1. Introduction

In recent years, especially after the COVID-19 pandemic in 2020, the trend of integration between movies and streaming media has become stronger [1-2]. Research has found that movie commentary video has a positive effect of 3% on the playing volume of the original movie [3-4]. It can be seen that movie commentary videos have great appeal to netizens in terms of content, and the development of movie commentary videos can also be conducive to the development of the film industry.

As of December 2023, the total number of short video accounts in China has reached 1.55 billion, and movie commentary has gradually become a popular content in short videos. For video creators, while it is important to choose popular content, it is equally valuable to determine the properties of the video itself. Compared with determining the content of the video, determining the properties of the video itself is cheaper and has higher utility. Video creators can improve the popularity of the video by changing the properties of the video itself. With regard to the popularity, influence and communication effect of videos, existing literatures have conducted researches on various types of short videos such as popular short videos in libraries [5], short sports videos [6], interactive videos for science popularization [7], etc. By studying heuristic cue such as publisher characteristics and title information and systematic cue such as content theme and video duration, it is found that video description length and title detail have a significant positive influence on video popularity [8], but due to the different types of videos studied, the final results are somewhat different. For example, for official videos such as short videos of People's Daily's Douyin number, popular short videos in libraries, and government affairs videos, viewers will prefer specific content [9]. The above literatures provide useful references for understanding the impact of the video's own attributes on the popularity of videos. However, most of the relevant literature do not analyze the movie narration videos. How should the movie narration videos adjust their own attributes to achieve a higher number of likes? In this respect, there is no certain theoretical basis.

In view of this, this study intends to analyze the current movie narration videos on the Douyin platform, summarize the characteristics of these short videos, and finally put forward suggestions according to the research conclusions, so as to provide experience reference for short video creators to better make movie narration videos and help the dissemination of movie narration videos.

2. Literature Review

2.1. Heuristic Systematic Model

Heuristic-systematic Model (HSM) is a kind of dual processing theoretical model proposed by psychologist Chaiken [10] in 1980 to explain individual information behavior process, aiming to describe how people combine Heuristic and Systematic strategies when processing information. Heuristic behavior refers to the tendency of individuals to rely on simple rules, experience or external cues to make judgments or decisions when processing information. Heuristic cues are non-content and situational cues contained in the information itself, such as title information, publisher characteristics, cover type, etc. Systematic behavior means that when individuals process information, they will invest more cognitive resources to conduct in-depth and detailed analysis and evaluation of information content, which requires users to think in more detail. Systematic cues are the content characteristics of information itself, such as video duration, content theme, narrative type, interaction motivation, interaction frequency, etc.

HSM is already being used by researchers. Zhou Tao et al. [11] collected video data from website B, and took the features of publisher, title and cover as heuristic clues, and the content theme, video duration and video introduction as systematic clues to empirically explore the effect of HSM on the dissemination of knowledge videos. Wang Yan et al. [12] took cover type and title information as heuristic clues, and interaction frequency, interaction motivation and interaction frequency as systematic clues to study the dissemination effect of knowledge-based videos.

Based on the above literature, HSM can provide theoretical support for users' information decision analysis of movie narration videos, but it has not been applied to the analysis of influencing factors of the number of likes of movie narration videos at present. Although the number of labels of movie narration videos is a specific value, it is more indirectly to guide user behavior by affecting user search and recommendation experience. Based on the number of tags and content, users may determine the theme, style, or appeal of a video and decide whether to click to watch it. This judgment relies on the user's personal experience and intuition, so the number of labels can be regarded as a kind of heuristic cue. The video duration as a specific value, users can directly obtain and judge the length of the video content, so as to decide whether to watch or continue to watch. Therefore, the video duration is consistent with the characteristics of systematic clues. Therefore, this study takes the number of tags as a heuristic cue and the length of video as a systematic cue to study its influence on the number of likes of movie narration videos.

3. Research hypotheses and models

Based on HSM, this article mainly examines the impact of video duration and title length in title information on video like volume. Therefore, this paper takes the duration of the video and the number of tags of the title as independent variables, and takes the number of video comments, the number of video collections, the number of video shares, the number of fans of the publisher, the number of likes of the publisher, etc. as control variables, to discuss the factors affecting the amount of video likes as a whole, and build a model of factors affecting the amount of likes of TikTok movie commentary videos (as shown in Figure 1).

Figure 1: Influential factors model of TikTok movie commentary video rating.

