The Issue of Scales for Measuring Parasocial Interaction and Parasocial Relationship between Followers and Influencers on Social Media

Abstract

Social media has introduced influencers and influencer marketing (IM), which is becoming increasingly popular among marketers. The effectiveness of IM is significantly influenced by parasocial interactions (PSI) and parasocial relationships (PSR) that develop between followers and influencers. Historically, a variety of scales have been used to measure PSI and PSR, raising concerns about the comparability of research results. Given the recent growth of IM and the importance of PSI and PSR within it, we aimed to examine how these constructs are measured in IM. We conducted a literature review, analyzing 72 studies focused on the empirical measurement of PSI and PSR. We found a significant heterogeneity in the measurement of PSI and PSR, identifying 26 scales for PSI and 29 scales for PSR, with two scales being used for measuring both PSI and PSR. This high degree of variability among scales that are supposed to measure the same constructs raises questions about the comparability of the results. We identified a critical need for clearer conceptual and empirical differentiation between PSI and PSR, which should be reflected in the development of measurement instruments. It is essential to develop reliable and valid scales that account for these differences and distinctly measure PSI and PSR in IM.

1. Introduction

1.1. Research Background

With the emergence of social media in the first decade of the 21st century, individuals who have become more recognizable, visible, and influential on these platforms have also appeared—we call them social media influencers. Dhanesh and Duthler [1] (p. 1) describe influencers as follows: “Social media influencers, a type of opinion leader, engage in self-presentation on social media, accomplished through the creation of their online images employing rich multimodal narration of their personal, everyday lives, and leverage these images to attract attention, and a large number of followers”.

Due to their public image [1], which they have created on social media, the content they publish [2,3], their expertise [3], interesting lifestyle [2], or similarity to their followers [4], they have become very intriguing to other social media users—the followers.

Influencers’ ability to attract and maintain the attention of thousands to millions of followers [5], who are also consumers, has drawn the attention of marketers, especially at a time when traditional forms of advertising are losing their effectiveness [6,7].

This has led to the development of a new form of marketing called influencer marketing. Influencer marketing is based on the collaboration between influencers and brands, where influencers include the promotion of products, services, or viewpoints in their content in exchange for payment. The goal is to gain a favorable attitude from followers towards the promoted product, service, or viewpoint [8,9,10].

Due to its effectiveness, more companies are adopting influencer marketing [8], contributing to the growth of the global influencer marketing market, which was valued at $9.7 billion in 2020 and is expected to reach $22.2 billion by 2025 [11].

Numerous studies agree that interaction and relationship between influencers and followers are among the most critical factors for the effectiveness of influencer marketing [1,12,13,14,15]. Previous research has explained this phenomenon using the concepts of PSI and PSR [16], which have been measured using specifically created scales.

Until 2016, the most established scale for measuring PSI was the PSI Scale [17], with 20 items, and its shorter version with 10 items was published two years later [18]. Despite its popularity, the scale has been heavily criticized for primarily measuring PSR instead of PSI, leading to confusion in the field due to the lack of distinction between the constructs. In response to these criticisms, several other scales have been developed over the years, including the Audience Persona Scale [19], the PSI-Process Scale [20], the Multiple PSR Scale [21], and the Experience of Parasocial Interaction—EPSI Scale [22], each operationalizing PSI and PSR differently.

Liebers and Schramm [23] note that the measurement of PSI and PSR from 1956 to 2016 was extremely diverse, with various scales being used, making it difficult to compare results. Past measurements of PSI and PSR have also faced significant criticism, as many scales were found to measure constructs other than PSI or PSR and often lacked construct validity [24].

Given that calls for more unified measurement of PSI and PSR, considering construct validation, have been made for almost a decade, we aim to investigate whether these initiatives have been implemented in recent studies, particularly those in the field of influencer marketing.

1.2. Research Purpose

Liebers and Schramm [23] highlight the extreme heterogeneity in methods used to measure PSI and PSR in the past, while Dibble, Hartman, and Rosaen [24] are critical of the scales used for measuring PSI and PSR. The PSI scale, which was the most frequently used scale from its inception until 2016 [23], lacks content validity, as it is inconsistent with the original definition of PSI provided by Horton and Wohl, as well as with subsequent updates [24]. Furthermore, it includes items that more accurately measure PSR rather than PSI [24].

