character Similarity |
|---|---|
| Persuasion principle (liking vs authority) | Character similarity (high vs low) |
| High – Liking | Low – Liking |
| High – Authority | Low – Authority |
Additionally, the findings will be of importance to the overall communication goals of the company as including visuals containing characters similar to the ones the brand wants to attract is part of the overall branding strategy (Keller, 1993). That is why it is so important to figure out what role the factors play in the content message.
Procedure
The participants were randomized across the groups to control for lurking variables.
After being redirected to the group condition, the respondents were first introduced to the story of the character and their visual features, then they were invited to answer the questions measuring the effect and the manipulation check.
Next, they were introduced to the second character and invited to answer the same questions. After that, in a new section, the final demographic and psychoanalytical items were provided.
Distribution
To ensure the closest fit to the target persona, the same outdoor and adventure Facebook groups from the previous research have been used as well as some new groups where people were interested in hiking, SUP, mountaineering, and climbing. A post inviting the group members of each Facebook group was created with a link leading to the survey.
Methods and Results
The data for the dependent variables and manipulation was coded with values from 1 to 5 and the data for the psychographics and demographics was given other values corresponding to the measurement.
Then questions 1,3, and 4 were added together to form one solution for the dependent variable ‘enjoyment’, 2 and 6 for ‘informative content’, 8,9 and 10 from the manipulation check for ‘viewer identification’, and 11 and 12 for ‘narrative transportation. Those combinations were performed for all groups.
Following, an SPSS input file was created for each group with 8 dependent variables measuring perceived enjoyment, content information, intent to visit, and trustworthiness of the content. Since each group had 2 prototypes, 4 dependent variables were analyzed for the first prototype and 3 items were analyzed for the manipulation check. In the end, the psychographic and demographic questions were analyzed using frequency tables and percentages.
Firstly, the dependent variables were analyzed for a central tendency using the mean, mode, and median of the frequency tests. Overall, the median and the mode were identical numbers (between 3 and 4 points) while the mean was close (ranging between 2.50 to 3.50 points on the 5-Lickert scale). Those results indicated the normal distribution of the data points for the two-way ANOVA analysis (Bryman, 2012).
Before analyzing the effect though, the means were analyzed for the dependent variables and the manipulation check to observe a pattern between the different groups and conditions. Based on those results, the ANOVA analysis was performed for a few of the results where there was a difference for the same variables across the conditions.
The manipulation check was successful when 50% of all respondents gave a positive answer to the two combined questions for viewer identification on a scale up to 5, so 4 and 5 were considered positive, and the other two questions, nominal items, to measure the likeability of the story for the liking conditions and whether the expertise of the guides in the authority condition. In the following table, you can see an overview of the variables and tests being performed.
| Independent variables | Measurement | Analysis | Expected effect |
| Character similarity | Nominal | Two-Way ANOVA | Change in the perception of the story and intent to visit based on the conditions |
| Persuasion | Nominal | Two-Way ANOVA | Change in the perception of the story and intent to visit based on the conditions |
| Dependent variables | |||
| Enjoyment | Scale | Means, Frequencies, Two-Way ANOVA | If the story was enjoyable based on the conditions |
| Repetition Enjoyment | Scale | Means, Frequencies, Two-Way ANOVA | If the story was enjoyable based on the conditions |
| Information | Scale | Means, Frequencies, Two-Way ANOVA | If the story was informative |
| Repetition Information | Scale | Means, Frequencies, Two-Way ANOVA | If the story was informative |
| Trustworthiness | Scale | Means, Frequencies, Two-Way ANOVA | If the story was convincing |
| Repetition trustworthiness | Scale | Means, Frequencies, Two-Way ANOVA | If the story was convincing |
| Intent to visit | Scale | Means, Frequencies, Two-Way ANOVA | If the conditions had an impact on the intent to visit the destination |
| Repetition intent to visit | Scale | Means, Frequencies, Two-Way ANOVA | If the conditions had an impact on the intent to visit the destination |
| Manipulation check | |||
| Degree of likeability of the story | Scale | Means, Frequencies | The respondent liked the story |
| Degree of expertise in the story | Nominal | Frequencies | The respondent is aware of the expertise |
| Viewer identification | Scale | Means, Frequencies | The respondent can see and understand the character’s point of view and emotions |
| Narrative transportation | Scale | Means, Frequencies | The response is mentally involved in the story |
| Psychographics | |||
| How often do you take part in an outdoor activity? | Ordinal | Frequencies | It measures the respondent’s fit with the persona |
| Preferred adventure activity | Nominal | Frequencies | It measures persona fit for outdoor activities |
| Demographics | |||
| Age | Nominal | Average age per group | Measures the average age per group to see if it is fit with the persona |
| Gender | Nominal | Frequencies | Measures sample gender per group |
| Education | Ordinal | Frequencies | Measures sample education per group to make sure respondents fit with the target audience |
Sample Description
Overall, the sample fits with the persona’s level of physical activity, education, and preferred adventure activities. However, the age does not correspond with the target persona. It was slightly higher than the desired age range.
Manipulation Check
The mean scores were evaluated to make a judgment about the manipulation check. A result that is 3.5 or above was considered positive, and the manipulation was successful. Percentages were used in some cases where means could not be calculated. If the result was over 50%, it was considered successful.
The manipulation check was successful for the female prototypes in the low similarity conditions except for liking (low similarity/liking group). Female prototypes in the high similarity conditions failed the manipulation check for viewer identification and narrative transportation, but the authority level was recognized in the high and low similarity conditions.