3.1. The impact of short video duration on video like volume

The biggest feature of short video is the short time, usually between a few seconds and a few minutes. This time limit requires that the content must quickly capture the audience's attention and deliver the core message, conveying the main content in the simplest way through a refined narrative and intuitive presentation. But for movie commentary videos, people have a higher demand for video duration. Scholar Ma Tengfei pointed out that short video commentary can decompose a 90-120 minute movie in 3-5 minutes or even less, but the content is relatively simple [13]. Xu Hailong also found that the total length of movie commentary videos is mostly 6-9 minutes or 12-18 minutes [14]. The popularity of super-long videos such as commentary videos and collection videos also proves that people have higher requirements for short video duration. For a brief introduction or quick review, short videos may be more effective. However, for some complex movies or content that requires in-depth analysis, longer videos may be more appropriate. For movie commentary videos, the audience may have higher requirements for the length of time. Based on the current duration data of movie commentary videos on TikTok, viewers may not like too long or too short movie commentary. Based on this, the following hypothesis is proposed:

H1: There is an inverted U-shaped relationship between the duration of a video and the number of likes it receives

3.2. The impact of title information on video like volume

The title information has the nature of social communication [15]. Scholars found that the tag-based video recommendation mode of Station B can attract users to watch videos and strengthen the role of heuristic cues [16]. TikTok also uses a similar algorithmic recommendation mechanism to push relevant content based on users' interests and behaviors. The label in the title is an important basis for the algorithm to recognize video content, classify and push. The more and more accurate the tags, the more the system can push the video to the users who are interested in its content, thus increasing the exposure and click rate of the video. For movie narration videos, the title contains popular keywords related to the movie, which helps the video get higher exposure rate in search engines and platform recommendation systems, and the increase in exposure rate means that more viewers can see and like the video. At the same time, videos that capture the current hot topic or movie release time and reflect those hot elements in the title will get more likes. Based on this, the research hypothesis is proposed:H2: The number of tags in the title has a significant positive impact on the number of likes in the video

4. Data

According to the analysis data of the development trend of the new media industry [17], by September 2023, the number of de duplication active users of the five typical new media platforms, Tiktok, Kwai, Xiaohongshu, Bilibili and Weibo, has reached 1.088 billion, with a penetration rate of 88.9%. Tiktok is a typical representative in the field of short video, and has been developing rapidly since its establishment in 2015. Now it has become one of the most influential short video platforms in China.

This study selects the film commentary videos of the Tiktok platform for analysis: First, Tiktok is the most popular short video platform in China at present, with more than 640 million users since 2020. It is also the platform with the largest number of film commentary videos, covering a variety of video types, themes and styles. This richness provides sufficient sample resources for the study, which can comprehensively analyze the likes of different types of film commentary videos. Secondly, the Tiktok platform is highly interactive, and users can express their views and feelings on movie commentary videos through likes, comments, sharing and other ways. These interactive data provide important basis for studying the liking factors of movie commentary videos. By analyzing the content of users' comments, we can gain a deeper understanding of the specific reasons why they like, whether it is due to the fun and professionalism of the commentary, or the editing effect of the video. At the same time, users' sharing behavior can also reflect the dissemination and influence of videos, further revealing the relationship between liking factors and video dissemination.

Taking comedy movies and suspense movies as examples, this study collected data on movie commentary videos from Tiktok platform from January 2020 to August 2024 (for example, video duration, ratings, number of author fans, number of authors' ratings, etc.) according to the keywords "comedy movie commentary", "comedy movie recommendation", "suspense movie commentary", and "suspense movie recommendation". The sample includes 767 comedy movie commentary videos and 976 suspense movie commentary videos. These two types of movies have a wide audience, distinct genre characteristics, and completely different styles. By comparing and analyzing the liking factors of these two types of movie commentary videos, valuable references can be provided for movie commentary creators.

5. Empirical analysis

5.1. Variable measurement

The likes in the dependent variable movie commentary video can be directly obtained from the data set on TikTok. The independent variables include the number of labels in heuristic cues and the duration of labeled videos in systematic cues. The number of labels is measured by regular expressions. The video duration is obtained from the dataset. The specific variable description declarations are shown in Table 1.

Table 1: Declaration of variable descriptions.

5.2. Descriptive Analysis

Descriptive statistics were conducted for each variable to understand its data characteristics, as shown in Table 2 and Table 3 for details. The duration of TikTok’s movie commentary video spans from a few seconds to several hours, but the duration is mainly within five minutes. The title length is concentrated in 10-30 words, and most titles have 1-5 tags (Videolablenum). Conduct correlation analysis on control variables, independent variables, and dependent variables to explore the degree of closeness of the impact relationship between each variable. The number and percentage of cases in different categories are shown in Table 4 and Table 5.

Table 2: Summary statistics of suspense movie explainer videos.

Table 3: Summary statistics of comedy movie commentary videos.

Table 4: The number and proportion of variables in different categories of suspense movie explainer videos.

Table 5: The number and proportion of variables in different categories of comedy movie commentary videos.

5.3. Multiple Linear Regression Analysis

The dependent variable of regression analysis is the number of likes of the movie narration video. In this paper, linear regression analysis is adopted to verify the hypothesis, and correlation analysis is carried out on the control variables, independent variables and dependent variables to explore the closeness of the influence relationship between each variable. The sample regression results are shown in Table 6.

Likes=β₀+β₁Duration)²+β₂Duration+ε_i(1)

Table 6: Estimation Results of the Model of Likes.