Dibble, Hartmann, and Rosaen [24] also criticize other scales used later, arguing that they do not define the constructs they aim to measure with sufficient validity and precision and have not undergone adequate construct validation testing. They emphasize that PSI and PSR have undergone significant theoretical and conceptual refinements since their inception, but the measurement tools have not reflected these developments.

Given that studies highlighting the extreme diversity in PSI and PSR measurement [23] and pointing out conceptual and empirical confusion in measurement methods [24] were published before the rise of influencer marketing, we are interested in understanding how PSI and PSR are measured in influencer marketing. Considering the critical role of PSI and PSR in the effectiveness of influencer marketing, it is essential to understand how PSI and PSR are measured in this context, especially given the empirical and conceptual ambiguities that existed before the rise of influencer marketing.

Therefore, we aim to examine how the challenges regarding empirical differentiation are addressed in the measurement of PSI and PSR in influencer marketing.

The purpose of this study is to determine the following:

• Which scales are used to measure PSI and PSR in influencer marketing?
• Are there predominant scales for measuring PSI and PSR in influencer marketing?
• Do the measurement instruments for PSI and PSR in influencer marketing account for the distinction between PSI and PSR?

2. Theoretical Framework

2.1. Parasocial Interaction and Parasocial Relationship

The concepts of PSI and PSR were first mentioned in 1956 in an essay by sociologist Richard Wohl and anthropologist Donald Horton titled “Mass Communication and Para-Social Interaction: Observations on Intimacy at a Distance”, published in the journal Psychiatry [16].

Horton and Wohl [16] defined PSI as a simulated form of conversational interaction that viewers experience while engaging with media characters, such as watching TV content featuring a television persona. It is a one-sided, mediated form of social interaction between the viewer and the media character [23]. The viewer perceives it as an illusion of a “face-to-face” interaction with the performer.

Viewers experience PSI as an intimate, reciprocal personal interaction, even though they are aware that it is merely an illusion [24]. Thus, PSI is similar to personal interaction between two individuals in real life, with the crucial difference being the lack of reciprocity and two-way communication—PSI is entirely one-sided.

PSI is triggered and managed by the media persona or TV character during their performance, specifically during the viewer’s exposure. This interaction begins with behaviors that indicate the character’s awareness of the audience, such as physical and verbal engagement, which create the illusion that the TV character is directly addressing the viewer. Experiments have demonstrated that this engagement includes actions like looking straight into the camera, mimicking real-life eye contact, turning towards the camera, and directly addressing the audience verbally [16,22,24,25,26]. The behavioral patterns and tactics used in real interpersonal interactions also contribute to the formation of PSI [20]. According to Horton and Wohl [16], physical and verbal engagement with the audience is crucial for establishing PSI.

In the article “Mass Communication and Para-Social Interaction: Observations on Intimacy at a Distance” [16], Horton and Wohl also discussed PSRs, describing them as long-term, illusionary relationships and one-sided intimacy that viewers develop with media personas from a distance. While PSR begins with PSI, it persists even when the viewer is no longer directly exposed to the media character and extends beyond a single medium. The key aspect of PSR is its duration, which significantly outlasts the period of exposure. According to Horton and Wohl [16], PSR results from repeated exposure to the media character, whereas PSI involves mutual awareness that only lasts during the exposure to the media persona.

Both PSI and PSR involve illusory interactions or relationships experienced unilaterally by viewers or fans of media characters. Initially, the focus was on TV casters, but it later expanded to include media personas and now encompasses influencers. The crucial difference between PSI and PSR is the duration. PSI occurs only during moments when the person experiencing PSI is directly exposed to the media character, interacting with them. In contrast, PSR is a longer-lasting feeling that often begins with PSI but continues and is felt even when the viewer or person experiencing PSR is no longer directly exposed to the media character.

2.2. Measurement of Parasocial Interaction and Parasocial Relationship

PSI and PSR received little attention until 1985, being mentioned in fewer than ten scientific articles [23]. The topic gained significant momentum after 1985, when Rubin, Perse, and Powell [17] published the PSI Scale. This scale enabled the theoretical concepts to be translated into numerous quantitative studies, allowing for the measurement and examination of the effects of PSI and PSR. The scale, in its original and various adapted forms, became the standard for measuring the parasocial phenomenon in subsequent periods.