The manipulation check for the male characters across all groups failed for viewer identification and narrative transportation and got neutral scores for liking and authority in both high and low-similarity groups.
In brief, it was concluded that the manipulation was valid in the groups where the participants could relate to the character. And it was noted that the female characters in the low similarity group performed better when the age of the identifiable characters was much higher than the one in the high similarity groups.
As a side note: That was contrary to what I initially expected. However, I couldn’t control for age in this study due to time and budget constraints.
Besides, in the liking conditions, it was found that the participants generally didn’t enjoy the story as much and gave a neutral score.
That seems consistent across all conditions, so the story in the liking conditions was not well designed and accurate judgments cannot be made based on that manipulation.
In comparison, the authority manipulation was understood by the participants across the female prototypes in both low and high similarity conditions. That makes the manipulation of authority valid. Accurate conclusions cannot be made about the male prototypes.
Descriptive Statistics and ANOVA Results
The dependent variables for the male character prototypes did not show any significant results. Only in the low similarity/authority condition did the participants find the content to be informative for the male character.
Though, the result is not consistent across groups for all male characters and stories, so it was considered not reliable.
The results of the information effect from the two-way ANOVA were insignificant as the p-value for the factors and their interaction was over 0.05 alpha.
And the results are not surprising as the participants in the test could not relate to the male characters in the prototypes as shown in the manipulation check.
The results of the female character prototypes are a bit more promising. The mean results showed that in 3 out of 4 prototypes the content was found to be informative and in the two low similarity groups the content was found to be trustworthy in both low/liking & authority groups and enjoyable in the low liking/authority condition.
Then, the two-way ANOVA tests were run for the dependent variables that had positive mean scores (greater than or equal to 3.50 out of 5). In all groups, the persuasion principle didn’t have a significant effect on the dependent variables. However, the character similarity condition was significant for the variables of trustworthiness.
Discussion
The clear winner among the groups is the low similarity/authority. Those results seem a bit surprising since the participants could identify with the character in the story despite the attempt of the researcher to create dissimilarity in low conditions.
In the introduction of the character was stated that the character prefers leisure activities that are calmer and is not a risk-taker in comparison to the active lifestyle of the participant. So, the character in the story was generally not experienced with outdoor activities in contrast to the target audience.
So, it seems like the visual of the character in the prototype including the age of the character, the outdoor background on the image, and sports clothing had a greater effect on creating similarity rather than the dissimilarity that was intended through the story.
Even though the liking principle failed to provide significant results, it was found that the participants enjoyed the content because of the character similarity manipulation. Those results are in line with previous research on relatedness and its effects on liking (Alves, Koch, & Unkelbach, 2016). More research is needed to make a reliable conclusion about the enjoyment effect, as the second low/liking group didn’t show that effect despite the successful character identification and narrative transportation.
When it comes to the persuasion principles, the results do not clearly say which one performed better as that was the goal set earlier in the research.
From the mean scores of the dependent variables, it seems like the groups in the authority condition have slightly better scores and advantages for both male and female characters in comparison to the liking conditions. And although the manipulation was successful for the authority principle for female characters, the results are insufficient to make reliable conclusions and the principles didn’t influence the results significantly.
In conclusion, the participants found the content to be informative, trustworthy, and enjoyable.
ANOVA results show that the effect of trustworthiness and enjoyment is due to character similarity manipulations.
The research shows that using identifiable characters in visuals and text is beneficial as it increases the credibility of the brand. The client is safe to use the format to improve the credibility of the brand as it was suggested.
Limitation and Recommendations
- Authority principle and trustworthiness
One particular study (Griffin, Wiley, & Thiede, 2008) argues about the effects of “working memory and self-interpretation as a self-explanation reading strategy” on improved metacomprehension of texts. Thus, to improve the understanding of the narrative, it should include real-life examples of situations in which the character needed the expertise of the guides to overcome an obstacle. As it was the case in the second story where the character shares how they would not have made it on their own and how they managed to keep warm in below-zero temperatures with the instructions of the guides. More research could be taken into consideration about the message framing and comprehension to ensure the best outcomes from the narrative.
- Liking principle
Including more positive feelings about the experience and the fun, the participant has on the hike could improve the perceived enjoyability of the stories. To improve the research further, the appeal of the story and the character should be measured together as part of the liking principle. Currently, the research only measured the appeal of the story, however as research shows when participants found the character to be like them, they like the content (visual + text) in general despite the story not being so interesting. That was a limitation of the current research.
- Design
The expectations for age and gender in the sample could not be met in the current study and perhaps younger audiences and participants of different genders can show different results. That also made it impossible for the repetitions to be compared as intended and get a greater level of confidence in the obtained results.
So, to make the results more reliable for comparison, the design of the study should test only male or only female participants and compare them to a control group.
That would give more reliable and objective results for comparison of both female and male characters and their effect on the targeted audience (male or female).
The one-size-fits-all in the current design was limited. Although some research suggests that demographic criteria don’t influence results (Igartua, Rodríguez-Contreras, Ramos, Garay, & Frutos, 2020), it was found that it is a limitation when it comes to identification and narrative transportation.
Behind-the-Scenes ↗
See how I ran the studies to collect the primary data that is used for decision-making in this work.
Inspiration ↗
A few analyses inspired me to think about the message and innovate.
About & Contact ↗
Get to know my experience and contact me.