Note:^*p<.05. ; ^**p<.01. ; ^***p<.001.

Based on the regression results of models one and two, explain and analyze from the following two aspects:

(1) The regression coefficients of video Duration are all positive (35.811, 23.046, respectively), and the significance is less than 0.05 (0.000, 0.045, respectively), which means that video duration has a significant positive impact on the number of video likes. The regression coefficients are all negative (-0.0098 and -0.0028 respectively), indicating that there is a direct inverse U-shaped relationship between video duration and video likes. With the increase of video duration, video likes first increase and then decrease, that is, there is a maximum point for video likes. The duration of videos that maximized the number of likes in suspense movie narration videos and comedy movie narration videos is 0.56 hours and 1.08 hours, respectively, assuming H1 is valid.

(2) The number of labels (Vln) presented different results in the two models. In the suspenseful movie commentary video, the regression coefficient of Vln was negative, with a significance greater than 0.05 (0.573); In comedy movie commentary videos, the regression coefficient of Vln is positive, with a significance of less than 0.05 (0.039). This means that in suspense movie commentary videos, the number of tags does not have an impact on the number of video likes, while in comedy movie commentary videos, the number of tags has a significant positive impact on the number of video likes. This may be because the two types of movies have different content characteristics. Suspense movies are characterized by complex plots and frequent suspense, and audiences are more inclined to be attracted by the content itself rather than too many tags. Too many and disorderly tags may distract the audience's attention and even have negative effects; Comedy movies are characterized by humor and lightheartedness, and audiences are more easily influenced by the relaxed and pleasant atmosphere while watching or listening to commentary. In this case, more tags mean more potential audiences, and the number of tags may become a factor affecting video exposure and likes. Thus, the H2 part is established.

5.4. Research Conclusion

Based on the heuristic systematic model, this study examines the factors influencing the likes of Tiktok movie commentary video from the attributes of the video itself in the heuristic clues and systematic clues, and conducts regression analysis on the likes of the video from the video duration, title length, and the number of tags in the title, respectively. In terms of the content of movie commentary video, two types of suspense movies and comedy movies are selected to build Model 1 and Model 2.

The analysis results indicate that: 1. Video duration can have an impact on video like volume. A 0.56-hour suspense movie commentary video and a 1.08-hour comedy movie commentary video are more likely to attract audiences and receive higher likes. Reasonably planning the time of a video based on its content can maintain audience interest and effectively convey information, which is similar to the research conclusions of scholars such as Ji Haixiang and Ren Nan. However, this article has a clearer calculation of the reasonable time for suspense movie commentary videos and comedy movie commentary videos, providing accurate time recommendations for these two types of movie commentary videos, which can directly help creators plan their time reasonably when making videos. The relationship between the number of tags and the number of likes varies depending on the content type. In suspenseful movie commentary videos, the number of tags is not a key factor; In comedy movie commentary videos, increasing the number of tags appropriately may help increase the exposure and likes of the video. However, regardless of the type of video, attention should be paid to the quality and relevance of the tags, avoiding misuse and abuse of tags, to ensure that the target audience can be accurately guided to discover the video and generate positive interactive behavior.

6. Conclusion

Based on HSM, this article studies the factors affecting the number of likes in movie commentary videos. Through analyzing 767 comedy movie commentary videos and 976 suspense movie commentary videos collected, the research results show that both video duration and number of tags have an impact on the number of likes. A 0.56-hour suspense movie commentary video and a 1.08-hour comedy movie commentary video are more likely to receive higher likes, and the number of tags in comedy movie commentary videos has a significant positive impact on video likes.

The theoretical contributions of this study include: 1. Based on the HSM model, this paper investigates the influencing factors of likes in movie commentary videos, including heuristic clues and systematic clues. Heuristic cues include title features, while systematic cues include video duration. The research findings reveal the impact of title features and video duration on movie commentary videos; This study found a U-shaped relationship between the number of likes and video duration in movie commentary videos, equivalent to videos that are too long or too short. Film commentary videos of appropriate length are more likely to be welcomed by audiences; This study found that the relationship between the number of tags and the number of likes varies depending on the content type. The impact of the number of tags on the dissemination of different types of movie commentary videos varies. Therefore, the research findings of this article contribute to the dissemination of movie commentary videos.

This study expands the research in the field of video duration and number of tags, clearly indicating the optimal duration range for suspense movie commentary videos and comedy movie commentary videos and revealing the differences in the role of tag number in different types of movie commentary videos. Creators can adjust tag strategies according to the type of video content to achieve the best results.

This research also has some limitations. The research only selects the movie commentary videos on Tiktok for analysis and exploration; Neglecting the subjective emotions of video users and their impact on the number of likes, there is a lack of qualitative analysis of factors such as text content and language style in comments and bullet comments; It also ignores the interaction between video comment volume, collection volume, and video like volume. This may lead to research bias, so future research can further explore and improve the above-mentioned shortcomings.