While the PSI Scale [17] revitalized the topic, it also introduced some confusion. The individual items in the scale did not distinguish between PSI and PSR [24]. Furthermore, the definition of PSI provided by Rubin, Perse, and Powell [17] in their article—described as the viewer’s interpersonal involvement with the media persona, including seeking guidance from media characters, viewing them as friends, imagining being part of their social world, and wanting to meet them in real life—did not distinguish between PSI and PSR. Rubin, Perse, and Powell [17], following the uses-and-gratifications theory, regarded PSI and PSR as forms of social involvement or gratification that users seek from media [24]. While Horton and Wohl [16] explicitly stated that PSI exists only at the time of the viewing experience, Rubin, Perse, and Powell [17] combined PSI and PSR into a broader concept of social involvement, encompassing interaction, identification, and long-term identification with television characters, contrary to Horton and Wohl’s original distinction.

The PSI Scale and its items convey feelings of long-term friendship, long-term affinity, or general positivity towards the character. Rather than focusing on the simulated conversational exchange or the one-sided sense of mutual awareness that occurs only during viewing, as described by Horton and Wohl [16], the PSI Scale measures these constructs differently [24]. Dibble, Hartmann, and Rosaen [24] conclude that the PSI Scale expresses more positive affection towards the media persona rather than the mutual awareness critical for PSI, making it more suitable for measuring long-term PSR. Tukachinsky [21] criticizes the PSI Scale for capturing PSR in only half of its items, with the remaining items measuring perceived realism, affinity, and identification.

As a result, after 1985, there was considerable ambiguity in precisely defining PSI and PSR. The term PSI was often used to describe what we now understand as PSR, and sometimes the terms were used interchangeably or as synonyms [20,23]. At the turn of the millennium and beyond, several researchers called for clearer definitions and distinctions between PSI and PSR [21,22,26,27]. These calls for distinction and more precise definitions led to a clearer differentiation between PSI and PSR, primarily focusing on the duration of each parasocial phenomenon. Consequently, new scales for measuring PSI and PSR were developed.

The Audience Persona Scale [19], with its 22 items, measured positive long-term involvement with a favorite media character. It consists of four subscales: Identification with Favorite Character, Interest in Favorite Character, Group Identification/Interaction, and Favorite Character’s Problem-Solving Ability. Schramm and Hartmann [20] evaluated the Audience Persona Scale as suitable for measuring PSR with a favorite character but unsuitable for measuring parasocial engagement with a disliked character and for PSI.

Schramm and Hartmann [20] were critical of previously published scales, asserting that they all measure PSR instead of PSI. Consequently, they developed the PSI-Process Scale [20], which measures PSI or users’ parasocial processing during exposure. The authors define PSI as parasocial processing, encompassing users’ cognitive, affective, and behavioral responses to media characters.

Tukachinsky’s [21] 24-item Multiple-PSR Scale defines PSR as para-friendship, imagined support and intimacy between the media persona and the viewer, and para-love, where the viewer perceives the media persona as physically attractive and desires the media character. Tukachinsky [21] notes that, similar to social relationships, PSR is not uniform and varies in both intensity and quality (friendship vs. love). Tukachinsky describes PSR as a complex, multi-dimensional phenomenon that cannot be entirely measured by the PSI Scale.

In response to the key shortcoming that the PSI Scale does not measure PSI as originally defined by Horton and Wohl [16], Hartmann and Goldhorn [22] created the Experience of Parasocial Interaction—EPSI Scale. This scale focuses on reciprocal awareness, attention, and adjustment with the media performer. Dibble, Hartmann, and Rosaen [24] claim this scale to be more suitable for measuring PSI than the PSI Scale.

The considerable number of different scales for measuring PSI and PSR, which have treated PSI and PSR differently both as conceptual constructs and empirically, has resulted in great diversity in measuring PSI and PSR over the 60-year period from 1956 to 2016 [23].

Liebers and Schramm [23] found that between 1956 and 2016, the parasocial phenomenon, which includes various parasocial concepts (e.g., PSI, PSR, parasocial break-ups, etc.), was measured in different ways. It was sometimes measured using one of the established scales, such as the short and long versions of the PSI Scale [17], the Audience-Persona Interaction Scale [19], and the PSI-Process Scale [20]. Others approached measurement by combining several different established scales, while some added items exclusively developed for their study to the established scales, and still others used individually developed scales unique to their study.

Liebers and Schramm [23] point out that the heterogeneity in the methods of measuring parasocial phenomena stems from the lack of differentiation between individual parasocial phenomena, making it difficult to compare different studies. They emphasize the need for standardized and more differentiated measurement instruments for parasocial phenomena, which would ensure the approximate comparability of empirical findings in parasocial research.

2.3. Parasocial Interaction and Parasocial Relationship in Influencer Marketing

Research examining PSI and PSR between influencers and followers on social media within influencer marketing consistently finds that parasocial phenomena have a crucial impact on the effectiveness of influencer marketing. This effectiveness is typically measured by purchase intention [28,29,30], brand trust [31], brand interest [32], and engagement with brand content [33].

Followers who developed higher PSI with a vlogger showed stronger purchase intentions for the product presented in the video [34]. PSI between French followers and fashion bloggers positively influenced purchase intentions for the products they showcased [4]. A study on a sample of Chinese individuals aged 18 to 40 also confirmed that PSI between them and the influencers they followed positively affected purchase intentions [35]. Followers who developed a more intense PSI with a beauty YouTuber expressed higher purchase intentions for the products featured in the videos, indicating that PSI directly influences purchase intentions [36]. Women who developed stronger PSI with a fashion vlogger also showed greater purchase intentions for the featured products [37,38]. Farivar, Wang, and Yuan [9] found that PSR had a stronger impact on purchase intentions than followers’ perception of the SMI as an opinion leader.

In their study, Reinikainen [31] found that PSR between a vlogger and her followers increased trust in the brand she represented. Similarly, PSR developed by followers towards a YouTuber positively influenced their opinions about the products he presented [39]. PSR between followers and influencers also positively affects followers’ interest in the brand being promoted [32]. A study among social media users who follow various influencers showed that PSR positively impacts all levels of engagement with brand content [40]. Likewise, followers who developed a stronger PSR with a fitness YouTuber exhibited a stronger and more enduring intention to perform the exercises and anticipated more positive physical effects from the workout compared to those who did not develop PSR with the influencer [41].

All these studies demonstrate that PSI and PSR play a very important role in influencer marketing.

3. Methodology

Selection of Studies

We conducted a systematic review using the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) [42] of empirical studies that included the measurement of PSI and PSR in influencer marketing. We searched for articles in which the constructs PSI or PSR were among the variables in different models. These variables (PSI and PSR) were measured using various scales.

In the identification phase, we selected the Scopus database for our article search, as it is recognized as the largest multidisciplinary database of peer-reviewed scientific articles in the social sciences [43]. We accessed Scopus through the Digital Portal of the University of Primorska.

The literature review was carried out in December 2023 and included articles published up to 14 December 2023. We initiated the search using two keywords. The first keyword was “parasocial”. To encompass two forms of the parasocial phenomenon (relationship, interaction), we did not add any additional terms to this keyword, as doing so would have limited the search.

Since our study focuses on influencer marketing on social media, we chose “social media” as the second keyword. For the first 60 years, from 1956, when the term parasocial was first used [16], until 2016, researchers primarily focused on film and television [23]. The field of social media has become relevant only in recent years. Therefore, we limited our selection to studies examining the parasocial phenomenon on social media. The selected keywords were connected using the Boolean operator “AND”.

In the initial search and identification phase, we retrieved 3677 results. Since we focused exclusively on the Scopus database, there were no duplicates to remove during the identification phase. In the next step, the screening phase, we limited the results to fully published, peer-reviewed scientific articles in English, reducing our selection to 1903 articles ready for eligibility assessment.

In the eligibility assessment phase, we assessed the content relevance of the articles. We read the titles and abstracts to determine whether the articles covered the scope of our study: (1) the relationship between influencers and followers on social media and/or (2) the field of influencer marketing on social media. After this review, we excluded 1718 articles, leaving 185 for further review.

We then thoroughly read the remaining articles and included quantitative studies that empirically examined the parasocial phenomenon between followers and influencers on social media using one of the scales for measuring the parasocial phenomenon. We excluded any quantitative studies that measured PSI or PSR but did not provide sufficiently detailed information on the scales used. Specifically, we excluded studies that did not provide enough precise data to determine which items and how many were included in the PSI and/or PSR scales they used. At the end, 72 articles were included. The selection of articles is shown in Figure 1.

We were left with 72 studies for a detailed analysis of scales used to measure the parasocial phenomenon. The list of all examined studies is provided in Appendix A, and the lists of all scales is in Appendix B.

After selecting studies suitable for examining the scales, we read each one in full and extracted the data of interest from each study. The data included the following:

• Authors
• Title of the article
• Journal
• Year of publication
• The parasocial phenomenon measured in the study (PSI or PSR)
• Questions in the scale
• Number of questions in the scale
• The predominant scales from which the questions were derived
• Whether the scale includes questions not derived from established scales

4. Results

Out of the 72 studies examined, 32 focus on PSI, and 40 focus on PSR. See Figure 2.

The first study included in our review (Scopus database) was conducted in 2015. From 2015 to 2019, three or fewer studies were conducted annually. However, the number of studies began to increase sharply thereafter, with the highest number of studies conducted in the most recent year. See

Table 1. Number of studies conducted per year.

Year	Total Number of Research	Research Measuring PSI	Research Measuring PSR
2015	1	0	1
2016	1	1	0
2017	1	0	1
2018	3	2	1
2019	3	1	2
2020	12	6	6
2021	15	6	9
2022	12	5	7
2023	24	11	13

.

In the 72 studies examined, 55 different versions of scales were used, varying in both the number of items and the content of the items. We defined scales as different if they differed in the number and content of items. If two scales contained the same number of items but differed in the content of only one item, we treated them as two different scales.

If items differed only in the subject’s reference, for example, “my favorite soap opera character makes me feel comfortable, as if I am with a friend” and “the YouTube blogger makes me feel comfortable, as if I am with friends”, we treated this as the same item. However, if the items also differed in meaning, for example, “I think the YouTube blogger is like an old friend”, we treated them as two different items and consequently as two different scales. Given that different items can indicate different operational definitions, we also classified scales that differ by only one item as distinct scales.

4.1. Parasocial Interaction Scales

In the 32 studies that measured PSI, we identified 26 different scales. The list of all scales and the studies in which they were used is provided in Appendix B.

The 26 examined PSI scales vary significantly in length, containing between two and 13 questions. The most common scales have five questions (four scales), followed by those with three, four, eight, or nine questions (three scales each). Only two scales contain two questions, and one scale contains 13 questions. See

Table 2. Number of questions in the PSI scales and the number of scales containing that many questions.

Number of Items	The Number of Scales Containing This Many Questions	Scales Containing This Many Questions
2	2	PSI 4, PSI 20
3	3	PSI 16, PSI 17, PSI 23
4	3	PSI 11, PSI 18, PSI 22
5	4	PSI 1, PSI 7, PSI 13, PSI 19
6	2	PSI 6, PSI 8
7	2	PSI 9, PSI 24
8	3	PSI 2, PSI 21, PSI 26
9	3	PSI 3, PSI 14, PSI 15
11	2	PSI 12, PSI 25,
12	1	PSI 10
13	1	PSI 5

.

The 26 different scales were derived from various established scales (PSI Scale, Audience Persona Scale, PSI-Process Scale, Multiple PSR Scale, EPSI Scale) and included a wide range of items. The scales combined items from established scales and items specifically developed for the particular scale in different ways:

• Thirteen scales derived all their items from one of the established scales.
• Five scales combined items from an established scale with items specifically developed for the scale or taken from other less established scales.
• Five scales drew items from multiple established scales.
• Two scales combined items from multiple established scales and added items specifically developed for the scale or taken from other less established scales.
• One scale consisted solely of items not derived from any established scales.

Among all the scales:

• Twenty-three scales included items from the PSI Scale [17].
• Eight scales included items from the Audience-Persona Scale [19].
• Two scales derived items from the EPSI Scale [22].
• One scale derived items ns from the PSI-Process Scale [20].
• One scale consisted solely of items that were not part of any established scales.
• Eight studies used additional, specially developed questions for the study alongside one or more established scales.

4.2. Parasocial Relationship Scales

In the 40 studies that measured PSR, we identified 30 different scales. The list of all scales and the studies in which they were used is provided in Appendix B.

The PSR scales also varied significantly in length. The number of items ranged from three (one scale) to 22 items in one scale (one scale). A detailed breakdown of the number of items in the scales is presented in

Table 3. Number of questions in the PSR scales and the number of scales containing that many questions.

Number of Items	The Number of Scales Containing This Many Questions	Scales Containing This Many Questions
3	1	PSR 18
4	1	PSR 12
5	3	PSR 10, PSR 13, PSR 20
6	7	PSR 3, PSR 7, PSR 8, PSR 15, PSR 17, PSR 23, PSR 24
7	3	PSR 21, PSR 26, PSR 27
8	4	PSR 4, PSR 22, PSR 28, PSR 29
9	2	PSR 5, PSR 25
10	1	PSR 1
12	1	PSR 6
13	5	PSR 9, PSR 11, PSR 14, PSR 16
16	1	PSR 19
22	1	PSR 2

.

In the 40 different scales measuring PSR, we identified items derived from established scales or specifically developed for the particular scale.

• Twelve scales derived all their items from one established scale.
• Twelve scales combined items from one established scale with additional items specifically developed for the particular study or derived from other less established scales.
• One scale derived items from multiple established scales.
• Three scales combined items from multiple established scales with additional questions specifically developed for the particular study or derived from other less established scales.
• One scale consisted of items that were not found in any previously established scales.

The scales used to measure PSR between influencers and followers on social media drew questions from the following established scales:

• Fourteen scales drew items entirely or partially from the PSI scale [17].
• Four scales drew items entirely or partially from the Audience Persona Scale [19].
• Two scales drew items entirely or partially from the Multiple-PSR Scale [21].
• One scale drew items from the PSI Process Scale [20].
• One scale consisted of items that did not appear in any of the established scales.

4.3. Similarities in Parasocial Interaction and Parasocial Relationship Scales

In measuring both PSI and PSR between influencers and followers in influencer marketing, we find that both phenomena are assessed using a large number of different scales. PSI is measured with 26 different scales, and PSR with 29 different scales, which vary in the number and origin of items.

We observe that the scales for measuring PSI and PSR draw from the same established scales: 23 PSI scales and 15 PSR scales draw from the PSI Scale [17], eight PSI scales and four PSR scales draw from the Audience-Persona Scale [19], and one scale for each PSI and PSR draws from the PSI Process Scale [20].

Upon detailed examination of all PSI and PSR scales, we found that two scales were identified as both PSI and PSR scales. In some studies, these scales were used to measure PSI, while in others, they were used in identical form to measure PSR. See

Table 4. Scales and their items used for measuring both PSI and PSR.

Articles Where Scale Was Used to Measure PSI	Scale/Items	Articles Where Scale Was Used to Measure PSR
[30,36,44,45,46,47,48]	PSI 2 = PSR 4
	I look forward to watching the YouTube blogger on her YouTube channel.	[31,49,50]
	If the YouTube blogger appeared on another YouTube channel, I would watch that video
	When I’m watching the YouTube blogger, I feel as if I am part of her group.
	I think the YouTube blogger is like an old friend.
	I would like to meet the YouTube blogger in person.
	If there were a story about the YouTube blogger in a newspaper or magazine, I would read it.
	The YouTube blogger makes me feel comfortable, as if I am with friends.
	When the YouTube blogger shows me how she feels about the luxury brand, it helps me make up my own mind about the brand.
[51]	PSI 8 = PSR 7	[52]
	The unboxer makes me feel comfortable, as if I am with a friend.
	When I interact with the unboxer, I feel included.
	I can relate to the unboxer.
	I like hearing what the unboxer has to say.
	I care about what happened to the unboxer.
	I hope the unboxer can achieve their goals.

.

5. Discussion

The aim of this study was to review and examine the scales used to measure PSI and PSR between influencers and followers in influencer marketing and to answer the main research questions:

• Which scales are used to measure PSI and PSR in influencer marketing?
• Are there predominant scales for measuring PSI and PSR in influencer marketing?
• Do the measurement instruments for PSI and PSR in influencer marketing account for the distinction between PSI and PSR?

To achieve this, we reviewed the existing seminal literature on the measurement of PSI and PSR. We found that there has historically been considerable theoretical and empirical confusion in the conceptualization and measurement of PSI and PSR. Since their first mention, PSI and PSR have undergone several refinements and conceptual enhancements, as have the methods for measuring these constructs. Liebers notes that over 60 years of research, there has been significant heterogeneity in the ways PSI and PSR have been measured.

Despite Horton and Wohl [16] distinguishing between PSI and PSR, this distinction has become blurred, particularly after the publication of the PSI Scale, with the terms sometimes being used interchangeably. Rubin, Perse, and Powell [17] defined PSI as social involvement, which includes both interaction and long-term identification. The items included in their 20-item scale from 1985 and the shorter 10-item version from 1987 do not distinguish between the concepts of interaction and long-term relationship. Dibble, Hartmann, and Rosaen [24] emphasize that the scale primarily measures PSR rather than PSI.

At the beginning of the new millennium, several scholars began emphasizing the need to distinguish between PSI and PSR and developed new scales that incorporated recent theoretical advancements in these concepts. With the emergence of social media influencers and influencer marketing, PSI and PSR have become highly relevant concepts for describing interactions and relationships between influencers and followers. Research has shown that PSI and PSR between influencers and followers play a crucial role in the effectiveness of influencer marketing [34,35,37,38]. Therefore, we decided to examine the measurement instruments for PSI and PSR in influencer marketing.

We found that both PSI and PSR are measured using a considerable number of scales. In the 40 studies that examined PSR between influencers and followers on social media, 29 different scales were used. Researchers studying PSI in 32 studies used 26 different scales. For both PSI and PSR, the scales vary in content and length.

The shortest scales for measuring PSI have only two items, while the longest have 13 items. The shortest scales for measuring PSR have three items, while the longest have 22 items. The scales also differ in content, drawing from various established scales developed in the past. They may also include items from other less established scales or items specifically developed for the particular study.

The scales used to measure PSI draw their items from the PSI Scale [17], Audience-Persona Scale [19], EPSI Scale [22], PSI Process Scale [20], and several other less established scales, or they are entirely composed of items that do not belong to any established scale.

A similar practice is observed in the measurement of PSR. The scales for measuring PSR between influencers and followers on social media draw from the PSI Scale [17], Audience-Persona Scale [19], Multiple PSR Scale [21], PSI Process Scale [20], and various other less established scales. In this case, we also found one scale entirely composed of items that do not belong to any established scale.

The degree of differentiation among the 26 different scales for measuring PSI and the 29 different scales for measuring PSR varies. Some scales are very similar and differ by only one item. For example, the PSI 26 and PSI 15 scales differ by just one item. Similarly, the PSR 4 and PSR 15 scales both contain six identical items, with PSR 4 having an additional two items, all derived from the PSI Scale.

On the other hand, scales can be entirely different, such as PSI 1 and PSI 5, which differ both in length and content. PSI 1 has five items, while PSI 5 has 13 items, with no overlapping items between the two scales. A similar situation is observed in PSR measurement; for instance, PSR 18 has three items, whereas PSR 2 has 22 items, with no overlapping items. This raises questions about whether a scale with just two items can measure the same phenomenon to the same extent as a scale with 22 items.

We found that there is a significant lack of standardization and uniformity in the measurement of PSI and PSR in influencer marketing. Both PSI and PSR are measured using a large number of different scales that vary in content and length. There could be two reasons for this diversity. First, the nearly 40-year period of developing different scales has provided researchers with a wide range of tools for measurement, which they further adapted and combined in ways that suited their specific studies. It seems that the established practice among researchers studying PSI and PSR in influencer marketing is to select one or more established scales, choose the items they find suitable for their specific study, and add other items as needed. As a result, PSI was measured with 26 scales in 32 studies, and PSR was measured with 29 scales in 40 studies.

Another reason for such heterogeneous measurement of both concepts is the lack of clear and logical conceptual clarification of PSI and PSR and, consequently, the empirical clarification of what PSI and PSR actually are in influencer marketing and which dimensions they encompass.

Due to the significant heterogeneity in measurement, it raises the question of whether the findings from studies that measure PSI and PSR in such diverse ways are even comparable.

The answer to the research question of if there are predominant scales for measuring PSI and PSR in influencer marketing is no. As explained earlier, the measurement tools for PSI and PSR in influencer marketing are numerous and highly varied. The only predominant feature we observed in both PSI and PSR scales is that the majority of scales we studied draw from the PSI Scale. Among the 26 scales measuring PSI, 23 derive items from the PSI Scale, and among the 29 scales measuring PSR, 14 draw from the PSI Scale. Thus, we can conclude that there is no single predominant scale for measuring PSI and PSR in influencer marketing, but items from the extensive 20-item PSI Scale frequently appear in many scales. Therefore, the scale that has been the subject of much criticism for allegedly measuring PSR instead of PSI, as its name suggests, still serves as a common basis for developing new scales for both PSI and PSR measurement.

The reasons for drawing from the same scales are likely simplicity and trust. Researchers rely on frequently used scales from the past when measuring the concepts included in their studies.

In conclusion, we find that the scales used to measure PSI and PSR in influencer marketing do not differentiate between the two distinct constructs of PSI and PSR, despite calls from numerous authors over the past 20 years to do so. While many of the articles we reviewed differentiate between PSI and PSR at the theoretical level and briefly explain the main differences in the introduction or literature review, this distinction is not reflected at the empirical level.

As noted above, the scales for measuring PSI and PSR often draw from the same sources. Most frequently, scales draw from the PSI Scale, followed by the Audience-Persona Scale—eight scales for measuring PSI and four scales for PSR. Additionally, one scale for measuring PSI and one scale for measuring PSR draw from the PSI-Process Scale.

The issue extends further. Dibble, Hartmann, and Rosaen, as well as Schramm and Hartmann, emphasize that the PSI Scale primarily measures long-term PSR rather than PSI. The Audience-Persona Scale, according to its authors, measures “positive long-term involvement”, yet items from this scale appear in eight scales for measuring PSI. Simultaneously, one scale for measuring PSR draws from the PSI-Process Scale, which was designed to measure PSI.

We also identified two scales that are used in identical form to measure both PSI and PSR. The first scale was used to measure PSI in seven articles and PSR in three articles. The second scale was used once to measure PSI and once to measure PSR.

This raises questions about whether these scales truly measure two different constructs or what they are actually measuring. Do the scales intended to measure PSI between influencers and followers, which draw from scales originally meant to measure PSR, genuinely measure PSI? Do these scales, based on such items, possess sufficient content validity?

To conclude, the current problem with measuring both PSI and PSR between influencers and followers is twofold. Firstly, the measurement of both constructs is overly heterogeneous, as it involves a large number of highly varied scales, differing so much that it questions the comparability of study results.

Secondly, there is a lack of conceptual and empirical differentiation between the constructs, leading to two different constructs being measured by scales that can be very similar or even identical. We believe that it is not possible to measure two different constructs with the same or very similar instruments.

Limitations and Future Research

In our literature review, we limited our scope to the Scopus database and to scientific articles published in English. We used two key search terms in our investigation; however, employing a broader set of keywords might have captured a larger number of scales and uncovered additional measures used for assessing PSI and PSR. Regarding the field, we focused specifically on influencer marketing. Future studies could expand the scope of investigation to include other areas. It would also be worthwhile to explore how PSI and PSR on social media differ from PSI and PSR in traditional media, as social media platforms offer significantly more opportunities for interaction compared to traditional media.

We suggest that future research should focus on creating clearer differentiation between PSI and PSR and on developing reliable and valid measurement scales that accurately measure the constructs of PSI and PSR. These scales should then be used in empirical studies in as unchanged a form as possible. This approach will enable easier comparability of research results that use the same scales. Additionally, a comprehensive literature review focusing on the conceptualization, dimensionality, itemization, and underlying theoretical perspectives of PSI and PSR in influencer marketing would be valuable in further refining and standardizing these constructs.

Currently, both constructs are measured with too many varied scales, which are too similar across constructs.